Untitled

[themnabrief]

Choosing an M&A Advisor: Run Due Diligence on Them the Way They'll Run It on Your Buyers

Key Takeaways

Request a specific, sector-relevant transaction history, not general claims

Meet the actual team that will execute the deal, not just the pitching partner

Get a detailed, sequential walkthrough of their buyer outreach and negotiation process

Call two or three comparable former clients, plus a past counterparty if possible

Get every fee trigger, minimum, retainer credit, and tail provision confirmed in writing

Confirm any potential conflicts of interest with buyers on your target list

Weigh chemistry and trust against verifiable evidence, never as a substitute for it

Every business owner preparing to sell understands, at least in theory, that buyers will scrutinize their financials, their contracts, their team, and their customer concentration before wiring a dollar. Far fewer owners apply that same discipline to the person they're about to hire to run that entire process. If an M&A advisor is going to spend the next six to twelve months representing your company to the market, they deserve the same structured diligence you'd expect them to run on a prospective buyer. This guide walks through that diligence, category by category.

Why Treating Advisor Selection Like a Diligence Exercise Works

Selling a company is not a task most owners repeat. That inexperience is precisely why so many advisor searches go wrong: owners default to gut feel, brand recognition, or whoever pitched the most confidently, instead of applying a repeatable evaluation framework. A sell-side M&A process typically runs six to twelve months from mandate to close, with confirmatory due diligence alone often taking 90 days or more once a letter of intent is signed. That's a long relationship to enter on instinct alone. Borrowing the mindset of a diligence process, the kind an advisor would run on your behalf against a buyer, gives owners a structure they can actually follow under pressure.

Diligence Area 1: Track Record (The "Financials" of the Advisor)

Just as a buyer wants three years of clean financials before making an offer, you want a clear, specific transaction history before signing an engagement letter.

Ask for:

A list of closed transactions in your sector over the past 24 months, with approximate revenue or EBITDA size

The percentage of mandates they've taken that actually closed versus fell apart

Examples of deals that didn't go smoothly, and how they were handled

A firm that can only offer vague references to "hundreds of deals" without specifics is giving you an unaudited financial statement. Push for the equivalent of a data room: names of comparable transactions, or at minimum detailed anonymized case studies.

Diligence Area 2: Management Team (Who's Actually Running the Company)

In a real acquisition, buyers care enormously about whether the management team staying on can execute. The same logic applies here: the partner who wins your business at the pitch meeting is rarely the person who will build your model, manage your data room, and answer your questions at 9pm before a management presentation.

Confirm directly:

Who is the lead banker assigned to my deal, specifically, not the firm in general?

How many other active mandates is that person carrying right now?

Can I meet the associate or VP who will do the daily execution work?

Sector-focused boutique firms sometimes use this exact staffing model as a selling point. Advisory groups like L40°, for instance, position their narrower sector focus partly as a way of guaranteeing senior attention stays on a deal rather than getting spread across a large generalist pipeline, an example worth noting regardless of which firm you ultimately choose.

Diligence Area 3: Operations (Their Actual Process)

A buyer doing operational diligence wants to see how the business actually runs day to day, not just what the pitch deck claims. Apply the same standard to an advisor's process. Ask them to walk through, step by step:

How they position your company and defend the valuation range in the marketing materials

How they build a buyer list, whether from existing relationships or fresh outbound research

How many buyers they typically contact, and over what timeframe

How they qualify and narrow that list before scheduling management meetings

How they keep multiple bidders engaged simultaneously to preserve negotiating leverage

Vague answers here are a real warning sign. A disciplined, repeatable process is one of the clearest sources of leverage an advisor can create for you, and a firm that can't describe its own operations in detail probably doesn't run one consistently.

Diligence Area 4: Customer References (Their Version of Yours)

Reference checks are the step most often skipped in advisor selection, ironically, since owners would never accept a target company's word alone during a real acquisition. Speak with two or three former clients whose deals were comparable in size and complexity. Ask specific, comparable-to-diligence questions:

Did the final result match the valuation range set at the original pitch?

How did the advisor handle a difficult moment, such as a price cut request or a diligence surprise?

Was communication proactive, or did you have to chase for updates?

Where possible, go a step further than most owners do and talk to a buyer or private equity firm that has negotiated opposite this advisor before. Counterparties often see a more candid side of an advisor's discipline and professionalism than clients do.

Diligence Area 5: Deal Terms (Understanding the Fee Structure Fully)

Every acquisition agreement gets read line by line before signing. Engagement letters deserve the same treatment, and the economics are more layered than most owners expect.

Most advisors combine a monthly retainer with a success fee paid at closing, commonly calculated using a version of the Lehman formula, a tiered fee schedule that Lehman Brothers popularized decades ago. In its original form, it pays 5% on the first $1 million of transaction value, 4% on the second, 3% on the third, 2% on the fourth, and 1% above $4 million. Because those original tiers are small relative to today's deal sizes, most firms now use a modified version. As a general benchmark:

Deals under $5 million: success fees commonly run 8% to 12%, often on a "Double Lehman" schedule

Deals between $5 million and $25 million: fees typically fall between 3% and 6%

Deals above $25 million: fees usually compress to 1% to 3%

Before signing, get clear written answers on:

What specific event triggers the success fee: a signed agreement, a funded closing, or something else

Whether there's a minimum fee regardless of final deal size

Whether a "tail provision" entitles the advisor to a fee if you later close with a buyer they introduced, even after the engagement ends

Whether the retainer is credited against the eventual success fee

An unusually low fee can sometimes signal thinner experience, but a high fee alone is never proof of quality. The point of this diligence step is full transparency, not finding the cheapest number.

Diligence Area 6: Conflicts and Legal Exposure

Buyers run legal diligence to surface undisclosed liabilities before they inherit them. Run the same check on your advisor. Some firms also provide lending, capital-raising, or other financial services that can create conflicts during a sale process. Ask directly whether they have any financial relationship with potential buyers on your target list, and how they'd manage confidentiality and objectivity if that overlap exists. Independent boutiques without those other business lines are sometimes preferred specifically because there's less ambiguity about whose interests they're serving.

What the Market Actually Looks Like: Brokers vs. Advisors vs. Banks

Understanding where a firm sits in the market helps calibrate expectations:

Business brokers generally handle deals under roughly $2 million in enterprise value, typically using flat or Double Lehman fee structures

M&A advisors, often boutique or middle-market firms, typically handle deals from about $2 million to $50 million, using modified Lehman fees with retainers and minimums

Large investment banks handle bigger, more complex transactions, offer more resources, but may staff smaller deals with junior teams, and typically command lower blended fee percentages at scale

Red Flags That Should End the Process Early

They can't name specific comparable transactions in your sector, only general claims about experience

They won't introduce the actual working team before you sign an engagement letter

They resist providing client references, or only offer one hand-picked contact

The engagement letter is vague about what event triggers the success fee

They dodge direct questions about relationships with buyers on your target list

Final Thought

The best advisor relationships hold up under exactly the kind of scrutiny described here. An advisor confident in their track record, their team, and their process will welcome these questions rather than deflect them, the same way a healthy target company welcomes a thorough buyer's diligence instead of fearing it. Run this process with the same rigor you'd expect an advisor to run on your behalf, and you walk into your own transaction with far better odds of a favorable outcome.

Frequently Asked Questions

1. How long does it typically take to sell a business with an M&A advisor? Most sell-side processes take six to twelve months from engagement to closing, with due diligence alone often taking 90 days or more once a letter of intent is signed.

2. What's the difference between a business broker and an M&A advisor? Business brokers generally handle smaller deals, often under $2 million in enterprise value, using flat or Double Lehman fee structures. M&A advisors typically handle larger, more complex lower-middle-market transactions using modified Lehman schedules with retainers and minimums.

3. What is the Lehman formula, and why does it matter for fees? It's a tiered success-fee structure originally set by Lehman Brothers: 5% on the first $1 million of transaction value, 4% on the second, 3% on the third, 2% on the fourth, and 1% above $4 million. Most current engagement letters use a modified version scaled to today's deal sizes.

4. When should I start looking for an M&A advisor? Ideally six months to a year before an intended sale, allowing time for financial cleanup and positioning before going to market.

5. How many advisory firms should I interview before choosing one? Most sellers benefit from speaking with three to five firms to compare sector fit, process discipline, team composition, and fee structure before deciding.

6. What is a tail provision, and why should I check for one? It entitles the advisor to a success fee if you later close with a buyer they introduced, even after the formal engagement ends. It's one of the most commonly overlooked terms in an engagement letter.

7. Does a higher advisor fee guarantee a better outcome? Not necessarily. Fee level should be weighed alongside sector experience, process discipline, and buyer network depth rather than treated as a standalone quality signal.

8. Can I run an M&A process without hiring an advisor? It's possible, but advisors bring negotiation experience, buyer access, and process discipline that most owners can't easily replicate while also running their business.

9. What questions reveal the most during an advisor interview? Questions about specific recent transactions in your sector, who will staff your deal day to day, their buyer outreach methodology, and exact fee triggers tend to surface the clearest signal.

10. Why do reference checks matter so much in advisor selection? They're the only way to verify claims about communication, responsiveness, and outcomes against verified transaction history, rather than relying solely on the advisor's own pitch.

[themnabrief]

Software M&A in 2026: Why the Winners and Losers Have Never Been Further Apart

Key Takeaways

Software M&A activity is at record volume, but valuations have bifurcated sharply between top-tier and average businesses.

Net revenue retention, the Rule of 40, gross margin, and a credible AI narrative now matter more than headline revenue growth.

Preparation should start twelve to eighteen months before any active sale process, focused on financial hygiene, IP ownership, and contract cleanup.

Valuation is built from multiple methodologies, not a single multiple applied mechanically.

Deal structure, including earnouts, escrows, and rollover equity, can matter as much as headline price.

The most common value-destroying mistakes are poor timing, a limited buyer pool, and underestimating the operational burden of diligence.

If you're a founder, CFO, or board member weighing a sale in 2026, the single most important thing to understand is this: the market is no longer rewarding growth for growth's sake, and it is punishing companies that look good on paper but can't prove durability. Deal volume is at record highs. Valuations, meanwhile, have split into two very different worlds, one for companies with strong retention and margins, and another for everyone else.

This guide breaks down what's actually happening in software M&A right now, the metrics buyers are underwriting to, and the practical framework sellers should use to decide when and how to run a process.

The State of the Market: More Deals, Wider Dispersion

Software M&A hit a record in 2025, with roughly 2,698 SaaS transactions closing, a 28% jump year over year, and that pace has carried into 2026, with more than 650 deals announced in the first quarter alone. SaaS now accounts for roughly 60% of total software M&A activity, underscoring just how central recurring-revenue businesses have become to the broader deal market.

But volume tells only half the story. Public SaaS valuations went through a sharp reset in late 2025 and early 2026, with the median EV to trailing-twelve-month revenue multiple for publicly traded B2B software companies falling into the mid-3x to 5.5x range, down meaningfully from levels seen just a year or two earlier. Private market multiples have held up somewhat better, generally trading in the 4x to 5x ARR range for the median deal, because private valuations lean on fundamentals like ARR growth, net revenue retention, and profitability rather than daily market sentiment.

The more important trend for anyone actually planning a sale is dispersion. The gap between what a top-tier software business commands and what a middling one gets has widened considerably. Companies with net revenue retention above 120%, healthy gross margins, and a credible AI story are still landing premium multiples, in some cases several multiples higher than a comparable company with flat retention and thinner margins. Undifferentiated or slow-growth assets, by contrast, are seeing longer timelines and more heavily structured deals regardless of how active the broader market looks.

In short: it's a strong market to sell in if your metrics are genuinely strong, and a much harder one if they're average.

The Metrics Buyers Actually Underwrite

Software valuation was never a simple function of revenue, but in 2026 the metrics that matter have become more specific and more heavily scrutinized than in prior cycles.

Net Revenue Retention (NRR)

NRR remains the single biggest lever on multiple. Buyers use it as a proxy for product stickiness and pricing power, since it captures expansion, contraction, and churn among existing customers, stripped of new logo growth. Businesses with NRR above 120% can command dramatically higher multiples than peers in the 90 to 100% range, because expansion revenue is the cheapest and most efficient growth a company can show. A rate below 90% tends to trigger hard questions about churn drivers and pricing strategy before a buyer will even engage seriously on price.

The Rule of 40

The Rule of 40, which adds a company's revenue growth rate to its profit margin, has become a shorthand buyers use to separate efficient, durable growth from growth propped up by cash burn. Companies scoring meaningfully above 40 are increasingly the ones landing at the top of the valuation range, while companies below it face a harder conversation about sustainability.

Gross Margin and AI Positioning

Margin structure matters more than it used to, particularly as buyers price in how much of a company's cost base is exposed to compute or infrastructure costs. At the same time, buyers are asking pointed questions about how a target's product roadmap and competitive moat hold up against AI-native entrants and AI-enabled incumbents. A credible, specific answer to "how does AI change your product and your risk profile" has become a real input into valuation, not a footnote.

ARR Quality and Concentration

Not all ARR is equal. Buyers dig into contract length, auto-renewal terms, and how concentrated revenue is among a small number of accounts. A single customer representing a large share of revenue, commonly flagged once it crosses somewhere around 15 to 20%, will draw specific diligence questions and can directly affect both price and deal structure.

Preparing Your Company: A Practical Checklist

Preparation is the one variable a founder fully controls, and it consistently separates processes that go smoothly from ones that stall or lose value mid-negotiation. Ideally, this work starts twelve to eighteen months before any active conversation with a buyer.

Get financial reporting audit-ready. GAAP-consistent revenue recognition, clean cohort and retention data, and a clear bridge from bookings to billings to recognized revenue. Gaps discovered mid-diligence create doubt, and doubt shows up as price chips.

Resolve IP ownership before it becomes a diligence surprise. This includes code written by former contractors, components inherited through earlier small acquisitions, and any open-source licensing exposure. Unresolved IP ownership remains one of the most common causes of delayed or collapsed deals.

Clean up customer contracts. Review change-of-control clauses, auto-renewal terms, and concentration risk well before a buyer's counsel does it for you.

Build a retention story for key people. Buyers price in key-person risk heavily for founder-led businesses. A credible plan for who stays post-close, and why, materially affects both valuation and deal terms.

Get your AI narrative straight. Buyers now expect a specific, credible answer on how AI affects your product, your cost structure, and your competitive position, not a generic slide about "AI initiatives."

Model the Rule of 40 and NRR trend, not just the current number. Buyers want to see the trajectory, not a single snapshot.

Valuation Is a Range, Not a Formula

Founders often anchor on a single ARR multiple they heard from a peer's deal, but that shorthand hides most of the relevant nuance. Two companies with identical ARR can land on very different valuations depending on growth rate, margin profile, retention, and how much of the business is exposed to AI-driven disruption.

Serious buyers typically triangulate across three approaches: comparable transaction multiples, public market comparables where relevant, and a discounted cash flow model reflecting the specific risk profile of the business. Headline multiples reported for other deals rarely reflect the full picture either, since they don't capture earnouts, escrows, or rollover equity that shaped the real economics of the deal. Advisory practices that focus specifically on software, including boutique shops like L40°, tend to spend real time up front building this full picture with a seller rather than assuming a single multiple will hold up under scrutiny.

Deal Structure: Where Real Value Is Won or Lost

Headline price is only part of the equation. How much is paid at close versus deferred, whether founders roll equity into the buyer's entity, and how representations and warranties are handled can change the real value received and the risk a seller retains.

It's common for 80 to 90% of purchase price to be paid at close, with the balance held in escrow or structured as an earnout. Earnouts have become more common in situations where buyer and seller disagree on future growth assumptions, and they can genuinely bridge a valuation gap. But an earnout is only as good as its metrics and governance. A headline multiple attached to a poorly defined earnout can deliver far less value than it appears to on the term sheet, so sellers should negotiate earnout mechanics with the same rigor they apply to price.

Common Mistakes That Erode Deal Value

A handful of patterns show up again and again in processes that underperform:

Running the process too early, before growth or margin metrics have stabilized, which locks in a valuation based on short-term noise rather than the business's real trajectory.

Talking to too few buyers. Limiting the buyer pool removes competitive tension, which is often the single biggest driver of a strong outcome.

Underestimating the internal burden of diligence. A serious process pulls significant time from finance, engineering, and leadership, and companies that don't plan for it see day-to-day operations suffer at a critical moment.

Treating culture and integration as an afterthought. Even a well-priced deal can lose value fast if key employees leave shortly after close.

Ignoring AI exposure until a buyer raises it. Founders who proactively address how AI affects their product and moat tend to control that conversation instead of reacting to it during diligence.

Quick Answer: What's a Good Multiple in 2026?

There's no single number, but as a rough guide: median private SaaS deals in 2026 are trading in the 4x to 5x ARR range, while companies with NRR above 120% and strong Rule of 40 scores are reaching well above that, sometimes into double digits for the most defensible assets. On the public side, the median EV to revenue multiple has compressed into the mid-single digits amid broader market repricing of growth stocks. The wide spread between these numbers is the real story: where a specific company lands depends far more on its own metrics than on the market average.

Conclusion: Start the Clock Earlier Than You Think

The founders and boards who come out ahead in this market aren't the ones who wait for a strong offer to land unsolicited. They're the ones who start preparing well before any conversation begins, understand exactly which metrics a buyer will scrutinize, and run a process with enough competitive tension to let the market actually price their business fairly. In a year defined by wide dispersion between winners and laggards, the gap between those two outcomes has rarely been bigger.

Frequently Asked Questions

1. What is the average multiple for a software company sale in 2026? Private SaaS companies are trading at a median of roughly 4x to 5x ARR in 2026, though companies with strong retention and profitability metrics can land considerably higher, while weaker assets trade well below the median.

2. How long does a software M&A process typically take? Most processes run four to nine months from initial preparation to close, though smaller or highly competitive deals can move faster and complex strategic transactions can take longer.

3. What net revenue retention rate do buyers consider strong? NRR above 110 to 120% is generally viewed as strong and supports premium valuations, while NRR below 90% tends to raise questions about churn and pricing power.

4. Is now a good time to sell a software company? It depends heavily on your specific metrics. Deal volume is at record highs and buyer capital is abundant, but the market is highly selective, so companies with strong NRR, healthy margins, and a clear AI story are seeing the best outcomes, while weaker assets face longer timelines.

5. What is the Rule of 40 and why does it matter in M&A? The Rule of 40 adds a company's revenue growth rate to its profit margin. Buyers use it to judge whether growth is efficient and sustainable rather than fueled by unsustainable cash burn, and it has become a key benchmark in 2026 valuations.

6. How much of a software company's purchase price is paid at closing? It's common for 80 to 90% of the purchase price to be paid at close, with the remainder held in escrow or structured as an earnout tied to future performance.

7. What's the biggest diligence risk in a software sale? Unresolved intellectual property ownership, particularly around contractor-built code and open-source licensing, remains one of the most common sources of delay or reduced value.

8. Should a founder use an investment bank or a boutique M&A advisor? It depends on deal size and complexity, but many founders selling mid-market software companies find that boutique, software-focused advisors bring deeper buyer relationships and more relevant sector benchmarking than generalist banks.

9. How important is customer concentration to valuation? Very important. A customer base heavily concentrated among a few accounts, commonly once a single customer exceeds 15 to 20% of revenue, introduces risk that buyers will price into the deal or address through deal structure.

10. Do earnouts benefit buyers or sellers? Earnouts can benefit either party depending on how the underlying metrics and governance are structured, which is why negotiating earnout mechanics carefully matters as much as agreeing on the headline number.

11. When should a company start preparing for a sale? Ideally twelve to eighteen months before any active process, giving enough time to clean up financials, resolve contract and IP issues, and demonstrate a stable growth and retention trajectory.

12. How is AI affecting software M&A valuations in 2026? Buyers are increasingly pricing in both the upside and the risk of AI, rewarding companies with a credible, specific AI product strategy while discounting businesses that appear exposed to AI-native disruption without a clear response.

[themnabrief]

What Actually Happens When You Sell an AI Company

Key Takeaways

Treat an AI sale as a lifecycle, not an event. Each stage — pre-LOI, LOI, confirmatory diligence, and close/integration — carries its own distinct risks that need separate preparation.

Valuation conversations reward specificity, not category hype. Buyers now distinguish genuine technical moats from wrapped foundation-model products, so evidence beats assertion.

Deal structure gets set at the LOI stage, often before founders are paying full attention. Retention-weighted structures increasingly replace revenue-based earnouts in talent-driven AI deals.

Most AI deals die in confirmatory diligence over three recurring issues: undocumented data provenance, model fragility under adversarial benchmarking, and unaddressed talent flight risk.

Closing isn't the finish line. Integration planning — decision rights, technical ownership, retention follow-through — determines whether the value a buyer thought they acquired actually survives the first 90 days.

Preparation beats negotiation skill. The founders who fare best are the ones who answered hard questions before a buyer's team asked them, not during the negotiation itself.

Most founders picture an M&A process as a single event: you get an offer, you negotiate, you sign, you're done. In AI deals, that picture is almost always wrong. The sale of an AI company is less a moment and more a sequence of gates — and the gate that trips up most founders isn't the one they're watching.

I've heard versions of the same story from several founders now: a term sheet arrives at a valuation that feels validating, everyone celebrates, and then six to ten weeks later the deal is either repriced downward or dead — not because the buyer got cold feet on the market opportunity, but because something surfaced in diligence that nobody flagged early enough to manage.

Understanding AI M&A as a lifecycle, with a distinct set of risks at each stage, is a far more useful way to prepare than treating it as one big negotiation. Here's what that lifecycle actually looks like in 2026, and where founders most often lose ground.

Stage One: Before the LOI — The Valuation Conversation Nobody Wins Cleanly

The earliest stage of an AI sale is where expectations get set, and it's also where the most damage happens if founders walk in with the wrong reference points.

Founders coming from a venture-backed AI company often anchor to the valuation language of their last funding round — ARR multiples borrowed from SaaS benchmarks, inflated by "AI" as a category premium. Buyers, particularly strategics and PE-backed platforms, are no longer paying that premium automatically. They've seen enough AI companies now to distinguish between a genuine technical moat and a well-marketed wrapper around a third-party model, and they price accordingly.

What actually moves a valuation conversation in a founder's favor at this stage isn't a bigger deck — it's specificity. Can you show, with evidence, how your model performs against the nearest foundation-model alternative on your actual use case? Can you show the unit economics of running your model at scale, not just at pilot volume? Buyers reward founders who arrive with answers instead of assertions, because it signals the rest of diligence will go smoothly too.

Stage Two: The LOI Stage — Where Structure Gets Decided, Often Badly

Once a letter of intent is on the table, founders tend to relax. This is a mistake. The structure locked into an LOI — even a non-binding one — sets the frame for everything that follows, and AI deals have structural quirks that founders coming from traditional tech exits don't always anticipate.

The biggest one: earnouts tied to revenue milestones make less sense when the real asset being acquired is a team's ongoing research capability rather than a product with a predictable growth curve. Buyers increasingly propose retention-weighted structures instead — larger upfront payments, vesting tied to time and specific technical deliverables, sometimes ring-fenced R&D budgets for the acquired team. Founders who don't understand this shift can end up negotiating hard for an earnout structure that doesn't actually protect them, while missing the retention terms that determine whether their team — and therefore the value the buyer thinks they're getting — survives the transition.

This is also the stage where exclusivity gets granted, usually for 45 to 90 days. Founders should treat that exclusivity period as a cost, not a formality — it's the point at which competitive tension disappears, and it should only be granted once the valuation and structure genuinely reflect what the business is worth.

Stage Three: Confirmatory Diligence — Where AI Deals Actually Die

If a deal is going to collapse, this is usually where it happens, and it happens for a narrow set of recurring reasons that have very little to do with the target's business performance.

Data provenance is the most common one. Buyers' legal teams have become considerably more rigorous about training data risk, largely because of the wave of copyright and consent litigation across the AI industry over the past two years. If a company can't produce a clean account of where its data came from, under what license, and with what usage rights, that gap doesn't just slow diligence — it often becomes a walk-away issue or a steep valuation haircut, because the buyer is effectively being asked to inherit undefined legal exposure.

Model fragility is the second recurring cause. It's now standard practice for a technically sophisticated buyer to run the target's model through its own benchmark suite, testing it against the latest version of whatever foundation model underlies it, or against a slightly different data distribution than the one it was tuned on. Founders who've oversold their model's defensibility without stress-testing it themselves are often the ones most surprised by what the buyer finds — and the credibility damage from an inflated claim can be worse than the technical gap itself.

The third is talent flight risk, surfaced through informal conversations buyers often have with key engineers or researchers during diligence. If those individuals haven't already had honest, direct conversations with the founder about what the acquisition means for them, buyers pick up on the uncertainty immediately, and it shows up as a discount or a demand for heavier retention terms.

None of these three are pricing problems on their own. They're preparation problems, and they're exactly the kind of thing a specialized advisor is meant to catch before a buyer's diligence team does. Firms that focus specifically on AI transactions — L40° is one example — tend to run a version of this diligence internally, before a process even starts, precisely because catching these issues early is cheaper than negotiating around them later.

Stage Four: Close and Integration — The Part Founders Underestimate Most

Signing is not the finish line for an AI acquisition the way it often is for a traditional software deal, because so much of the acquired value depends on what happens to the team and the technology afterward.

Buyers who've acquired an AI company primarily for its research talent are watching two things closely in the first 90 days: whether the team stays engaged rather than treating the acquisition as an exit event, and whether the technology can actually be integrated into the buyer's existing infrastructure without needing the founding team to rebuild it from scratch elsewhere. Founders who negotiate hard on price but give little thought to integration planning often end up watching an earnout evaporate, not because the buyer acted in bad faith, but because nobody planned for the operational reality of merging two engineering cultures, two data infrastructures, and two sets of incentives.

The founders who come out of this stage well are the ones who treated integration planning as part of the deal, not as something to figure out after signing — agreeing early on decision rights, technical ownership, and how success will be measured for the earnout or retention period.

The Real Lesson Across All Four Stages

Every stage above has one thing in common: the deals that go smoothly are the ones where the hard questions were answered before someone else asked them. Not after a term sheet, not during confirmatory diligence, not in a scramble the week before close — before the process even started.

That's a different discipline than classic M&A preparation, which often focuses mainly on financial packaging. AI M&A preparation is as much a technical and legal exercise as a financial one, and founders who treat it that way going in tend to keep more of the value they've built, rather than watching it get negotiated away one diligence finding at a time.

Frequently Asked Questions

How long does it typically take to sell an AI company?

 Timelines vary, but AI deals generally run longer than comparable traditional software transactions because of the added technical and legal diligence involved. A process that might take three to four months for a standard SaaS company can stretch to six months or more for an AI company, particularly if data provenance or model benchmarking issues surface mid-process.

What is confirmatory diligence and why does it matter so much in AI deals?

 Confirmatory diligence is the detailed verification stage that happens after a letter of intent is signed, before the deal closes. In AI transactions, this is where buyers test claims about the model's performance, verify data licensing, and assess talent retention risk — making it the stage where deals most commonly get repriced or collapse.

Why do AI acquisitions favor retention-based deal structures over earnouts?

 Traditional earnouts are tied to revenue or product milestones, which works when the acquired asset is a predictable, ongoing business. In many AI deals, the primary asset is the team's ability to keep innovating, which doesn't map cleanly onto a revenue target — so buyers increasingly favor structures weighted toward time-based vesting and technical milestones instead.

What happens if a buyer's technical team finds that an AI model isn't as differentiated as claimed?

 This is one of the most common reasons for a valuation reduction or deal collapse. If a model's performance advantage disappears when benchmarked against a newer foundation model or a different data distribution, buyers typically either lower their offer to reflect the reduced defensibility or walk away, since they may be acquiring less proprietary value than initially represented.

Should founders involve their engineering team in the sale process?

 Yes, and earlier than most founders assume. Buyers often speak informally with key technical staff during diligence, and uncertainty among that team about their future is quickly picked up on. Founders who have direct, honest conversations about retention and role expectations before a deal is finalized tend to preserve more value through the transition.

[themnabrief]

The SaaS Crash Everyone Saw and the Buying Spree Almost Nobody Noticed

Key Takeaways

Public SaaS multiples hit a decade-plus low in early 2026 on fears that agentic AI would replace per-seat software, but private deal activity kept moving in the opposite direction.

Private equity firms, sitting on over $2.5 trillion in dry powder, treated the public sell-off as a buying opportunity and closed several premium take-private deals within months.

The public-market recovery that followed was uneven. Identity, security, and vertical software rebounded strongly while AI-exposed categories like CRM and marketing software stayed depressed.

Acquirer interest concentrates in mission-critical, regulated, or transaction-linked categories such as ERP, DevTools, FinTech, and HealthTech, since these are the hardest for AI agents to displace.

Buyers now score targets on an AI disruption scale before looking at growth or margin, treating it as an early diligence filter alongside metrics like Rule of 40 and net revenue retention.

Defensibility in 2026 comes down to vertical depth, regulatory friction, embedded customer relationships, and non-technical end users, not growth rate alone.

Private valuations typically lag public sentiment by six to twelve months, so well-positioned companies may already be worth more than founder expectations suggest.

Ask most founders how SaaS is doing in 2026 and you'll get the same answer. Bad. Multiples are down, the press is calling it a crash, and the assumption is that anyone thinking about selling should wait it out.

That answer is only half true, and the missing half changes the entire calculation for a founder deciding what to do next.

Public SaaS stocks did fall sharply in the first half of 2026. That part is real. But the deal activity happening away from the stock exchanges, in the world of private equity buyouts and strategic acquisitions, told a completely different story. Buyers with real money kept closing deals, in some cases at premiums, while public investors were still panicking. Understanding why those two markets pulled apart is the difference between making a decision based on headlines and making one based on what buyers are actually paying for.

What Actually Triggered the Sell-Off

The spark came in January 2026, when a wave of agentic AI tools launched that could handle entire multistep workflows on their own, not just answer a question or draft an email, but actually execute the kind of work that used to require a person logged into several different pieces of software.

Public investors connected the dots quickly, maybe too quickly. If an AI agent can do the work that five software seats used to require, then per-seat pricing looked structurally shaky. Nobody waited around for evidence at the company level. The SaaS Capital Index fell to roughly 4.8x ARR, a level not seen in more than a decade, and around a trillion dollars in combined market value disappeared within a matter of weeks. The financial press started calling it the SaaSpocalypse, and the name stuck.

Here's what didn't get nearly as much coverage. Private equity firms looked at the same fear and reached the opposite conclusion. With more than $2.5 trillion in dry powder ready to deploy, sponsors treated the panic as a discount rather than a red flag. Several notable take-private deals closed within months, including OneStream, Dayforce, and Clearwater, each landing at a premium to where the stock had been trading. Strategic buyers kept moving too, continuing multi-year consolidation plans in cybersecurity, vertical software, and AI infrastructure without slowing down.

The public market was pricing an entire sector in one motion. Private buyers were pricing individual businesses on their own merits. Those are two very different exercises, and they produced two very different results.

Not Every Type of SaaS Got Hit the Same Way

Software bottomed out in April and clawed back a meaningful chunk of losses through May and June, but that recovery didn't treat every category equally.

Identity and security software led the rebound, with some of the strongest names up close to 20% for the year. Vertical software, meaning tools designed around the specific rules of a single industry rather than a broad horizontal use case, performed in a similar range. On the other side, categories most exposed to AI substitution kept sliding. CRM tools and sales and marketing platforms were the weakest of the bunch, with some names down almost 45%. The distance between the best and worst performers in the same asset class stretched to roughly 65 points.

That spread matters more than any single headline number. It shows that SaaS stopped behaving like one uniform category somewhere in the last six months. It split into software that AI agents genuinely struggle to replace, and software that resembles exactly the kind of task an agent was built to automate. If you run a company anywhere in this space, figuring out which side of that split you fall on tells you far more about your near-term prospects than the overall market headlines do.

Following the Money: Where Buyers Are Actually Looking

Broad lists ranking the "top buyers in SaaS" rarely tell a founder anything useful. The better question is which categories are pulling real acquirer interest right now, and what those categories have in common.

L40°, an M&A advisory firm, maps this through a network of 172 private equity buyers spread across 5,780 portfolio companies in 28 mapped verticals. The pattern that shows up in that data is consistent. Buyers are chasing software that is mission critical, regulated, or tied directly to a transaction.

ERP and back-office platforms sit near the top of that list. These are systems of record, and ripping one out means rebuilding how a company operates internally, which is not a decision anyone makes lightly. DevTools and infrastructure software draws a similar level of interest, since it's software that other software relies on. That kind of embedded usage is nearly impossible for a competitor, AI-powered or otherwise, to quietly displace.

FinTech and payments platforms follow closely behind, largely because their revenue is tied to transaction volume and layered with regulatory complexity that keeps new entrants out. HealthTech and medical software benefits from that same dynamic, where compliance requirements act as a natural moat against fast-moving AI-native challengers that haven't cleared the same regulatory bar yet.

A step down from those, data and analytics tools built on proprietary information still command a real premium, especially when the underlying data can't be easily replicated by a generic model. Consumer and retail software has seen renewed interest too, mostly in cases where network effects or brand loyalty are doing the defensive work. Marketing software rounds out the list. Deals are still getting done here, but this category faces the most scrutiny of the group, precisely because it sits closest to the type of repetitive workflow an AI agent was designed to handle.

The thread connecting every category with strong buyer interest is the same. It's genuinely hard for a general-purpose AI agent to step in and take over. The categories under the most pressure are the ones where that isn't true.

Buyers Are Diligencing Differently Than They Did a Few Years Ago

Talk to anyone running acquisitions at a growth equity fund this year and a pattern comes up again and again. They're scoring targets on an AI disruption scale before they even look at growth rate or margin. One fund described ranking companies from 1 to 8 based on how likely they are to get replaced, whether by a competitor's internal tooling, a new AI-native entrant, or a more advanced feature added to an existing product.

That's a genuinely new step in the process. A few years ago, the opening questions in diligence were about churn, CAC payback, and net expansion. Those still matter, but they aren't the first filter anymore. The first filter now is whether the business could quietly become irrelevant after the next model release.

A handful of traits keep showing up in acquirer conversations as the things that ease this concern. Vertical depth is one, meaning workflows shaped around the specific, often unglamorous rules of a single industry that a horizontal competitor can't shortcut with a fine-tuned general model. Regulatory and infrastructure friction is another. Compliance obligations, real-time dependencies, or critical-infrastructure status make "just switch to an agent" a much tougher sell to a risk committee. Embedded relationships help as well, since software woven into a network of customers, vendors, and support teams tends to survive disruption better than a simple interface sitting on top of a database. And finally, having non-technical end users seems to reassure buyers. One mid-market acquirer explained it simply: when the person using the product isn't especially tech-savvy and deal sizes are moderate, the business feels sturdier because there's less appetite among that customer base to experiment with building an AI workaround themselves.

None of this replaces the fundamentals. A benchmark set of 101 software companies shows a median Rule of 40 score near 35% and net revenue retention around 109%, with top-quartile companies clearing 47% and 113%. Multiples across that group range from 2x to 10x revenue, and companies above $10M in ARR consistently price higher than those in the $5M to $10M range. Growth and margin are still the floor. AI defensibility is the new ceiling.

What Founders Should Actually Do With This Information

Private valuations tend to lag public sentiment by six to twelve months. That gap matters right now, because much of the worst public headline noise may already be priced out for the categories buyers actually want, even though plenty of founders are still anchoring their expectations to a narrative that private buyers stopped believing in months ago.

A few things worth acting on if you're weighing a sale in the next year or two.

Figure out which side of the AI exposure line your category falls on before setting price expectations. A strong CRM business and a strong ERP business with identical growth numbers are not competing in the same buyer pool this year, and treating them the same way sets up a disappointing process.

Build a real, specific answer to the AI question rather than treating it as a slide to skip past. Buyers will ask how your product holds up in an agentic world, and a bullet point about "AI-powered features" won't hold up under real diligence. A grounded explanation of your actual moat will.

Stop watching the public market for a signal to act. Private deals are negotiated, not traded on an exchange, and waiting for a green day on the Nasdaq before starting a process misreads how this side of the market actually works.

Run an actual process instead of hoping the right buyer finds you first. Targeted outreach, real competitive tension between a small group of qualified acquirers, and a clean data room routinely produce a better outcome than fielding one unsolicited offer. It's the kind of buyer curation firms like L40° build their sell-side work around, focusing conversations on the acquirers most likely to actually pay for a specific type of business rather than casting the widest possible net.

The Real Story of 2026

SaaS M&A didn't slow down this year. It split into two markets that happened to share a name. One is trading on fear about what AI agents might eventually do to software as a category. The other is trading on evidence about which specific businesses AI agents genuinely can't replace, and it has kept paying real premiums for those businesses the entire time.

The founders who come out ahead won't be the ones waiting for the headlines to sound better. They'll be the ones who already have an answer to the question every serious buyer is now asking first, which isn't how big is your market, but why does this business still need to exist when an agent can write code, book meetings, and close support tickets on its own. Get comfortable with that answer, and the rest of the process tends to sort itself out.

[themnabrief]

Your SaaS Company Won't Sell in Six Months — And That's Okay If You Plan Right

Key Takeaways

The full exit timeline is 12–24 months, not six to nine. The widely quoted figure only covers the live advisor-managed process.

Preparation is the clock you own. Once the live process starts, pace is set by buyers, advisors, and lawyers — not you. Preparation is different: it's entirely founder-controlled and usually the more impactful of the two phases.

The live process runs through four phases — Prepare & Position, Targeted Outreach, Drive Negotiations, Execute & Close — each with distinct time demands and founder involvement levels.

Disorganised information is the single biggest deal drag. Clean reporting, an organised data room, and a defensible metrics narrative consistently produce faster, better-valued outcomes.

Owner dependency is a valuation risk that can't be fixed during the process. It needs to be addressed years in advance.

Start the preparation conversation 12–16 months before you plan to go to market — not to rush the process, but to understand exactly what needs to be ready before it begins.

Let's start with a question most founders never ask until it's too late: when did you actually start preparing to sell?

Not when did you decide to sell. Not when you hired an advisor. When did you start preparing — cleaning up your finances, documenting the things only you know, building a business that doesn't fall apart the moment a buyer looks under the hood?

For a lot of SaaS founders, the honest answer is "not early enough." And that single gap — between when preparation should start and when it actually does — is responsible for more deal delays, lower valuations, and failed exits than almost anything else.

This isn't a process article dressed up as strategy. It's a straight look at what selling a software company actually takes, how long it really runs, and where founders consistently lose time they didn't know they were losing.

The Six-to-Nine Month Myth (Sort Of)

Six to nine months is real. A well-run, competitive sell-side process — managed by an experienced advisor, with qualified buyers, moving through proper stages — genuinely does take six to nine months from engagement to close.

The myth isn't the number. The myth is thinking that's when the work begins.

Before a process starts, there's a preparation phase that's entirely on the founder. Getting financials buyer-ready. Reducing how much the business depends on you personally. Organising a data room that doesn't make buyers nervous. Building a metrics story that holds up under scrutiny.

That phase typically takes six to eighteen months. Sometimes longer.

Add it all together and the realistic window from "I want to sell" to "deal closed" is twelve to twenty-four months. That's not a worst-case scenario. That's what a well-executed exit looks like when founders are honest about the work involved.

The founders who closed in six months didn't find a shortcut. They did the preparation work a year before anyone saw their CIM.

Why Founders Get This Wrong

There's a psychological pattern here worth naming.

Selling a company is a massive, emotionally loaded decision. Most founders sit with it privately for months — sometimes years — before they tell anyone. And during all that internal deliberation, preparation isn't happening. The clock isn't running.

Then one day the decision feels final. They reach out to an advisor, set a target close date mentally, and work backwards — and the math just doesn't work. They're already behind.

This isn't a failure of intelligence. It's a failure of framing. Preparation feels like an admin. It doesn't move a revenue metric or close a customer. It's invisible work that pays off later, in a process you haven't started yet, in a negotiation you can't fully imagine. That makes it very easy to defer.

The problem is that the preparation clock is the only one you fully control. Once the live process starts, pace is shaped by buyers, advisors, lenders, and lawyers. You're a participant, not the driver. Preparation is different — it's entirely yours to run, which makes it the highest-leverage time investment in the entire exit.

What a Live Process Actually Involves

Once preparation is done and a process launches, it runs through four phases. Here's what each one looks like on the ground.

Prepare and Position — 4 to 6 weeks

Your advisor takes everything you've built and turns it into a buyer-facing story. They pressure-test the financials, stress-test the metrics narrative, build the Confidential Information Memorandum, and map the buyer universe. If your information is clean and organised, this phase moves efficiently. If it isn't, it extends — and you feel the drag for the rest of the process.

Targeted Outreach — 4 to 8 weeks

NDAs go out. Buyers receive the teaser, then the CIM. Interest gets qualified. Indications of interest come in. The duration here varies more than any other phase, and a lot of it comes down to buyer type.

PE-backed acquirers who run platform strategies have done this dozens of times. They know exactly what they're looking for, they move fast, and they come prepared. First-time strategic buyers — a larger company acquiring software for the first time — build conviction slowly. They ask more questions, need more context, and take longer to commit. Neither is a problem, but knowing which type dominates your buyer pool shapes your timeline expectations.

Drive Negotiations — 4 to 6 weeks

Management presentations. Moving from indications of interest to Letters of Intent. Building the kind of competitive tension that protects your leverage on price and terms. Selecting a preferred buyer and entering exclusivity.

This is emotionally the hardest phase. Decisions are coming fast, the stakes are high, and there's rarely a clearly right answer. A structured, competitive process — rather than a single inbound offer you're reacting to — is what gives you options when it matters most.

Execute and Close — 8 to 12 weeks

Confirmatory diligence. Definitive agreement. Legal negotiation. Funds transferred. This is typically the longest phase, and the one most sensitive to how well-prepared your information was from the start. Customer concentration issues, complicated contracts, equity rollover provisions, earnout structures, or cross-border complexity can all extend the close significantly.

This is also the phase where you're doing all of this while still running the business. It's the stretch that separates founders who planned for it from those who didn't.

The Things That Slow a Deal Down

Two SaaS companies at similar revenue and growth can run processes that differ by six months or more. The gap almost always comes down to a short list of factors.

Messy or incomplete information. This is the biggest controllable drag, and it shows up everywhere — in Phase 1 when the advisor is building the CIM, in Phase 2 when buyers are forming their initial view, and in Phase 4 when diligence questions need answers. Buyers assess risk through information quality. Gaps don't just slow the process — they create doubt, and doubt gets priced in.

High owner dependency. If the business relies on you to function — you carry the key customer relationships, you're the only one who understands the product roadmap, decisions stall without you — buyers see a transition risk. They manage that risk with earnouts, valuation adjustments, or extended diligence. All of those slow the deal.

Complex deal structure. Earnouts, equity rollover, customer concentration, international entities — each one adds negotiation surface and legal review time. Some complexity is unavoidable. Some of it can be anticipated and simplified before the process starts.

Buyer mix and market timing. A process that launches in November runs into holiday schedules. A buyer pool dominated by first-time strategic acquirers takes longer to mature. These aren't disasters — they're variables. The founders who account for them plan around them.

The Preparation Work That Actually Moves the Needle

If the live process is largely outside your control, preparation is entirely within it. Here's where that time is best spent.

Financial reporting that buyers can actually use. Not just top-line revenue — a proper ARR waterfall showing new expansion, contraction, and churn. Gross margin by cohort. CAC and LTV trends over time. A clean bridge from bookings to recognised revenue. If you can't produce this today without significant effort, that's the gap to close first.

An organised data room. Customer contracts, IP assignments, employee agreements, vendor dependencies — all of it accessible, labelled, and current. A data room that's already structured when the process starts signals maturity. One assembled mid-diligence signals the opposite.

Reduced owner dependency. This is slow work, and it can't be rushed. Building out a management team that can operate independently. Documenting institutional knowledge that currently lives in your head. Strengthening customer relationships that exist at the company level, not just with you personally. It takes time measured in years, not months.

An honest read on your metrics story. Not the version you'd pitch to an investor — the version that holds up when a buyer's diligence team runs the numbers independently. Firms like L40°, which runs structured sell-side processes for technology companies, consistently find that the founders who arrive with a metrics narrative they can defend under pressure close faster and at better terms than those who can't.

When to Start — And What "Starting" Actually Means

The standard advice is to start preparing twelve to sixteen months before you plan to go to market. That's right, but it undersells the point.

The real question isn't when to start preparing. It's when to start talking to someone who can tell you what to prepare.

An M&A advisor who works in your space can look at your business today and tell you what it would look like to a buyer — what the strengths are, where the gaps are, and which of those gaps are worth closing versus accepting. That conversation is worth having well before you're ready to run a process. It reframes preparation from a vague to-do list into a specific plan with a specific payoff.

Most founders who do this find that the gaps are bigger than they expected, and that closing them takes longer than they'd like. Which is exactly why the conversation is worth having early.

What the Full Timeline Looks Like

Put it all together and here's what a realistic exit timeline looks like for a SaaS founder who starts on time:

12 to 18 months before going to market: Engage with an advisor. Get an honest read on the business. Identify and start closing preparation gaps — financial reporting, owner dependency, data room organisation.

6 to 12 months before going to market: Complete the preparation work. Metrics narrative is clean and defensible. The data room is organised. Key dependencies are reduced. Financials tell the story buyers need to see.

Go to market: Live process begins. Six to nine months: Prepare and Position → Targeted Outreach → Drive Negotiations → Execute and Close.

Total from first serious intent to closed deal: Twelve to twenty-four months.

That's the honest answer. Not the clean six-to-nine-month version that looks good in a slide. The full picture that reflects what a well-run exit actually requires.

[themnabrief]

The Ownership Math SaaS Founders Don't Run Until It's Too Late

Key Takeaways

Cumulative dilution compounds faster than any single round suggests — Carta's 2026 data shows median founder ownership falling from 56% after seed to roughly 27% by Series B for AI-native companies, even when each individual round looked reasonable.

Capital efficiency directly determines ownership retained — AI-native companies reaching $100M ARR in 1.5 years at lower burn multiples dilute less per milestone than traditional SaaS companies taking 7-plus years to get there.

Compressed valuation multiples raise the real cost of equity — Series B SaaS rounds now trade at roughly 5–7x ARR, down from 12–18x in 2021, meaning each percentage point of equity buys more of the company than it used to.

Option pool top-ups dilute founders beyond the headline round terms — and rarely get modeled into a founder's mental math until after the round closes.

Venture debt's record growth ($68.8B in 2025) is a rational response to pricier equity, not just a trend — it shifts some growth spending off the dilution curve entirely.

Later-stage protective provisions can matter more than ownership percentage at exit — liquidation preferences and participation rights determine who actually gets paid first, independent of the headline cap table.

Ask a founder mid-raise how much of the company they'll own after the round closes, and most can answer within a point or two. Ask the same founder how much they'll own by Series C, or at exit, and the answer gets vague fast. That gap — between knowing the next round's dilution and knowing the cumulative dilution three rounds out — is where a surprising amount of founder wealth quietly disappears.

It's not because founders are bad at math. It's because each financing decision gets evaluated in isolation, against the round in front of them, rather than as one link in a chain that compounds. A founder who carefully negotiates 18% dilution at Series A and another careful 15% at Series B has, without anyone doing anything wrong, already given up close to a third of the company in two transactions that each looked individually reasonable. Add a Series C, an option pool top-up, and a debt facility with warrants, and the cumulative math gets a lot less forgiving than any single round ever suggested.

This is the conversation worth having before the next term sheet arrives, not after: not "is this round's dilution fair," but "what does my ownership curve look like three rounds from now, and is there a cheaper way to get some of this capital."

What the Curve Actually Looks Like in 2026

Carta's 2026 founder ownership data gives a clearer picture of this curve than most founders have in their heads. The median founding team retains roughly 56% of the company after a seed round. By Series A, that median drops to 36%. By Series B, for AI-native companies specifically, the median founding team holds just 27.3% of fully diluted equity. Each step looks like a normal, market-standard round in isolation — 15 to 30% dilution at Series A, another 10 to 20% at Series B. Stacked together, they tell a different story: most founders are statistically likely to be minority owners of their own company well before any exit conversation begins.

There's real variation underneath that median, and it's not random. Founders in digital, capital-light businesses retain meaningfully more equity per round than founders in physical, asset-heavy ones — roughly 37.5% versus 30.5% at Series A on comparable terms. Capital efficiency compounds the same way: AI-native companies are now reaching $100 million ARR in roughly 1.5 years versus 7-plus years for traditional SaaS, and are doing it at burn multiples of 0.8 to 1.2x versus 1.2 to 1.5x for traditional SaaS peers. That efficiency gap shows up directly in ownership retained, because companies that need fewer total dollars to hit the same milestones dilute less to get there, even at identical per-round dilution percentages.

The practical implication isn't "raise less" as a blanket rule. It's that capital efficiency is now a direct, quantifiable input into how much of the company a founder ends up owning, not just a nice-to-have operating discipline.

Why Round Size Compression Is Changing the Calculus

2026 has added a second wrinkle to this math that didn't exist a few years ago: valuation multiples have compressed meaningfully at growth stages. Series B SaaS rounds are now trading at roughly 5 to 7x ARR, down from 12 to 18x in 2021. That compression cuts two ways for founders thinking about dilution.

On one hand, lower multiples mean a given dollar amount raised costs more in ownership percentage than it would have a few years ago — a $20 million Series B at a $100 million pre-money valuation today dilutes founders more than the same check would have against a richer 2021 valuation. On the other hand, it's pushing more founders to look harder at non-dilutive alternatives for the portion of capital that doesn't strictly need to be equity, precisely because each percentage point of equity now buys a lender or investor more of the company than it used to.

This is exactly the environment in which venture debt's growth makes sense as a rational response, not just a trend. U.S. venture debt closed 2025 at a record $68.8 billion, with SaaS accounting for more than $28 billion of that. None of it is free — venture debt typically runs an all-in cost of 8 to 15% annually plus warrant coverage that dilutes, just less and less permanently than a priced round does. But against a backdrop where every percentage point of equity costs more than it used to, the case for financing at least some growth spending through debt rather than another priced round has gotten stronger, not weaker.

The Round Nobody Models Properly: The Option Pool Top-Up

Here's a dilution mechanic that rarely gets the attention it deserves, because it isn't framed as part of the round — it's the option pool refresh that typically accompanies a priced round, and it dilutes founders before the new investor's money is even counted.

A Series A term sheet that requires expanding the option pool to 10% of the post-round fully diluted share count doesn't just dilute founders by the investor's stated ownership percentage. It dilutes them by that percentage plus the pool expansion, layered on top. A round that looks like 20% dilution on the headline term sheet can land founders several points lower once the pool refresh is included — which is exactly the gap between the simple math most founders run in their heads and the real number Carta's data captures.

The fix isn't avoiding option pools — they're necessary to hire the team that earns the next round. It's negotiating pool size deliberately, sized to actual hiring plans rather than a round number the investor proposes by default, and accounting for it explicitly when estimating post-round ownership rather than discovering the gap after the round closes.

What the Bootstrapped Counterexamples Actually Prove

It's worth being honest about the other end of the spectrum, because it's instructive rather than just inspirational. Zoho turned down a $10 million VC offer in 2000 and has since built a business serving hundreds of thousands of companies, fully founder-controlled three decades later. Mailchimp raised no venture capital across twenty years of operation and was acquired by Intuit for $12 billion in 2021, with founders capturing the full equity value because none of it had ever been diluted away.

These aren't arguments against raising capital — most SaaS companies competing in venture-backed categories don't have the luxury of bootstrapping at that scale, and trying to force it can cost more in lost market position than the dilution it avoids. What they actually prove is narrower and more useful: dilution is a tool a founder is choosing to use, not a tax that's simply owed at each stage of growth. The founders who end up with the most value at exit, across both the bootstrapped and the venture-backed examples, are the ones who treated every dilutive dollar as a deliberate decision rather than a default next step.

How This Plays Out at the Exit Line

The reason cumulative ownership math matters more than any single round's dilution is that it's the number that determines what a founder actually walks away with — and the path to get there changes that number in ways that aren't always visible from inside a single financing decision.

Debt taken on along the way is repaid first and senior in any sale, which makes its impact on exit proceeds straightforward to calculate: a fixed amount comes off the top regardless of how the remaining equity is structured. Equity dilution is where the real complexity compounds, because later-stage rounds increasingly carry protective provisions — participation rights, anti-dilution protection, liquidation preference stacking — that can change who gets paid first independent of straightforward ownership percentage. A founder holding 25% after four rounds of standard, clean 1x non-participating preferred is in a meaningfully different position at a $100 million exit than a founder holding the same 25% behind several rounds of stacked participating preferred.

This is precisely the kind of analysis the advisory team at L40° runs with founders well before a sale process begins: modeling the actual cumulative effect of every round and every debt facility on the cap table, against real exit scenarios, rather than evaluating each financing decision purely on its own terms. The round that looked perfectly reasonable in isolation can look very different once it's properly weighed against the three or four rounds that came after it.

Three Questions Worth Running Before the Next Round

What does my ownership look like two rounds from now, not just after this one? Model the next raise on top of this one using realistic stage-appropriate dilution, including the option pool top-up most founders forget to include. The number is almost always lower than it feels in the room.

Is this specific dollar better financed with debt than equity, given where multiples sit right now? With Series B multiples compressed to roughly 5–7x ARR, every percentage point of equity raised buys an investor more of the company than it would have a few years ago — which raises the bar for what equity dollars need to be financing.

What terms are riding along with this round besides the headline dilution? Liquidation preferences, participation rights, and pool expansions don't show up in the percentage everyone quotes, but they're often the terms that matter most by the time there's an actual exit on the table.

The Bottom Line

No single financing round is usually the mistake. The mistake is treating each one as a standalone decision instead of one link in a chain that compounds faster than it feels like it should. A founder who runs the cumulative math before each raise — not just the round in front of them, but the curve it's part of — ends up making meaningfully different decisions about how much to raise, when, and through which instrument. That's the gap between founders who are surprised by their ownership percentage at exit and the ones who saw it coming three rounds earlier and planned around it.

[themnabrief]

Choosing a Boutique M&A Advisor in 2026: A SaaS Founder's Field Guide to a Record Market

Key Takeaways

2025 set a SaaS M&A record with 2,698 deals, but buyers in 2026 are far more selective about quality.

Median private SaaS trades near 4.5x ARR; elite metrics (Rule of 40 above 50) can reach 6–8x.

AI positioning now appears in ~72% of deals; a weak AI story is a real liability.

Independent advisers have tripled their fee share since 2000 by winning the specialized mid-market.

Test advisors on positioning instinct and AI fluency, not just their buyer Rolodex.

Confirm which senior partner actually runs your deal day to day; get names.

Expect 5–7% success fees on $5–10M deals, declining to 1–2% above $100M.

Beware valuations anchored to the vanished 2021 peak; they signal a failed process ahead.

Begin advisor relationships a year early; preparation, not timing, drives the premium.

There's a strange paradox shaping SaaS exits in 2026. On paper, it has never been a better time to sell. SaaS M&A activity hit a record in 2025 with 2,698 transactions, up 28% year over year, and the first quarter of 2026 kept pace with 659 announced deals. Private equity dry powder sits at historic levels, and financing costs are easing as interest rates moderate. The buyers are active, the capital is there, and the headlines are loud.

And yet, ask any founder who's been through a process recently and you'll hear the other half of the story: buyers have become ruthlessly selective. The market isn't lifting all boats anymore. It's a sorting machine. The 2021 peak, when public SaaS companies traded at a median of 18.6x revenue, is long gone, replaced by a more disciplined market where the gap between premium and average businesses has widened sharply.

That gap is the whole game now. And it's why the advisor you choose matters more than it did even two years ago, not because a good banker can conjure a higher number, but because the difference between a "must-have" narrative and a "nice asset" narrative is now worth several turns of revenue. Let's talk about how to pick one in this specific market.

First, understand what "selective" actually means in 2026

Buyers aren't being cautious for caution's sake. They've rewritten what they're willing to pay for. The median private SaaS company in the lower middle market now trades around 4.5x ARR, but companies with growth above 30%, net revenue retention above 110%, and a Rule of 40 score above 50 can command 6x to 8x. That's not a rounding error. On a $40M ARR business, the spread between a 4.5x and a 7x outcome is $100 million.

Two forces are driving the sorting. The first is consolidation fatigue among customers. Sixty-eight percent of tech leaders plan to consolidate vendors in 2026, and the average enterprise now runs 106 SaaS applications, down from a peak of 130 in 2022. Buyers want software that's so embedded in a workflow it can't be ripped out the kind with real switching costs and a defensible niche.

The second force is AI, and it's reshaping deal logic faster than anything I've seen. Roughly 72% of SaaS M&A targets in 2025 referenced AI capabilities in their positioning. Acquirers aren't just buying revenue anymore; they're buying training data and defensible AI features. The flip side is that an unconvincing AI story now reads as a liability. A founder who can't articulate how their product survives and benefits from the AI wave will get marked down for it.

So when you evaluate an advisor in 2026, the real question isn't "can you run a process?" Most competent firms can run a process. The question is: can you frame my company so it lands on the right side of the sorting line?

Why boutiques have quietly won the mid-market

It's worth pausing on a structural shift, because it explains a lot. Independent advisers have nearly tripled their share of global M&A advisory fees since 2000, reaching about 37% of a roughly $27.6 billion pool. The listed boutiques are booming: Moelis posted record revenue of $1.52 billion in 2025, up 28%, and Houlihan Lokey grew quarterly revenue 18% year over year. A Chicago boutique even filed to go public, the first such IPO since 2021.

This isn't a coincidence. In a market that Barclays aptly calls a "conviction cycle" rather than a volume cycle, where buyers do fewer, more deliberate deals, sector knowledge and senior attention are worth more than institutional reach. A bulge-bracket bank optimizes for the $500M-plus transaction. For a $20–200M founder-led SaaS deal, that scale can work against you: your account gets handed to analysts, and your story gets compressed into a template.

A specialist boutique inverts that. The partner who pitched you is the partner who works the deal. Firms like L40°, which focuses specifically on cross-border SaaS and technology transactions in the mid-market, exist precisely because a founder selling a vertical SaaS company with a complex retention story needs an advisor who already speaks that language, not one who'll learn your metrics on your dime.

A 2026-specific framework for evaluating advisors

The standard advice: check sector expertise, buyer network, references still holds. But it's table stakes. Here's how I'd pressure-test a firm in this market:

Test the positioning instinct, not the Rolodex. Anyone can list buyers. Ask a prospective advisor to tell you, in the first meeting, what your company's single strongest "must-have" angle is and which three acquirers would care most about it. A good advisor will have a sharp, slightly uncomfortable opinion. A weak one will flatter you and recite a generic process. In a sorting market, the firm that can argue your narrative is the firm that earns the premium.

Probe the AI thesis directly. Ask: "How would you position our AI story to a strategic buyer, and where are we exposed?" If the answer is buzzwords, walk. With AI now embedded in three-quarters of deal rationales, you need an advisor who can distinguish a defensible capability from "AI-washing" because buyers' diligence teams certainly can.

Confirm who's actually in the room. This is the oldest piece of advice in M&A and still the most violated. Ask precisely who runs day-to-day buyer interactions. Get names. Senior involvement is the entire value proposition of a boutique; if it's not there, you're paying boutique fees for bulge-bracket neglect.

Look at recent, relevant deals not lifetime totals. A firm's $2 billion in "career transactions" tells you nothing about whether they've closed a vertical SaaS deal in your ARR band in the last 18 months. Cross-border activity hit a four-year high in 2025, so if your buyer universe is global, ask specifically about international processes.

Judge them on the questions they ask you. The best advisors interview you as hard as you interview them. If a firm is ready to sign a mandate without understanding your retention cohorts, your concentration risk, or your churn drivers, they're selling a process, not an outcome.

Demystifying the fees (with real numbers)

Fee structures intimidate founders, but they follow recognizable patterns. The Lehman formula a declining 5-4-3-2-1 scale remains the most common structure, used by roughly 41% of firms, with the percentage dropping as deal size rises. In the lower middle market, the Double Lehman (10-8-6-4-2) is often referenced, because smaller deals require substantial work relative to their value.

In practice, here's the rough shape: deals between $5M and $10M generally command success fees of 5% to 7%, $20M to $50M deals most commonly fall between 2% and 3.9%, and $100M-plus deals land around 1% to 2%. Monthly retainers typically run $5,000–$10,000, usually credited back against the final success fee. Watch for the rising "accelerator" structure, where the percentage increases above a valuation threshold a genuinely founder-friendly twist that aligns the advisor with stretching for the top of your range.

The number to scrutinize isn't the headline percentage. It's the fine print: the tail period (how long after the engagement ends the advisor still earns on a closing), the definition of "transaction value" when earnouts or rolled equity are involved, and the expense policy. A clean engagement letter is itself a signal of a clean firm.

The traps that are specific to this market

Old red flags opaque fees, stale buyer networks, conflicts of interest all still apply. But 2026 has its own.

The most seductive is the anchored valuation. Some advisors will quote you a multiple that quietly references the 2021 peak to win your mandate. Premium multiples in 2026 go to businesses with durable growth, strong cash flow, and defensible AI capabilities not to everyone. An advisor who promises a number wildly above the 4.5x median without a specific, defensible reason your company deserves it is setting up a failed process and a bruised reputation with buyers.

The second is AI theater: a firm that papers over a weak AI story rather than honestly addressing it. Diligence in 2026 is sharper than ever; a narrative that doesn't survive contact with a technical buyer will cost you credibility and momentum.

The third is manufactured urgency. Yes, the market is strong. Nearly 73% of advisors expect deal flow to increase in 2026. But "you have to sell now" is a sales tactic, not strategy. The strongest position is readiness without desperation.

Start before you think you need to

If there's one takeaway founders consistently learn too late, it's this: the work that earns the premium happens before the process starts. The metrics that make a company attractive to acquirers clean financials, strong retention, efficient growth are the same ones that make it durable as an independent business. Getting your house in order doesn't just improve your valuation; it gives you the optionality to act on inbound interest without scrambling.

So treat advisor selection as a relationship you build a year out, not a vendor you hire a month before. Talk to two or three firms. Watch who sharpens your thinking and who just flatters it. In a market that rewards conviction and punishes vagueness, the advisor who challenges you in the first meeting is usually the one worth hiring for the last one.

[themnabrief]

The AI Spending Paradox: Why the "Bubble" Debate Asks the Wrong Question and What Founders Should Do Instead

Key Takeaways

The AI bubble debate is mis-framed. Infrastructure, hyperscalers, platforms, and applications each carry different risk profiles. Treating AI as one market obscures where volatility actually lives.

The capital-revenue gap is real. Hyperscalers are spending ~$725B on AI infrastructure in 2026, while pure-play AI vendor revenues remain a small fraction of that commitment.

Application-layer startups face the sharpest existential risk. Thin-wrapper products with no proprietary data, low switching costs, and below-SaaS margins are the most vulnerable when the cycle turns.

Margins are improving, but the structural gap persists. AI application gross margins are projected at 52% for 2026, up from 41% in 2024, but still well below the SaaS standard of 75–85%.

Defensibility comes from outcomes, data, and integration, not novelty. Founders who identified a painful problem first, then applied AI, consistently outperform those who started with the technology.

The "token tax" is real and growing. Inference consumes ~23% of revenue at scaling-stage AI companies. Dependency on subsidised compute is a structural risk that only becomes visible when pricing normalises.

The Trillion-Dollar Gap Nobody Wants to Talk About

Here is a number that should stop you mid-scroll. The world's largest cloud and technology providers  Amazon, Alphabet, Meta, Microsoft, and Oracle are confirmed, based on Q1 2026 earnings calls, to be spending approximately $725 billion on AI infrastructure in 2026 alone, up from $443 billion in 2025 and $256 billion in 2024. Morgan Stanley has since revised that figure upward to approximately $805 billion when accounting for the most recent guidance updates. Goldman Sachs projects that combined hyperscaler spending from 2026 through 2031 will total $7.6 trillion.

Set against that, the pure-play AI vendors are posting extraordinary revenue growth but from a base that remains a fraction of the infrastructure commitment being made on their behalf. OpenAI is generating approximately $24–25 billion in annualised revenue as of mid-2026. Anthropic, in one of the most remarkable corporate growth arcs in technology history, crossed $30 billion in annualised run rate in April 2026 (up from $9 billion at the end of 2025), and has since reported figures approaching $47 billion by late May. OpenAI, for its part, has committed to approximately $1.4 trillion in data centre infrastructure over the next eight years.

That is not a rounding error. That is a structural gap between capital deployed and returns so far realised.

Now, does this mean it all collapses? Not necessarily. The revenue trajectory at the model layer is genuinely extraordinary. Anthropic's growth from $87 million in annualised revenue in January 2024 to $30 billion by April 2026 at a pace CEO Dario Amodei said outstripped the company's own internal forecasts by a factor of eight is without historical precedent at that scale. Salesforce took roughly 20 years to reach $30 billion in annual revenue. Anthropic did it in under three.

But the capital commitment is still running well ahead of aggregate returns at the application layer, where most founders actually live. That gap is where the bubble risk actually sits.

Stop Thinking About AI as One Market

One of the most expensive mistakes founders and investors make is treating AI as a single asset class. It is not. It is a stack of four distinct businesses with very different risk profiles, and your exposure depends entirely on which layer you occupy.

The infrastructure layer (chip designers, GPU manufacturers, power and cooling providers) is generating real cash today. Nvidia reported net income exceeding $120 billion for fiscal 2026 and trades at a forward P/E of approximately 24, elevated, but a far cry from Cisco's 200x at the dot-com peak. Demand from every layer below is still described by hyperscalers as supply-constrained, not demand-constrained. The risk, as ever, is overcapacity and the DeepSeek shock of January 2025, when Nvidia lost $589 billion in market capitalisation in a single day after a Chinese startup demonstrated competitive model performance reportedly trained for under $6 million, showed just how fragile that "infinite compute" narrative can be.

The hyperscaler layer (Amazon Web Services, Google Cloud, Azure, Oracle Cloud) is caught structurally in the middle: paying unprecedented amounts for infrastructure while racing to lock in enterprise customers who are still in evaluation and deployment mode. Capital intensity capex as a percentage of revenue has reached 45 to 57% at the largest providers, ratios more commonly associated with industrial utilities and telcos than technology businesses. Margins depend entirely on utilisation. Empty racks remain expensive racks.

The model and platform layer (OpenAI, Anthropic, Google DeepMind) is where the most complex economics live. These companies are subsidising access to capture market share, betting that adoption today translates to pricing power tomorrow. The positive signal is unmistakable: revenue at the two leading pure-play vendors has grown faster than almost any software company in history. The structural challenge is that both remain deeply loss-making. OpenAI is estimated to have generated an operating loss of approximately $7 billion in Q1 2026 on roughly a negative 122% operating margin, with infrastructure costs of approximately $25 billion and a roughly $6 billion annual revenue share owed to Microsoft. The direction of travel is right. The absolute numbers still need to catch up.

The application layer the startups, the vertical AI tools, the SaaS products being rebuilt around AI  is where most founders reading this actually live. And this is where the conversation gets most interesting, and most perilous.

The Application Layer: Where the Real Volatility Lives

ICONIQ Capital's January 2026 State of AI report, based on a survey of roughly 300 software executives, found that AI application gross margins improved from 41% in 2024 to 45% in 2025, with projections of 52% for 2026. That is a genuine and meaningful improvement. But traditional SaaS businesses still routinely operate at 75 to 85% gross margins. Even at the projected 52%, AI-native products run 23 to 33 percentage points below the SaaS norm a gap the best operators are working to close through model routing, prompt caching, and infrastructure migration, but a structural floor that is unlikely to fully converge.

The deeper problem at the application layer is not margins alone. It is dependency risk.

A significant portion of AI applications remain, at their core, thin wrappers around foundation models. When the underlying model improves, the wrapper can become obsolete. When a hyperscaler decides to bundle a feature natively, and they do this constantly, an entire category of startups can be disrupted by a single product announcement. The ICONIQ data also reveals that companies are now using an average of 3.1 model providers (up from 2.8 a year ago), with OpenAI at 77% penetration, Google Gemini at 55% (up sharply from 43%), and Anthropic at 51%. Multi-model strategies are becoming standard. Single-model dependency is increasingly seen as a vendor concentration risk that sophisticated investors are now actively pricing into valuations.

What this means practically: defensibility at the application layer comes from a very specific set of sources. Proprietary data that cannot be replicated. Deep workflow integration that makes switching painful. Domain expertise that a generalist model cannot easily replicate. Customer relationships built on demonstrable, measurable improvements to the customer's business rather than novelty or convenience.

The companies building durable businesses are almost always the ones that started by solving a specific, painful, measurable problem before they figured out how AI made it better. The AI was the enabler, not the product. That distinction sounds subtle. It is what separates companies that survive market corrections from those that don't.

What the Dot-Com Comparison Gets Right (and Gets Wrong)

The dot-com analogy gets wheeled out constantly in bubble conversations. In 2026, it is both more instructive and more actively debated than at any point in the last two years.

What the parallel gets right: The Shiller CAPE ratio stood at approximately 41 in mid-2026, the second-highest reading in 125 years of US stock market data, exceeded only by the dot-com peak of approximately 44 in late 1999. The top ten S&P 500 companies now represent roughly 30% of the entire index, the highest concentration in half a century. Nvidia's market capitalisation reached approximately $4.3 trillion by early 2026, and the five largest AI-adjacent companies collectively hold valuations that rival the GDP of major economies. As Michael Burry noted in May 2026, the top ten stocks have surged roughly 784% over the past year compared to 622% in the equivalent period before the dot-com crash. Markets are pricing in futures that have not yet arrived.

What the analogy misses: The companies driving the AI rally in 2026 are, unlike the companies driving the dot-com rally in 1999, among the most profitable in corporate history. Nvidia earned $120 billion in net income for fiscal 2026. Meta, Amazon, Alphabet, and Microsoft are generating substantial free cash flow. The infrastructure being built data centres, power systems, semiconductor fabrication is physically real in a way that Pets.com was not. And critically, the current wave of investment is largely self-funded by profitable technology giants rather than debt-driven startups raising money on the promise of future eyeballs.

The honest read: What is more likely unfolding is a prolonged correction at specific parts of the stack, with the application layer seeing the most visible casualties and the infrastructure layer absorbing a slow-burning overcapacity risk if efficiency gains at the model layer reduce compute demand faster than capacity can be rationalised. AI capex as a percentage of GDP currently sits at approximately 0.8%, still below the 1.5% peaks of comparable technology booms over the past 150 years. The cycle likely has runway remaining, but the runway is not infinite.

Five Questions Every AI Founder Should Answer Honestly Right Now

The macro debate will resolve itself. What you can control is the resilience of your own business. These five questions are not rhetorical.

1. What happens to your unit economics when AI inference pricing normalises? Inference currently consumes approximately 23% of revenue at scaling-stage AI companies, according to ICONIQ's January 2026 data. Significant portions of the AI ecosystem are operating on below-market inference costs as large providers subsidise access to capture market share. When that pricing normalises, and it will, does your margin structure hold? The companies that can answer this question with numbers, not narrative, are the ones worth backing.

2. How deep is your data moat, actually? Everyone says they have proprietary data. Most have a customised pipeline built on top of accessible datasets. Real proprietary data comes from deep, sustained customer integration over time the kind that accumulates through workflow embedding, not one-time ingestion. The ICONIQ data found that fewer than expected companies believe their data quality is actually ready for the agentic AI applications they are building toward. How long would it take a well-funded competitor to replicate yours?

3. Would your best customers pay 30% more tomorrow? Pricing power is the clearest test of genuine value. If the answer is "probably not," that is information about how defensible your product actually is, regardless of current net revenue retention. The companies commanding premium pricing in 2026 are overwhelmingly those selling measurable outcomes labor hours saved, error rates reduced, conversion rates improved- rather than capability access.

4. Is your retention story about outcomes or switching costs? Switching costs protect revenue in the short term. They do not generate growth, and they do not survive the cycle if a better product emerges. Outcome-driven retention, where customers stay because the product demonstrably improves their business, drives expansion revenue and survives market contractions. With 37% of companies planning to change their AI pricing model in the next twelve months, according to ICONIQ, the transition from novelty pricing to value pricing is accelerating. Know which side of that transition you are on.

5. What does your roadmap look like if your primary model provider changes API pricing by 2x? This is not a hypothetical. The companies using an average of 3.1 model providers did not arrive at that number by accident. They arrived there after experiencing the consequences of single-provider dependency. The shift to multi-model strategies is now visible in ICONIQ's data. Founders who have not yet diversified their model stack should treat this quarter as the last comfortable moment to do so.

The Founder's Actual Takeaway

Markets overshoot in both directions. The AI boom has generated genuinely inflated expectations in some parts of the stack and genuinely transformative technology in others. Both things are true simultaneously, and the evidence for both has only become more compelling through the first half of 2026.

For founders, the practical implication is not to try to predict which part of the story wins. It is to build a company that survives the version of the future where expectations don't fully materialise, while still being positioned to scale in the version where they do.

That means margins matter, not because growth doesn't, but because margin is what gives you time when the cycle turns. It means defensibility matters not as a pitch deck slide but as a genuine answer to the question of why customers would pay, stay, and expand. And it means building for outcomes rather than features, because features get commoditised at a pace in 2026 that would have seemed impossible two years ago, and outcomes generate the kind of loyalty that survives both market corrections and competitive announcements.

The AI investment cycle will eventually equilibrate between capital deployment and economic return. Goldman Sachs projects $7.6 trillion in cumulative AI capex between 2026 and 2031. The revenue base required to justify that will either arrive or it won't. The companies that endure won't be the ones that timed the cycle correctly. They'll be the ones that built something worth enduring.

[themnabrief]

Earnout Negotiation for Founders: Why 79% of Deferred Dollars Never Get Paid

Key Takeaways

The base rate is the headline. Earnouts have historically paid only ~21 cents on the dollar across all deals, with ~28% contested (SRS Acquiom 2025). Negotiate the expected value, not the face value.

Metric choice = control. Rank by who can move the number after close: revenue/ARR (best for sellers, 62% of 2024 earnouts) > gross profit > EBITDA (22%) > net income. For software, defend ARR or total revenue.

Write a blocker for each failure mode: cost dumping → revenue metric/carve-outs; resource starvation → covenant floors; strategic redirection → non-frustration clause; the flip → acceleration on change of control.

Shape over duration. Keep periods to 12–24 months, use quarterly measurement with partial payouts, and structure tiered (proportional) rather than all-or-nothing payouts.

Mind the tax trap. An earnout tied to your continued employment can be taxed as ordinary income, not capital gain. Decouple payment from employment and involve tax counsel early.

Walk away when there are no covenants or no operational role for you to take cash or a seller note instead.

Timing is leverage. Settle the earnout framework at the LOI stage, inside a competitive process, before exclusivity erases your alternatives.

When a buyer slides an earnout across the table, it almost always arrives dressed as good news. "We love the business. We want you to share in the upside." The number looks like part of your price. Mentally, you bank it.

Here is the figure that should change how you read that sentence: across all M&A deals tracked in the SRS Acquiom 2025 Deal Terms Study, earnouts have historically paid out only about 21 cents on every promised dollar. Roughly four out of five deferred dollars never reach the seller. And about 28% of earnouts end up contested, meaning the disagreement you papered over at signing tends to resurface, magnified, while you and the buyer are working under the same roof.

That is not an argument against earnouts. They remain one of the few tools that can rescue a deal stuck on price, and in 2024 they appeared in roughly 22% of private, non-life-sciences transactions. The point is narrower and more useful: the face value of an earnout and its expected value are two very different numbers, and most founders negotiate the first while collecting the second.

This is a playbook for closing that gap. Not the legal boilerplate version the version that treats an earnout as what it actually is: a bet on someone else's behavior after they have your company.

What you're really agreeing to

An earnout is deferred purchase price. A slice of the total, commonly tied to revenue, ARR, or a growth milestone reached within 12 to 24 months, gets paid only if the business performs after close. In 2024, the median earnout outside life sciences ran around 31% of closing payments. So in a deal that uses one, nearly a third of what you expect to walk away with is riding on what happens after you've already handed over the keys.

The mistake is treating that third as money you've earned and merely have to wait for. You haven't earned it. You've earned the right to try to earn it, under operating conditions a new owner now controls. The whole negotiation is about narrowing the distance between "we both believe the targets are reachable" and "the contract makes them collectable even when the buyer's own decisions get in the way."

So before debating dollar amounts, run the earnout through one question.

The only question that matters: who holds the lever?

Every earnout is measured against a metric. The metric's real risk isn't whether it's "fair"; it's who can move it after close. Rank the common choices by how much control sits with the buyer, and the seller's preference order writes itself:

Revenue / ARR moved mostly by customers, not by the buyer's accounting. Hardest to manipulate, which is exactly why sellers want it. In 2024, 62% of earnouts used revenue as the primary metric.

Gross profit sits one layer down, exposed to cost-of-goods allocation decisions the buyer makes.

EBITDA is heavily buyer-controlled. Headcount, marketing spend, shared-service charges, and integration costs all flow through it. A buyer can leave your top line untouched and still shrink the number you're paid on. Only about 22% of 2024 earnouts used it, and there's a reason that figure is low on the sell side.

Net income the most exposed of all. Avoid where you can.

The pattern: the further down the income statement your metric lives, the more of it the buyer controls. For most software exits, ARR retention or total revenue are the cleanest lines to defend, because they map to how the buyer underwrote you in diligence and resist post-close engineering.

If a buyer insists on EBITDA, you haven't lost; you've changed the conversation. Now the deal needs guardrails: a minimum operating-budget commitment, a ban on dumping parent-company costs onto your unit, standalone financials prepared on agreed accounting principles, and the right to inspect them. An EBITDA earnout without those is not a price. It's an option the buyer holds to pay you less.

The four ways a clean-looking earnout quietly fails

Most earnouts don't blow up over fraud. They erode through ordinary post-close decisions that happen to flow against your number. Knowing the four common failure modes lets you write the blocker for each before you sign.

1. Cost dumping. The buyer routes shared services finance, HR, IT, a slice of group marketing into your entity, and your EBITDA quietly deflates. Blocker: a revenue-based metric, or explicit carve-outs and standalone financials that exclude allocated parent costs.

2. Resource starvation. Nobody breaks a promise; they just stop investing. Marketing spend drifts down, two open roles go unfilled, and growth stalls on schedule. Blocker: operating covenants with hard floors marketing as a percentage of revenue, headcount bands by function, not vague "commercially reasonable efforts" language, which Delaware courts have repeatedly singled out as the root of earnout disputes.

3. Strategic redirection. The buyer repositions the product, changes pricing, or reroutes your customers into a different SKU. Revenue moves to a line you're not measured on. Blocker: a non-frustration covenant plus carve-outs that protect the measured metric from buyer-initiated strategy changes.

4. The flip. The buyer sells your unit mid-period, and the new owner has no agreement with you and no reason to honor one. Blocker: an acceleration clause that makes the full remaining earnout payable on any change of control. This one is reasonable for any good-faith buyer and should be close to non-negotiable on your side.

Notice that three of the four are defeated by metric choice and covenants, which is why those two decisions deserve more of your energy than the headline number.

Structure beats duration

Standard earnout periods run 12 to 36 months; for mid-market software, the seller-friendly zone is 12 to 24, and the market has been drifting shorter in 2024; no deals in the SRS Acquiom sample stretched past four years. The logic is simple: time is the buyer's ally, not yours. Every additional quarter adds another integration decision, another leadership change, another market wobble you don't control.

When a buyer pushes for a longer runway, negotiate the shape, not just the length. A single make-or-break measurement at month 36 is the worst possible structure; one bad quarter near the end can erase everything. Quarterly measurement with partial payouts converts that cliff into a staircase: you collect incrementally, disputes get resolved while memories are fresh, and neither side is staring down an all-or-nothing calculation years out.

The same principle applies to the payout curve. An all-or-nothing target is a trap; hit 95%, and you get zero, which is also the structure most likely to end in litigation. A tiered earnout that pays proportionally 85% of target earns 85% of the earnout- removes the incentive to fight over a narrow miss and keeps both parties pointed at the same outcome. And if your growth case is genuinely strong, it's worth asking whether an uncapped structure is realistic, letting you share in performance above the target. That conversation only works when both sides are already optimistic; raise it on a deal where the buyer is nervous, and you'll create friction instead of alignment.

The tax trap most founders never see coming.

Here's a detail that rarely makes it into earnout articles and routinely surprises founders: if your earnout is contingent on you staying employed, you may have just converted capital-gains treatment into ordinary compensation income. Tax authorities have long taken the position that a payment conditioned on continued service looks like wages, not purchase price, and the rate difference can be brutal.

The fix is to decouple the earnout from your employment wherever possible: the payment should follow business performance, not your presence on the payroll. There are also installment-sale interest considerations on larger deferred amounts. None of this is something to improvise in the room loop in deal counsel and a tax advisor early, because the structure is far easier to get right at the term-sheet stage than to unwind later.

When the right answer is "no earnout at all."

Not every earnout can be negotiated into something you'd want. A few signals mean you should push for cash at close even at a lower headline instead:

The buyer won't commit to operating covenants. If they insist on full discretion with no floors on investment or headcount, the earnout isn't a bridge to value. It's the option to underpay you, and you've just described the 79% that goes uncollected.

You're leaving at close. With no operational role, you can't influence the targets. Take cash, or a seller note that carries a repayment obligation regardless of performance.

Size alone, notably, is not a red flag. A large earnout that lifts a deal to an above-market multiple can be a great trade if the targets are reachable under the conditions on offer and the protections you've negotiated. The question is never "how big," it's "how collectable."

The leverage you have before exclusivity and lose after

The single most overlooked move in earnout negotiation is timing. Once an LOI is signed and exclusivity begins, your leverage collapses: you're negotiating against one buyer, on the clock, with no alternative. The metric, the period, the covenants, the dispute mechanism all of it is far easier to shape at the LOI stage, before the buyer knows they're the only one at the table.

This is also where a competitive process earns its keep. When several credible buyers are interested, no single one has the leverage to load the deal with deferred risk. Boutique sell-side advisors like L40° tend to push earnout structure into the LOI conversation for exactly this reason: by the time the purchase agreement is drafted, the framework is mostly set, and the room to improve it has narrowed to drafting nuance. The earnout you can actually collect is usually decided weeks before anyone starts arguing over clause language.

The reframe worth keeping

Stop reading an earnout as the back half of your price. Read it as a probability-weighted bet, and a $10M earnout with a market base rate behind it is worth far less than $10M until your terms move the odds. Everything in this playbook exists to move them: pick a metric the buyer can't quietly move, write covenants that bind, accelerate on a flip, shorten and stagger the clock, decouple from employment, and settle the structure before you give up your alternatives. Do that, and you stop hoping the earnout pays. You make it collectable.

[themnabrief]

Two Founders, Identical ARR, a 7x Gap: What Actually Decides Your SaaS Exit in 2026

TL;DR: In 2026, two mid-market SaaS companies with the same revenue can sell for valuations that differ by 5x to 7x. The cause isn't market timing, luck, or a better banker's Rolodex. It's a small set of things buyers now price ruthlessly: retention quality, AI defensibility, founder independence, and the discipline of the process you run. This is a guide to closing that gap, written for founders with roughly $5M to $100M in ARR who are thinking about selling in the next 6 to 24 months. Key Takeaways

The spread is the story. Two mid-market SaaS companies with identical ARR can sell 5x–7x apart in 2026. The gap is earned through fundamentals, not assigned by the market.

Selective, not scarce. Public multiples corrected, but private buyers anchor on fundamentals, and roughly $4.3 trillion of PE dry powder plus easing rates keeps quality assets competitively bid.

The premium moved. Retention durability, real AI defensibility, vertical depth, and founder-independent operations are what now command 10x–15x. OneStream's ~8x forward ARR (March 2026) shows quality still gets paid.

AI is now a deal-risk variable. In a K-shaped market, AI-exposed targets can draw no bid at all. Have a one-sentence answer to what a foundation model does to your moat.

Curate the buyer room. 30–100 qualified, well-fit buyers beat a 200-name auction above $10M ARR. Match your narrative to whether your best buyer is a strategic, a PE platform, or a rollup.

Structure is where value leaks. Earnout thresholds, working-capital pegs, SaaS-specific reps, and common-vs-preferred rollover quietly erode outcomes. Negotiate them as hard as price.

Prepare, don't time. You can't call the cycle. You can fix two of five readiness gaps in nine months, and that's worth more than waiting.

Picture two SaaS companies. Both do $30M in ARR. Both grow about 25% a year. On a spreadsheet they look like twins. One sells for roughly 5x ARR. The other clears 11x. Same revenue, a $180 million difference in outcome.

That gap is the defining feature of the 2026 mid-market. The correction that started in 2022 is over, but it didn't restore the old playbook; it replaced it. Multiple expansion, the rising tide that carried mediocre companies to good exits between 2020 and 2024, is gone. What's left is a market that pays a genuine premium for a short list of provable qualities and applies a brutal discount to everything else.

The good news, if you're preparing well, is that the gap is mostly earned, not assigned. Almost every variable that separates the 5x founder from the 11x founder is something you can influence in the year before you go to market. This piece is about which variables actually move the needle, and how to use the months before a process to land on the right side of the spread.

First, kill the myth that this is a bad market to sell into

It's tempting to read the headlines and conclude you should wait. Don't confuse public-market drama with private-deal reality.

When the so-called "SaaSpocalypse" hit in early 2026, public software multiples compressed hard and tens of billions in market cap evaporated in days. But private mid-market M&A barely flinched, because private buyers anchor on fundamentals: ARR durability, net revenue retention, profitability rather than the daily sentiment that whips public tickers around. Global M&A deal value reached roughly $4.9 trillion in 2025, the second-highest on record per Bain, and software stayed firmly in the mix, capped by Google's $32 billion acquisition of Wiz.

The capital backdrop reinforces the point. Private equity is sitting on something in the order of $4.3 trillion of dry powder globally, and 2025 saw the second-highest private-equity take-private volume in a decade, according to Morgan Stanley. Sponsors face real pressure from their own investors to deploy that capital and to exit aging portfolios. Add a rate environment that's drifting lower, and you have buyers who need to transact.

So the market isn't scarce. It's selective. That distinction is the entire game.

The 60-second self-assessment before you do anything else

Before you talk to an advisor, polish a deck, or answer that flattering inbound email, run yourself through five questions. Answer them honestly the way a buyer's analyst would, not the way your board deck does.

Is your net revenue retention above 110%? This is the closest thing to a single number that buyers trust. Above 120% and you're in premium territory; below 95% and buyers start pricing in decline.

Is your Rule of 40 above 30 and does it hold up cohort by cohort? Aggregate scores no longer pass diligence. Buyers want to see whether the healthy blended number is hiding a weak recent cohort.

Can you describe, in one sentence, what AI does to your moat? Not "we've added AI features." The real question is whether a foundation model could replicate your core value. If you can't answer it cleanly, neither can a buyer's investment committee.

Is any single customer under 15% of revenue? Concentration is the silent deal-killer. A $5M ARR company with 40% of revenue in one logo can trade at half the multiple of an identical one with no client above 10%.

Can the business run for 90 days without you? If you are the architecture, buyers apply a key-person discount that quietly removes 20% to 40% of value.

Three or more honest "no" answers mean you have a preparation project, not a process. That's not bad news; it's the most valuable thing you can learn a year out, because most of these are fixable in 9 to 12 months if you start now.

What buyers are actually paying for in 2026 (it changed)

The premium has migrated. Here's where it sits today.

Durability over raw growth. A company growing 25% with 120% NRR now beats a company growing 40% with 95% NRR in most buyers' models. Expansion revenue from a sticky base is worth more than new-logo growth that leaks out the back door.

AI as defensibility, not narrative. This is the sharpest change of the cycle. PwC has described the current environment as K-shaped: companies riding an AI tailwind command outsized multiples, while genuinely AI-exposed targets can find no bid at all. Strategic acquirers are now walking away from otherwise attractive businesses when the diligence team concludes a model could eat the product. AI-native positioning, where it's real and measurable, has been associated with valuation premiums in the 40% to 80% range. The premium attaches to defensibility, not to the word "AI" in your pitch.

Vertical depth over horizontal reach. Software that has become the system of record for a specific industry, especially a regulated one, keeps outperforming generalist tools on multiple. Depth is harder to replicate than breadth.

Founder-independent operations. The further the company runs without you, the lower the perceived transition risk, and the higher the multiple.

Map these against the real distribution of outcomes and the spread is stark:

For a concrete anchor on what quality still commands: Hg and General Atlantic's OneStream deal closed at roughly 8x forward ARR in March 2026, clean proof that PE buyers will still pay up for a defensible asset even in a "down" market. Where a founder lands inside this table, far more than when they choose to sell, decides the outcome. This is precisely the distribution L40° sees play out across its mid-market sell-side mandates, and it's why the firm's first conversation with a founder is almost always about positioning rather than price.

Build the right room, not the biggest one

A common and expensive instinct is to maximize the number of buyers. In the mid-market, signal quality beats coverage. A curated universe of 30 to 100 genuinely qualified buyers consistently outperforms a 200-name blast, because fit, sector fluency, and relationship depth move price more than raw bidder count once you're north of $10M ARR.

Three buyer types are active, and each is buying something different:

Strategic acquirers pay for capability gaps, AI defensibility, a customer base, or simply to take a competitor off the board. They close fast (often 45 to 90 days post-LOI) and pay mostly in cash, but they integrate hard and may restructure your team.

PE platforms pay for ARR durability and Rule of 40, folding you into a take-private vehicle. Expect cash plus 20–40% rollover equity and a 60–120 day close, with management usually retained but pushed on margin within 12 to 18 months.

Sponsor-backed rollups pay for vertical fit and a repeatable playbook, with founder retention central to the thesis, typically cash plus rollover plus an earnout, and a 24-month-plus commitment from you.

Knowing which of these you're built for before outreach lets you tune the narrative to the buyer who will pay the most for what you actually have. It's also where advisor choice matters most: boutique mid-market specialists tend to carry deeper, more current relationships with the right 30 to 100 names in a given vertical than a generalist bank running your mandate as junior-staffed overflow.

Where deals leak value quietly: structure

Headline price is theater; structure is where the real economics live, and it's where mid-market founders most often lose money they thought they'd won. With PE buyers pushing more rollover into deals to hedge bifurcation risk, four leaks recur:

Earnouts with thresholds that look fair at signing but become unreachable once the buyer controls the levers: sales, hiring, marketing budget, the roadmap. Most founders badly underprice the option value a buyer extracts here.

Working-capital pegs set against a stale baseline, quietly shaving proceeds at close.

SaaS-specific reps and warranties, particularly how "recurring revenue" and contract terms are defined, that can claw back 5% to 10% of headline value through indemnity claims.

Rollover structured as common rather than preferred stock, leaving you exposed to dilution in later rounds with no participation rights.

A point of headline price negotiated well and then surrendered through a loose earnout definition is the most common own-goal in mid-market SaaS. This is the part of the deal where experienced representation usually pays for itself several times over.

What to do now, based on where you sit

12+ months out: Run the five-question self-assessment with zero flattery, then fix the two weakest answers. Most founders can flip two of the five from "no" to "yes" inside nine months, and each flip can be worth a full turn or two of multiple.

6 to 12 months out: Start advisor conversations, not to hire on the spot, but to pressure-test your timing, positioning, and buyer universe while you still have room to adjust. An early read on where you'd land in that tier table is worth more than any valuation calculator.

Under 6 months, or already holding inbound interest: Engage a sell-side advisor before you respond to any LOI. Warm inbound almost always understates the competitive process; you could otherwise run the single offer in your inbox is rarely the best one the market would produce.

The throughline across all three: in a selective market, preparation compounds and timing doesn't. You can't reliably call the cycle, but you can absolutely decide which side of the 7x gap your company sits on. That work starts long before the first buyer call.

Frequently Asked Questions

What counts as mid-market SaaS M&A? The sale of SaaS companies with roughly $5M to $100M in ARR, generally translating to $20M–$500M in enterprise value. It sits between micro-SaaS marketplace deals and large-cap software transactions, and it's the segment where boutique advisors and private equity buyers run curated, relationship-driven processes.

What multiples do mid-market SaaS companies sell for in 2026? Median exits cluster around 4x–6x ARR. Top-quartile companies with strong retention, real AI defensibility, vertical depth, and founder-independent operations reach 10x–15x, while bottom-quartile assets land at 1.5x–3x. The gap between top and bottom has widened to roughly its widest in a decade.

Is 2026 a good time to sell a SaaS company? For a well-prepared, quality business, yes. Despite public-market volatility, private mid-market M&A stayed active, supported by record PE dry powder and easing financing costs. The market rewards readiness over timing; quality assets are bid competitively, weak ones are not.

How long does a mid-market sell-side process take? Typically 6 to 9 months from advisor engagement to close, plus 6 to 12 months of preparation beforehand. Skipping that prep window is where founders most often surrender a turn or more of multiple to diligence re-trades.

Do I need a sell-side advisor? Above ~$5M ARR, the competitive tension, diligence management, and structure negotiation an advisor brings typically more than covers the fee. Founders who go direct commonly leave 15%–25% of value on the table and shoulder the diligence load themselves, and the gap matters more in the mid-market than at the top end.