I'm coming to COLORADO! Catch me in DENVER on Jan 22 at The Tattered Cover<, and in COLORADO SPRINGS from Jan 23–25 where I'm the Guest of Honor at COSine. Then I'll be in OTTAWA on Jan 28 at Perfect Books and in TORONTO with Tim Wu on Jan 30.
If Andrew "bunnie" Huang didn't actually exist, I'd swear he was a character out of a(n extraordinarily technologically well-informed) cyberpunk novel. Every time I interact with this legendary hardware hacker, he blows my mind with some incredible project or insight that permanently alters how I think about technology.
I first met bunnie when he came to EFF for help with the threats he'd received from Microsoft. At the time, bunnie was an electrical engineering grad student at MIT, and he'd taken the bootloader locks on the new Xbox platform as a personal affront and challenge. He applied his prodigious skill and talent to these digital handcuffs, and in short order, he had broken the Xbox and installed Linux on it. MIT's general counsel immediately washed its hands of any responsibility to defend this young grad student from bullying by a corporate monopolist, hanging him out to dry. So he turned to us – and we got his back. You can read all about it in Hacking the Xbox, his canonical work about hardware hacking and technological freedom (it's free!):
In the many years since, I've been lucky enough to count bunnie as a friend, colleague and comrade, albeit one I only physically run into every year or so, usually at some tech event or on the playa at Burning Man, where he still camps with the MIT crew at The Institute.
I just got to see bunnie in person again, over Christmas week at the Chaos Communications Congress in Hamburg. He gave a late-night presentation with his collaborator Sean "xobs" Cross, entitled "Xous: A Pure-Rust Rethink of the Embedded Operating System":
https://www.youtube.com/watch?v=BbWWGkyIBGM
Don't let the technical-sounding title intimidate you! This was a banger of a talk, and as with every bunnie Huang production, it left a pleasant and persistent aftertaste.
The background for this talk is bunnie's obsession with building a trustworthy computer. For decades, bunnie has been chasing the dream of a computer whose every component – operating system, drivers, firmware, and hardware designs – are open to inspection. Bunnie's reasoning here is that anything that can't be inspected (and, by extension, modified) by its users is a spot where bad guys can hide bad stuff, and where lurking bugs can fester until they are exploited by bad guys. Remember the spectacular (and still mysterious) claims that Apple's servers had all been compromised with minuscule hardware bugs? The single best explanation of that you will find comes from bunnie:
https://www.youtube.com/watch?v=RqQhWitJ1As
Bunnie was doing all this before there was an "open source hardware" movement, and he remains at its vanguard. His "Precursor" project is a reference hardware platform where every component is open to inspection and modification, from the chassis to the random number generator:
One area of especial concern and interest for bunnie is the promise and peril of the "system-on-a-chip" (SoC). This is exactly what it sounds like: a cheap chip that incorporates everything you need to do full-fledged computing, including interfaces and drivers for networks, screens, peripherals, etc. SoCs are ubiquitous. You find them in things like individual car engine components and inkjet printer cartridges, and each one is a whole-ass computer, capable of running some really ugly malware.
As bunnie explained back in 2020, there are two problems with SoCs: first, they are packaged such that the silicon traces inside of them can't be readily inspected, and second, they are so expensive to fabricate that someone like bunnie can't possibly come up with the millions needed to make an open, trustworthy, inspectable alternative:
That's where bunnie's CCC talk comes in. The chips that SoCs are etched upon have lots of space (relatively speaking – we're talking about nanometer-scale circuits, after all). Even after an SoC designer packs in a ton of extra traces to handle oddball applications, the chip is still mostly "dark matter" – blank silicon.
The first half of bunnie and xobs's talk concerns itself with "Xous," a secure operating system for an SoC, written in Rust. But the second half of the talk tackles the problem of procuring an SoC that you can trust to run Xous on. That's where this dark matter comes in.
Bunnie's day-job is consulting on extremely gnarly, high-stakes, high-value hardware design and manufacturing, so naturally, he's got lots of clients and contacts in the SoC manufacturing world. He approached one of these companies with a proposal: let me tape out a whole separate chip that fits in the dark matter for one of your upcoming chips. Adding these traces adds virtually no cost to the production, and adding bunnie's chips to the production run actually saves the manufacturer money, because the prices drop when the quantities increase.
The idea is to put two chips on the chip, and badge most of them with the OEM's branding, while a small rump of the chips will have bunnie's branding (he calls it the Baochip). On bunnie's chips, the traces to the OEM chip will be physically cut, meaning that the Baochips will just be Baochips – the original chip will be inaccessible and unusable.
What's more, bunnie didn't just fit one chip into the OEM's "dark matter" – he fit five separate, specialized SoCs into the unused space. Remember, the beauty of SoCs is that once they're taped out and sent to production, the cost of an actual chip is peanuts, meaning that these Baochips are cheap as hell.
Even better: the traces on these chips are scaled to be readily inspected using relatively low-cost equipment, meaning that many parties around the world can grab one of these chips, stick it in a machine, and compare the traces on the chip to the free, open sourcefile that was used to produce it, confirming that there are no nasty surprises lurking inside.
This was such an exciting talk, and as I sat through it, I had this nagging feeling that it reminded me of something else I'd learned about years before, though I couldn't quite place it. Finally, as bunnie and xobs were stepping off the stage, I had it – it reminded me of another bunnie talk I'd seen – this one at The Institute, the MIT Burning Man camp, more than a decade prior.
Back in 2015, bunnie designed and built a set of really cool, wearable radio-linked badges for his campmates, which would help them locate one another on the playa at night. These badges were really cool – they used a genetic algorithm to "have sex" with one another and mutate their color patterns. Bunnie even worked in a "consent" mechanism!
But the really cool part that stuck with me was the manufacturing story. Bunnie wanted to fabricate custom injection-molded plastic enclosures for these pendants, but injection molding – like chip design – is a mass production phenomenon, with sky-high setup costs and incredibly cheap per-unit costs thereafter.
So (and this might sound familiar) bunnie reached out to a die-maker that he worked with in China and said, "Hey, the next time you're contracted to mill out a die for a client, let me know if there's any extra space on the face of the die, and I'll provide you with a shapefile you can carve out of this 'dark matter.'" This doesn't add any cost to the die setup, and it means that bunnie can run just a couple dozen injection-molded, custom cases at a cost of pennies per unit.
I grabbed bunnie later that night and mentioned this old Burning Man project to him and he said, "You know, I haven't ever thought of it, but you're right, there's definitely a throughline between the two projects."
I asked him what he called this technique and he shrugged and said he didn't really have a name for it, but he thought of it as "piggybacking," which seems like a good name to me.
It seems to me that these two kinds of manufacturing can't be the only ones that can be "piggybacked" onto. That's what motivated me to write this post – to get people thinking about these high-setup/low-unit cost production types that might be piggybacked for small batch, delightful projects like bunnie's.
Well, that, and just to do one of my periodic bunnie Huang appreciation posts. If there's one person that I'd recommend people pay more attention to, it's him. He's also a terrific communicator, and an indecently great writer. My readers might be familiar with him thanks to the afterword he contributed to Little Brother:
https://craphound.com/littlebrother/download/
More recently, he wrote a fantastic intro for last year's Science Comics Computers: How Digital Computers Work, a brilliant middle-grades graphic novel that uses steampunk dinosaurs to explain digital logic and the building blocks of computation:
He also co-authored a fascinating research paper with Edward Snowden, after the two of them collaborated on a daughter-board that spots otherwise untraceable malware:
That's not bunnie's only sweet hardware hack, of course. Check out the insanely clever design for a contact-tracing dongle he prototyped for the EU in 2020:
But really, you owe it to yourself to read bunnie at book length, and his best book is 2016's The Hardware Hacker, a tour-de-force, lay-friendly exegesis on the theory and practice of hardware hacking:
If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
What Future Trends Will Shape the Microfluidics Market by 2033?
The Microfluidics Market is set for transformative growth, rising from USD 39.65 Billion in 2025 to USD 100.30 Billion by 2033 at a 12.3% CAGR. Future trends point toward integration with digital health, AI analytics, and wearable biosensors.
Insights from the global microfluidics market show increasing adoption in environmental testing, food safety monitoring, and forensic applications beyond healthcare.
The emergence of organ-on-chip systems, personalized diagnostics, and real-time monitoring tools will redefine medical research and treatment approaches. Smart microfluidic systems with embedded electronics will enable automated data collection and analysis.
Sustainable materials, reusable chips, and scalable manufacturing will further reduce costs and environmental impact. Collaboration between biotech firms, electronics manufacturers, and healthcare providers will accelerate commercialization.
By 2033, microfluidics will be a cornerstone technology across diagnostics, research, and industrial testing applications.
Extreme Ultraviolet (EUV) Lithography Market Key Opportunities and Risks 2025–2032
Extreme Ultraviolet (EUV) Lithography Market
Introduction
The Extreme Ultraviolet (EUV) Lithography Market is witnessing rapid growth as the technology emerges as a critical enabler for next-generation semiconductor manufacturing. EUV lithography utilizes extremely short wavelengths to achieve higher precision and finer patterning, addressing the growing demand for smaller, faster, and more powerful chips used in smartphones, data centers, artificial intelligence, and advanced computing applications. With the continuous push towards miniaturization and the integration of advanced node designs such as 5nm and beyond, EUV lithography has become essential for leading foundries and integrated device manufacturers. Growing investments from major semiconductor companies, advancements in lithography equipment, and increasing demand for consumer electronics and high-performance computing devices are driving the expansion of this market globally.
Extreme Ultraviolet (EUV) Lithography Market Size
Extreme Ultraviolet (EUV) Lithography Market size is estimated to reach over USD 36.11 Billion by 2032 from a value of USD 11.71 Billion in 2024 and is projected to grow by USD 13.25 Billion in 2025, growing at a CAGR of 13.4% from 2025 to 2032.
The Extreme Ultraviolet (EUV) Lithography Market encompasses the development, production, and deployment of advanced lithography equipment that enables semiconductor manufacturers to create highly integrated circuits with smaller nodes and greater efficiency. This market covers a wide range of applications, including logic devices, memory chips, and cutting-edge processors that power technologies such as 5G, artificial intelligence, cloud computing, and autonomous systems. The scope of the market extends across equipment suppliers, semiconductor foundries, and research institutions working on innovations to enhance EUV productivity and cost-efficiency. With increasing demand for high-performance and energy-efficient chips, the EUV lithography market is projected to witness significant growth, driven by technological breakthroughs, strategic collaborations, and expanding semiconductor manufacturing capacities worldwide.
Laser-Produced Plasma (LPP): The most widely adopted EUV light source technology, offering high power and efficiency for advanced semiconductor manufacturing.
Vacuum Spark: An emerging technology with potential but limited adoption due to lower output stability.
Gas Discharge: Used in research and pilot lines, providing alternative approaches to generating EUV light, though less commercially scalable.
By End-User:
Integrated Device Manufacturers (IDMs): Major semiconductor companies adopting EUV for in-house chip design and production of advanced nodes.
Foundries: Contract semiconductor manufacturers driving EUV adoption to meet global demand for advanced chips.
Memory Manufacturers: Increasing use of EUV in DRAM and NAND production to improve density and performance.
Research & Development Centers: Institutions focusing on EUV innovation, materials, and next-generation applications.
Regional Analysis:
North America: Dominated by strong semiconductor R&D and investments, particularly in the U.S. with key players in equipment and chip design.
Europe: Growth supported by leading lithography equipment manufacturers and EU initiatives in semiconductor self-reliance.
Asia-Pacific: The fastest-growing market, driven by large-scale semiconductor production in China, Taiwan, South Korea, and Japan.
Rest of the World: Gradual adoption supported by government initiatives and emerging semiconductor industries in developing economies.
Samsung to boost US investment for 2nm, advanced packaging capacity
August 11, 2025 /SemiMedia/ — Samsung Electronics is accelerating its semiconductor investment in the United States, with plans to raise total spending to as much as $50 billion to advance manufacturing capacity. The company’s wafer fab in Taylor, Texas, which had slowed construction due to weaker business momentum, is now regaining pace.
Last year, Samsung cut its planned US investment from $44…
July 3, 2025 /SemiMedia/ — Intel’s new CEO, Lip-Bu Tan, is considering a major shift in the company’s foundry strategy, potentially shelving the 18A process for external clients in favor of its next-generation 14A node, according to two people familiar with the matter.
Since taking the helm in March, Tan has moved quickly to cut costs and reshape Intel’s approach to reviving its struggling…
TSMC to reportedly delay 10nm risk production, first chips now expected by 2017
TSMC to reportedly delay 10nm risk production, first chips now expected by 2017
According to some rumors, TSMC may be planning to delay the risk production of 10nm chips by two quarters, delaying the launch of 10nm chips. Before TSMS, it was reported that Intel has decided to delay the mass production of 10nm chips. So it is starting to look like 10nm chips may not be here...
