The semiconductor industry's hardest problem is shifting from making chips faster to making them trustworthy. Three stories logged on this beat this month - a startup putting verification chips in luxury goods, another betting on AI to design chips, and a wave of films about tech founders - all point at the same thing. When chips end up in everything, the industry's standing with the public and its customers depends less on raw speed than on whether anyone believes what a chip says.
Veritas takes the chip out of the datacenter
As TechCrunch reported, Veritas is expanding its anti-counterfeit chip from its original territory into jewelry, trophies, and fragrance bottles. The company was founded by Luci Holland, a former Tesla technical staffer. Counterfeits of luxury goods are a longstanding problem for fashion houses, and counterfeiters are getting better at replicating brands' marks and passing off "superfakes" as genuine items. Veritas's answer is a chip that verifies goods.
On its face this is a luxury-goods story, several steps removed from the silicon business. It is not. It is the semiconductor industry's credibility being rented out to solve someone else's trust problem. A verification chip only works if buyers, retailers, and resellers believe the chip itself cannot be cloned or spoofed. That is a hardware security claim, and it sits on the same foundations - secure elements, unique identifiers, tamper resistance - that the chip industry has spent years struggling to harden everywhere else.
The expansion also matters for US technology companies because it shows the commercial path for security silicon outside the usual government and enterprise markets. If verification chips can find a paying home in jewelry and fragrance, the addressable market for authenticated hardware widens considerably, and the companies that supply those chips widen with it.
Ricursive bets on AI to design the chips
TechCrunch also reported on Ricursive Intelligence, whose founders detailed how their startup is developing AI that can design chips, learn from each iteration, and improve subsequent designs. The goal is to automate and accelerate the chip design process.
That is a claim about process, but it lands in the same place as Veritas: trust. Chip design is one of the few engineering disciplines where almost every step is still gated by human review, in part because the cost of an undetected error is enormous and the ability to trace who decided what is essential. An AI that learns from each iteration and changes subsequent designs is useful only if the people signing off can understand what it changed and why. The technical challenge of AI-designed chips is not just generating plausible layouts. It is making the output inspectable enough that a design team will stake a product on it.
The compute crisis framing Ricursive uses is real, and it creates the pressure. Demand for compute has outrun the industry's ability to design and verify new chips on the old timelines. Automation is one answer. But automation that cannot be audited simply moves the bottleneck from design to validation.
The trust problem goes both ways
Put the two startups side by side and a pattern emerges that is more interesting than either alone. Veritas is trying to make the provenance of goods verifiable through silicon. Ricursive is trying to make the provenance of designs - which decisions produced which chip - verifiable through process. Both are selling certainty in an industry where certainty has become scarce.
That scarcity has a cause. The Verge reported that 2026 has been unusually stacked with films taking a critical look at how Silicon Valley is reshaping reality, citing works including Aaron Sorkin's The Social Reckoning, which frames Mark Zuckerberg as a robotic villain whose ambitions are destabilizing society. The films are not about semiconductors. But the skepticism they dramatize is the environment in which US chip companies now operate, and it is a business condition, not just a cultural one.
