No Creds Notes #22
Fabs and Shredders
No Creds Notes #22
Hey guys!👋
Things are looking up over here in Southern California! Last weekend I got to visit with my whole family on their trip down here and this week I got my patio garden all set up for the summer. In particular I’m excited to be growing some Thai chilis and Holy basil, both of which are important ingredients in one of my favorite Thai dishes, Pad Kra Pao. Aside from that, this week we’ve got exciting stories on a semiconductor fab that fits in a shipping container and the first PROTAC drug ever approved by the FDA. Let’s dive in!
If this is your first time reading Uncredentialed, welcome! Subscribe below to receive regular updates from me on tech, startups, and strategy!
Fab in a Box
Researchers at MIT have built a functioning semiconductor fab that fits in a shipping container. They’re calling it InchFab, named for the 4-inch wafers it runs, and goes for a $5–15M price tag.
For context, a modern TSMC leading-edge fab costs north of $20 billion. Even a trailing-edge fab capable of running commodity automotive chips runs hundreds of millions. The direction of travel in chip manufacturing since the 1960s has been bigger fabs, higher capital costs, and fewer players but InchFab reverses that trend.
To be clear, it’s not capable of leading edge chips, instead aiming for trailing edge chips where urgency and/or integrity are of the utmost importance. Think defense contractors, biomedical device manufacturers, or aerospace suppliers where they aren’t building high end AI data centers, making a chip’s origin and turnaround time often matter more than whether it’s built on 3nm or 45nm.
Equipment for 4-inch wafer processes is decades old, well-understood, and cheap. It’s also outside the reach of modern export controls, which mostly concern leading-edge tooling. You can buy, deploy, and run a 4-inch process line without routing anything through TSMC or ASML.
The use cases include DARPA-funded programs and classified defense applications that don’t want custom chips running through a foreign supply chain and medical implant chips needing regulatory traceability (hard to guarantee across a multinational fab network).
The smarter bet for InchFab’s growth may not be selling chips at all, it’s selling the fab itself. If they can standardize the container unit and move enough of them, they’d ride learning curves on the manufacturing side even if each individual fab stays low-volume.
The Shredder Gets Approved
On May 1st, the FDA approved Veppanu for ER+/HER2- metastatic breast cancer in patients with an ESR1 mutation.
Veppanu is the first PROTAC ever approved by the FDA. A couple months ago, I wrote about Arvinas’ PROTAC for Parkinson’s disease crossing the blood-brain barrier, with Phase 1 data suggesting the mechanism could work in neurodegenerative disease. Today the same company got the first PROTAC across the FDA finish line. As a quick refresher on how it works, PROTACs are bifunctional molecules with 2 ends. One grabs the disease-causing protein and the other recruits an E3 ligase (part of the cell’s own protein disposal system). The ligase tags the target protein for destruction. The PROTAC gets released and goes back to recruit another one. Unlike drugs that block a protein’s function while leaving it present, PROTACs eliminate the protein entirely.
The reason this matters specifically in ESR1-mutant breast cancer is that the estrogen receptor has already done an end-run around the current drugs. Standard endocrine therapies (aromatase inhibitors, fulvestrant, etc.) work by blocking or suppressing the ER. In patients whose tumors have developed an ESR1 mutation, the receptor mutates to become constitutively active, driving tumor growth even without estrogen present, and it’s evolved to ignore blockers. Veppanu degrades the receptor rather than blocking something that has already learned to ignore being blocked, eliminating the protein altogether.
The VERITAC-2 trial ran Veppanu head-to-head against fulvestrant in ESR1-mutant and ER+/HER2- patients post-endocrine therapy progression, resulting in a 43% reduction in risk of disease progression or death.
The PROTAC concept has been in development since the early 2000s when Craig Crews at Yale published the first proof-of-concept molecules. 25 years from concept to first approved drug is roughly standard for a new mechanism class (monoclonal antibodies took about 20 years, checkpoint inhibitors about 15). What the approval actually unlocks isn’t just Veppanu. It’s the regulatory path. Every future PROTAC trial now has a precedent trial design, endpoint selection, and submission package. Arvinas has its Parkinson’s PROTAC and a prostate cancer PROTAC both in clinical development, all of which stand to benefit from this FDA approval.
No Creds Reads
Dylan Morris launched Stone Power
Noah Smith wrote on the AI messaging pivot
Eric Topol discussed medical AI
James Wang wrote on world models
If you enjoyed this post, please consider sharing with a friend! And if you haven’t already, subscribe to Uncredentialed!