SpaceX is spending $16.8 billion on a Texas chip fab Musk says will be '50 times the size of the Pentagon'
THE SO WHAT
A $16.8B, 100M+ sq ft Texas chip fab from SpaceX/Tesla is another data point that hyperscale compute buyers are turning into chip manufacturers. If you depend on third-party accelerators, assume your largest competitors may vertically integrate at the silicon level and plan for a world where access, not just price, is the constraint.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechQuantum Computing Inc. Revenue Jumps as Acquisitions Expand Commercial Business
Quantum Computing Inc. is buying its way into revenue with photonics and semiconductor adjacencies while still running operating losses—this is a platform-assembly play, not yet a proof of quantum product-market fit. If you’re considering quantum vendors, separate near-term photonics/semiconductor value from longer-dated quantum roadmaps in your contracts.
Deep & Emerging TechThe US government wants to make it easier to build and launch floating nuclear reactors, 'unleashing American energy dominance'
Floating nuclear reactors backed by friendlier international rules would shift baseload generation closer to ports and maritime corridors—energy security becomes a naval and regulatory question, not just a grid one. Industrial operators with coastal footprints should start mapping scenarios where cheap, mobile nuclear power changes site selection and process economics.
Deep & Emerging TechValve is warning Steam Machine buyers that scammers have their address
A logistics provider breach exposing Steam hardware customers’ addresses shows how your supply chain can quietly become your biggest data leak. If you ship physical product, extend your security audits to 3PLs and fulfillment partners—PII risk doesn’t stop at your own APIs.
Deep & Emerging TechThis Nvidia change could spell good news for Micron
If Nvidia is dialing down some chip capabilities in ways that drive up high-bandwidth memory consumption, the bottleneck shifts further from compute to HBM supply. Infra buyers should treat HBM availability and vendor concentration as a first-class risk variable in AI roadmap planning, not an afterthought to GPU allocation.