Nvidia Touts Progress Getting New Rubin Design to Customers
THE SO WHAT
Rubin moving into customer hands means Nvidia’s next-gen architecture is transitioning from roadmap to deployment, reinforcing its lead in high-end AI compute. If you’re planning large training runs in 2026–2027, lock in your Rubin access and power footprint now or risk queuing behind better-prepared buyers.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechWhy Micron and other chip stocks are bouncing back so strongly
The rebound in Micron and peers on the back of open-source AI demand is a reminder that memory is a leveraged bet on model scale and context windows. If you’re planning infra spend, assume DRAM and HBM constraints—not just GPU counts—will shape availability and pricing over your next 12–24 month roadmap.
Deep & Emerging TechBucking EV slowdown, Sila raises $300M to expand battery materials factory
A $300M raise to scale silicon-carbon anode production for 100,000+ EVs says the bottleneck is materials, not just vehicle demand. OEMs and fleet operators should be locking in next-gen cell supply now rather than assuming today’s lithium-ion chemistry and pricing will hold.
Deep & Emerging TechThis Former Intel CEO Wants to Jumpstart Moore’s Law With Light
Pat Gelsinger pushing light-based chips to extend Moore’s Law is a signal that traditional scaling alone won’t feed future AI demand. If you’re planning multi-year compute strategy, you should be tracking photonics and other non-classical approaches as potential step-changes, not just incremental node shrinks.
Deep & Emerging TechThe 8-Hour challenge: SK On says solid-state robot batteries must justify a 4x manufacturing cost jump to replace 'inferior' lithium-ion
SK On putting a 4x cost hurdle on solid-state batteries for robots clarifies the bar—runtime and performance gains have to be quantifiable in dollars per shift. Robotics buyers should be modeling TCO over an 8-hour duty cycle before paying a premium for next-gen packs.