Nvidia’s Huang Expects to Sell Twice as Many Chips Next Year
THE SO WHAT
Planning to double chip sales on the back of AI demand means the constraint is shifting from GPU availability to power, cooling, and capital. If you’re budgeting large-scale training or inference, assume capacity will exist but at a premium for those who haven’t already locked in long-term supply and energy.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechMicron, Intel and other chip stocks are furthering their strong comebacks
Chip equities rebounding as AI slowdown fears fade tells you public markets are re-rating semis as a structural, not cyclical, AI play. For operators, that means memory, storage, and general-purpose compute are back in the scarcity conversation—plan multi-year supply and pricing, not quarter to quarter.
Deep & Emerging TechAura data removal review: It chases data brokers, but needs some minding
Tools like Aura that automate data-broker opt-outs but still need human oversight show how messy privacy operations remain in practice. If your business depends on consumer data, assume more of your dataset will be actively contested and design acquisition and consent flows that can survive automated opt-out pressure.
Deep & Emerging TechUS data center natural gas consumption could surpass Germany and Japan combined by 2035
Projected US data center gas consumption rivaling Germany and Japan combined by 2035 makes AI an energy story as much as a compute story. If you’re planning large deployments, you now have to map gas and grid exposure alongside GPU supply—site selection and energy contracts become board-level topics.
Deep & Emerging TechHackers extracted a Flock camera’s software and key, 404 Media reports
Pulling firmware and encryption keys off a deployed Flock camera shows how brittle edge security can be once hardware is in the wild. If you deploy or buy connected sensors, assume physical access equals root and design your auth, logging, and data retention accordingly.