Area
Reporting and discussion about Compute & Energy, with links to the original sources.
On Moonshots, the panel picked apart a rental index showing H100 prices rising 22% in a single month to $3.28 per GPU-hour. Dave called it a reversal of a lifetime of chip depreciation; Emad Mostaque explained why better models make the same old Hopper worth more; and the warning for companies was that the compute they assume will be there later is already sold out.
On Moonshots #288, a 4 a.m. chart about DeepSeek's new V4.1-Flash model sent the panel from cache statistics to the shopping list for an AI data center. DeepSeek says the model's lookup memory needs a quarter of the expensive high-bandwidth memory and an eighth of the SSD cache storage of its previous generation. The panel's argument was about what that does to a buildout in which, by one panelist's estimate, 40% of American capital spending goes to that one component.
On Moonshots, the panel played a launch video for Redwood, an accelerator that Palo Alto startup Architect Labs says its AI designed end to end from a specification written by two architects. The company's paper reports two weeks to verified design and FPGA deployment, with a small language model running in a third week — while the headline 3.4-times efficiency figure comes from a projected Samsung 8-nanometer chip that has not been built. The panel, who disclosed they are investors and an advisor, argued the real story is a "designless" company and recursive self-improvement at the chip layer.
On Moonshots, Peter Diamandis reported back from meetings with SK hynix and Solidigm leadership with a claim that memory, not compute, now limits AI. The panel argued that a changed workload and a supplier industry scarred by past busts are pushing prices up faster than factories can respond — and that the fix may be new chip designs, including etching model weights into silicon, rather than simply paying more.
On Moonshots with Peter Diamandis, Salim Ismail argued that a Mac Studio with 512GB of unified memory changes AI spending from a perpetual per-token bill into a capital asset, with law firms and mid-sized healthcare organizations as the likely buyers. Two other panelists agreed the machine was worth having and still called Apple's AI record a long-running software failure.
On Moonshots EP #284, the panel read out the performance figures OpenAI published for Jalapeño, the inference chip it built with Broadcom, and argued that inference is moving off NVIDIA. One panelist went further, imagining an "OpenAI Compute" cloud that rents capacity to competitors — possibly even hosting an Anthropic model. Dave Blundin explained why CUDA no longer locks buyers in at inference time, and why NVIDIA's next defense is the networking between chips.
On Moonshots with Peter Diamandis, one speaker argued that hyperscalers could win communities over with cheaper power, school programs and better-looking buildings. A fellow panelist replied that the fight has already been politicized beyond what benefits can fix — and the viral water figures at the center of it describe the whole country, not the town being asked to host the building.
Helion is targeting initial plant operation in 2028, followed by a ramp-up to at least 50 megawatts under its Microsoft agreement. On Moonshots, energy investor Ramez Naam explained what still separates encouraging fusion experiments from dependable electricity: whole-plant energy gain, durable components and a price customers can afford.
Ramez Naam sees AI demand as a powerful source of nuclear financing, but doubts new small reactors can supply electricity within the five-year window he considers reasonably predictable for investment. His argument turns on what can be delivered sooner—and whether repeated construction and factory production can make later plants cheaper.
Ramez Naam argues that new AI data centers can go where solar power is abundant, rather than waiting for electricity to reach established demand centers. Falling battery costs strengthen that case, but storing energy overnight is a different business from saving summer sunshine for winter. Land access, permitting and legal protection for AI models also shape where the computing can go.
Ramez Naam puts himself between those who dismiss orbital data centers and those expecting an imminent boom. He estimates that launch prices need to fall to roughly a quarter to a tenth of current levels for space-based AI to compete on cost. But cheaper flights leave a separate hurdle: building and launching enough hardware, with permission and reliability to keep flying.
Ramez Naam passed on Panthalassa’s early Bitcoin-mining pitch, then invested twice in 2026 at much higher valuations. The company now proposes wave-powered AI computing, cooled by seawater and connected by satellite. Its $140 million Series B is intended to support an Oregon pilot factory and northern-Pacific pilots; cheap electricity and longer-lived chips remain prospective benefits.