Where the compute constraint sits
The limit on how much AI companies can train and run is shifting from chip design to whether they can find a site, win local approval to build there, and buy enough hardware.
-
SHIPOpenAI publishes first Jalapeño inference-chip results
1.5 to 1.9 times more work per watt and 1.7 to 3.6 times lower end-to-end latency on SemiAnalysis's InferenceX. Internal deployment only, starting by year end. OpenAI
-
Local consent, not capital or silicon, is becoming the binding constraint on compute
Zvi Mowshowitz, August 24: about 75% of Americans oppose data center construction, and he argues the stated objections are downstream of distrust rather than physical harm. “People don't believe you, and even when they do believe you they don't care because that wasn't their true objection.” Anthropic committed $45B to Nscale two days later. Zvi
-
WATCHAnthropic commits $45B to Nscale for computing power
one supplier contract larger than most labs have raised in total. It prices the floor under Claude for years. Bloomberg
-
SHIPAnthropic previews a Model Hardware Standard
a driver layer that lets agents read from and write to physical devices, reachable over MCP. Waitlist only: no spec, SDK or repo yet. Anthropic
-
Anthropic planned, then dropped, a roughly $7B purchase of chip startup MatX. Single source. Reuters
-
a16z closed a $1.1B Machine Age fund for chips, memory, cooling, robotics and data centers. a16z
-
WATCHOpenAI buys tens of thousands of Mac minis and Mac Studios
For reinforcement learning and computer-use agent training. Anthropic is renting Mac mini capacity through AWS. Datacenter compute is spoken for, so labs are buying desktops. Roundup
-
EuroHPC signed a 387.8 million euro contract with Bull for LUMI-AI, running in Finland from late 2027. HPCwire
-
WATCHAnthropic reportedly signs $35B compute deal with Lambda
Bloomberg reports a deal for roughly 350 megawatts at Hut 8's Beacon Point campus in Texas, citing people familiar with the matter. Neither Anthropic nor Lambda has confirmed it. Single source. Yahoo Finance
-
Google is opening its TPU stack in a direct play at Nvidia's CUDA lock-in
SemiAnalysis's September 7 report on Google's InferenceX program found TPUv7 Ironwood beating Nvidia's B200/B300 by up to 50% on performance per dollar, with Google partially open-sourcing the TPU software stack to outside developers. Anthropic has committed to over 1 million TPUs, roughly 400,000 bought directly and 600,000-plus rented through Google Cloud. SemiAnalysis
-
Sep 10, 2026 Updated this evening
The DOJ formally requested information from Nvidia on its $20 billion licensing agreement with Groq, investigating whether the deal was structured to evade antitrust review; CEO Jonathan Ross and COO Sunny Madra joined Nvidia after signing. Bloomberg
-
WATCHNvidia's Huang projects $3 to 4 trillion in AI infrastructure spend by 2030
At Goldman's Communacopia conference September 10, Huang forecast roughly 70% revenue growth next year and dismissed circular-financing bubble concerns, days after Anthropic disclosed a 1-million-plus TPU commitment to a rival chipmaker. TechCrunch
-
WATCHPositron raises $875M to challenge Nvidia on inference cost
Positron's Series C, led by NEA and SemiAnalysis Capital, values the inference-chip startup at $5B. Its Asimov chip uses commodity LPDDR5X memory instead of HBM, claiming over 90% memory-bandwidth utilization against Nvidia GPUs' typical sub-30%. SiliconANGLE
-
Ayar Labs added $150M for co-packaged optics, bringing 2026 funding to $650M, plus a new Bengaluru design center. SiliconANGLE
-
WATCHAnthropic signs $13.7B compute deal with Trump-linked Rumble parent
Anthropic committed to a six-year, $13.7 billion compute deal with RUM Group, parent of Rumble and Truth Social, plus options on 51 million RUM shares — the same week Amodei led the industry's public pacing campaign. The Information · PYMNTS