AI Circle Daily
https://aicircle.news/archive/2026-09-26?category=leaders&lang=en
Scan to read this page
Nvidia CEO Jensen Huang said in a New York Times interview with Ezra Klein that AI labs unable to control the systems they build should stop developing them, putting accountability on developers rather than regulators.
Take: He's dumping the blame squarely back on the labs. It hands regulators a loaded line: even the builders admit they can't contain this.
OpenAI's Sam Altman and Anthropic's Dario Amodei warned at the UN Security Council that AI could threaten humanity and called for greater coordination, with Altman saying the AI race is no excuse for rash decisions.
Take: Two bitter rivals sharing a stage on safety is about framing the rules and owning the narrative. Trump rejecting 'globalist' controls means this fight is just getting started.
Top AI executives attended Trump's White House state dinner for Xi Jinping, where Meta CEO Mark Zuckerberg argued AI labs do not need to coordinate on safety.
Take: While others preach coordination at the UN, Zuckerberg says skip it. That's a green light for his own fast iteration — and a bet nobody can force him to slow down.
Nvidia CEO Jensen Huang downplayed AI's erosion of children's basic math skills, saying foundational math ability doesn't matter in the AI era.
Take: The guy selling the shovels says you don't need to know how to dig. Coming from the biggest beneficiary of compute, that lands badly — educators should be furious.
Musk predicted SpaceX could reach 'pole position' in AI in about six months, citing its ability to bring computing power online quickly, and hinted it could match Anthropic's Fable and OpenAI's GPT-6.
Take: Everyone else is racing on models; Musk is racing on how fast he can plug in power. That's 'compute is the moat' stated out loud — and fuel for the SpaceX/xAI valuation story.
Musk said xAI's Colossus 2 cluster could more than double its Nvidia chip count by the end of the year, laying out his most detailed expansion timetable yet.
Take: This is a public heads-up on a massive Nvidia order. Who's hoarding GPUs and how many is now an open book — the compute arms race has no brakes.
Musk said SpaceXAI will add another 660,000 AI GPUs this year, with Colossus 2 receiving its first batch of 220,000 GB300s next week.
Take: The number is absurd, but that's just 2026. Nvidia wins biggest, with utilities a close second.
Boom Supersonic CEO Blake Scholl said its new stationary power plants were no longer in Crusoe's near-term plans, after Crusoe abandoned a $1.25B plan to use Boom turbines at AI data centers.
Take: AI data centers' power anxiety just hit supply-chain reality. The turbine option is out, exposing the structural gap where compute runs ahead and electricity drags behind.
Anthropic committed $11.6 billion over seven years to Akamai's cloud infrastructure, potentially growing to about $20 billion, with Akamai granting a potential stake of up to 5% that grows as Anthropic spends more.
Take: It's a bet on CPUs and a way to lash a cloud provider to its wagon. Tying equity to spend means the supplier's fate rides with Anthropic's.
A federal appeals court upheld the Department of Defense's designation of Anthropic as a supply chain risk in a 2-1 decision, after Anthropic argued multiple violations of its rights.
Take: No amount of safety rhetoric stops a government blacklist. This is a heavy blow to Anthropic's public-sector business — and a reminder that compliance isn't optional.