AI Circle Daily
https://aicircle.news/archive/2026-10-20?category=leaders&lang=en
Scan to read this page
On the BG2 podcast, Nvidia CEO Jensen Huang said roughly a third of the business — classical computing with no AI labs involved — is growing 30% to 40% a year, framing Nvidia as competitive across all frontier segments.
Take: This is a message to Wall Street: Nvidia is hedging against AI-lab dependence. The more enterprises build their own models away from OpenAI and Anthropic, the more Huang can argue the shovel seller bets on nobody.
Microsoft is telling enterprises to closely watch agents employees run on systems not controlled by AI labs like OpenAI and Anthropic, arguing frontier models can't be trusted.
Take: Microsoft sells Copilot while putting shackles on third-party agents — turning a security narrative into an enterprise procurement gate and walling off model vendors it doesn't control.
Claude maker Anthropic pledged 5% of its shares to a new public benefit fund, structured as an LLC for flexibility, intended to make large donations or direct distributions after its IPO and address AI job disruption.
Take: A smart pre-IPO move: pre-hedging the 'AI eats jobs' backlash, using an LLC for flexibility, and giving regulators and the public something to point at.
Musk posted that he'll give one SpaceX share to 'every kid in America' next year, calling it 'the first of many' — hours after Anthropic said it would put 5% of its shares into a public benefit fund.
Take: Two billionaires racing to hand out cake on the same day isn't charity — it's an arms race over legitimacy of AI-era wealth. Whoever owns the 'everyone gets equity' framing gets a policy chip.
Reuters reports Anthropic and Elon Musk laid out sharply different blueprints for sharing the spoils of the AI boom, a split that may shape how the gains get distributed.
Take: Corporate benefit fund versus personal share giveaway — the real fight is over whether institutions or individuals control AI wealth distribution. Whoever wins writes the rules.
Amazon launched the TorchNeuron dev tool to break Nvidia's hold on PyTorch, aiming to get more developers running AI apps on alternative chips like its own.
Take: A frontal assault on the CUDA moat: whoever controls the PyTorch execution path can pry developers out of Nvidia's ecosystem, finally giving cloud giants' custom silicon a software lever.
Nvidia launched new open models aimed at enterprises that want to run their own models in their own data centers, while pushing on-device private AI agents.
Take: Huang is paving a second lane: even if customers don't buy top-end training GPUs, as long as they run agents locally, Nvidia's software stack stays the front door.
Nvidia CEO Jensen Huang urged German manufacturers to combine their expertise in physical machines with AI to 're-engineer' the economy, saying digital tech won't replace heavy industry.
Take: Door-to-door sales: German industry is Nvidia's next big customer pool. Sell the 'AI upgrades you, not replaces you' story first, then the GPUs.
Hugging Face's Irene Solaiman said AI labs have staked reputations on benchmarks, but the methods behind evaluations are complicated and still maturing, so embedded AI evaluators won't be an easy fix.
Take: It punctures the authority of leaderboards: once evaluation itself is a black box, vendors' 'benchmark score' narratives collapse and regulators and buyers need another yardstick.
Founder and CEO RJ Scaringe said autonomy will let Rivian build radically different kinds of vehicles than the outdoorsy SUVs and trucks it makes today.
Take: This is the next equity story: once the steering wheel isn't mandatory, Rivian's lineup and valuation logic get repriced — and it wants to stake the ground before Tesla does.