Tuesday, 3 November 2026
Anthropic Tells Investors It May 'Imminently' Hit Its Own Threshold for AI-Driven Acceleration
Anthropic Tells Investors It May ‘Imminently’ Hit Its Own Threshold for AI-Driven Acceleration
A week after Anthropic first filed to list on Nasdaq, its amended filing from Monday puts a new sentence in the risk factors, and TechCrunch’s reading is that the passage is new. It says: “our internal estimates and our expectations regarding future models, rather than our current models, indicate that we may imminently meet our Responsible Scaling Policy’s definition of ‘dramatically accelerated progress attributable to AI.’” No current model has been designated as crossing that line, and Anthropic says it continues “to assess the underlying data.” The company also says it does not expect to be able to safely let Claude “fully autonomously design and build successive generations of frontier models without meaningful human supervision,” and that it is considering changes to its policy to keep humans in control. The same filing says the value of Anthropic’s own internal Claude use, at list prices, grew more than 250 times between January and September 2026, and in April passed its monthly staff costs. Sceptics point out that an S-1 is a sales document as well as a warning. Either way, the line to read twice is a different one: “we have in the past not identified, and may in the future not identify, a significant acceleration in our rate of development until after it has occurred.”
Source: Anthropic, amended S-1 filing with the SEC, November 2, 2026
Marvell Agrees to Buy Lumentum in a $115 Billion All-Stock Deal
The AI build-out’s next bottleneck is not the chip but the wire between chips, and Marvell has just bet $115 billion on it. Marvell, which designs custom AI and networking chips, will buy Lumentum, which makes the lasers and optical parts that move data between servers as light. Each Lumentum share becomes 4.5 Marvell shares, about $1,132 a share and a 23.9% premium to the October 28 close; Marvell’s holders end up with about 67% of the combined company. The two promise more than $1 billion a year in cost savings and name a $550 billion market. “AI infrastructure is undergoing a fundamental architectural shift toward optical connectivity,” their announcement says. Closing is expected in the first half of 2028 — which, at current rates, is three or four model generations away.
Source: Lumentum, 8-K exhibit filed with the SEC, November 2, 2026
Palantir’s US Commercial Revenue Overtakes Its US Government Revenue for the First Time
The week’s mood says AI spending won’t pay off; Palantir’s quarter says somebody is paying. In the third quarter its US commercial revenue reached $984 million, up 148% on a year earlier, against $953 million from the US government — the first time the companies have outspent the state. One quarter earlier it was $764 million to $809 million the other way. Total revenue was $2.313 billion, up 96%, net income $1.128 billion, and the full-year forecast was raised to about $8.53 billion, roughly 91% growth. Chief executive Alex Karp, never one for a quiet sentence, said “the Sovereign AI movement is imposing the most dramatic dichotomy between haves and have nots.” The haves, this quarter, are mostly Palantir.
Source: Palantir, Q3 2026 results, 8-K exhibit filed with the SEC, November 2, 2026
Y Combinator’s Garry Tan Says He Spends Up to $200,000 a Year on AI Tokens
The head of the world’s best-known startup accelerator has an AI bill most startups would call a seed round. On Walter Isaacson’s podcast The Builders, Garry Tan said: “I spend about maybe $100[,000] to $200,000 a year on tokens,” and called it “a crazy privilege that not everyone can do.” His agents write code, prepared him for the interview and help run his personal life: “It’s read all of my email and all of my Dropbox… even my therapy notes going back 20 years.” YC itself runs an internal agent called Stray Light with a “company brain” holding every founder, application and email. Tan’s own verdict on the price: “It’s too expensive right now.” Personal agents for everyone will have to wait until they cost less than the person.
Source: TechCrunch, November 2, 2026