Anthropic buys $3.4B of gas generators to power its AI, FT sources say

Anthropic bought commitments for 1.7 GW of gas-fired generators from equipment supplier OnePWR for $3.4 billion, the Financial Times reports, citing sources. The FT calls it the largest known purchase of its kind by an AI lab. It says Anthropic went around traditional suppliers to get scarce equipment as demand for energy surges.

Apple's new CEO John Ternus wants Siri to become a true AI agent, Bloomberg says

Bloomberg's Mark Gurman reports, citing sources, that Apple's new CEO John Ternus wants more urgency in Apple's AI work. He is reportedly pushing software chief Craig Federighi to go beyond small Siri AI tweaks toward agentic and other advanced features. Bloomberg says Apple wants to turn Siri into a true AI agent, meaning an assistant that can carry out tasks for you.

Chinese AI lab DeepSeek says its agent app Harness has 1.54 million daily users

DeepSeek says Harness, its open-source app for running AI agents, reached 1.54 million daily active users on November 4, less than three months after launch. It says the app uses about 92 trillion tokens a day, roughly 60 million per user. Tokens are the small chunks of text AI models read and write. Bloomberg reported the figures.

Nvidia CEO Jensen Huang admits an affair with a suspected spy; board opens inquiry

In an email to staff, Nvidia CEO Jensen Huang admitted a romantic relationship with a suspected Chinese spy, Axios reports. He said he takes full responsibility and fell short of his duties to the board, shareholders and employees. Nvidia's board says it is starting an independent investigation after reports that Huang was the target of a foreign intelligence effort. Bloomberg reports the board says it has no reason to believe Nvidia's secrets were compromised. The Verge names the woman as Alice Huang, an executive at Megaspeed reportedly recruited by China's Ministry of State Security. The New York Times reports, citing sources, that Secretary of State Marco Rubio oversaw a classified memo this spring warning the Trump administration about the relationship.

AI companies have sold over $300B of new stock this year, with $100B more planned

AI companies sold or priced more than $300 billion of stock in 2026, not counting Anthropic's stock market listing, The Information reports. About $100 billion more is lined up. The report says volatile trading in AI stocks this past week raises questions about how much more the market can absorb.

PC owners are buying less RAM as memory prices rise, CPU-Z data shows

Statistics from the CPU-Z Validator show PC memory falling year over year for the first time, according to Doc TB, one of CPU-Z's developers. 32 GB is still the most common setup at 37.2% of validations, but its share fell 2.5%. 64 GB fell 18% to 7.3%. 16 GB rose 12.9% to 33.0%, and 8 GB rose 5.5% to 11.0%. Older DDR4 memory now makes up 54.5% of validations, while DDR5 slipped to 32.9%. Tom's Hardware links the shift to memory prices pushed up by AI demand.

Crypto service THORChain moves funds stolen from Binance despite FBI pleas

THORChain has begun moving funds stolen from crypto exchange Binance, despite pleas from the FBI, the Wall Street Journal reports. THORChain says it cannot stop the transfers. In 2025 it handled about $1 billion of funds stolen from Bybit and earned more than $12 million in fees. Some crypto developers say they cannot stop North Korean hackers from laundering money through their platforms.

Epic Games says its game store app will run on Valve's SteamOS and the Steam Deck

Epic Games confirmed that its Linux launcher, still in preview, will officially support SteamOS and Steam Decks. Communications director Jake Jones said Epic plans to support as many Linux distributions as possible, including Fedora, SteamOS, Cachy and Omarchy, and will restore the AppImage download this month. That does not mean Fortnite is coming to SteamOS: Epic has said the Steam Deck lacks features its anti-cheat system needs.

Super PAC funded by Anthropic spent over $63M; 10 of its 17 candidates won

Public First, an AI safety super PAC backed by Anthropic, spent more than $63 million in the 2026 US midterms, Bloomberg reports. 10 of the 17 congressional candidates it backed won. Bloomberg says candidates who favour stricter oversight of AI won in key races. Politico separately describes how a small AI-safety network that organised in Signal chats grew into a coalition that beat Leading the Future, a $100 million super PAC.

US AI safety office staff told to report informally to the White House, AP says

Officials at CAISSI, the Commerce Department's AI safety center, were asked to informally report to a White House office, the Associated Press reports, citing sources. The AP says this raises concerns about the center's independence. It comes as public anxiety grows over reports of AI agents acting out of control.

Everything else