374 stories, newest first · page 2 of 14

6 stories

Policy & Law

Opponents win nearly all US data centre votes so far

Voters in 10 US states decided about 40 local questions on November 3. One early count shows 33 limits approved and one measure allowing data centres. In Texas, a candidate who promised a moratorium beat Governor Greg Abbott.

Early results from the US midterm elections show opponents of data centres winning almost every local vote. Ballotpedia counted 41 measures in 10 states before the vote. By 8 a.m. Eastern Time on November 4, Redbud Consulting counted 33 restrictions approved and one measure allowing data centres, Business Insider reports. Counting continues, and some votes are only advisory.

Chips & Infrastructure

AMD values its Taalas deal at about $2.9B, filing shows

Up to about $1.4 billion of that is contingent, on conditions AMD has not explained. The Toronto start-up builds chips around a single AI model, and AMD expects the deal to close by the end of 2026.

AMD’s quarterly filing, published on November 4, puts its pending purchase of Taalas at about $2.9 billion in cash and stock. That includes up to about $1.4 billion of contingent payment. AMD announced the deal in August but gave no price. Taalas, founded in Toronto in 2023, makes chips hard-wired for one AI model.

AI Models

Kimi K3.1 Pro’s weights are out, and Tencent launches Hy4

Moonshot’s 2.8-trillion-parameter model can now be downloaded, under a custom licence with limits for the biggest businesses. Tencent’s Hy4 adds vision and computer use, but its weights are only promised “in the coming days”.

Moonshot AI has published the full weights of Kimi K3.1 Pro on Hugging Face, keeping its promise to do so by November 4. Anyone can download them without asking. On November 4, Tencent launched Hy4 in its apps and through its API. Tencent says Hy4’s weights will follow soon, so for now only the August preview can be downloaded.

Safety & Security

UK cyber agency warns of AI ransomware attacks by 2028

The National Cyber Security Centre says AI-powered malware is “highly likely” to be used in a ransomware attack on UK critical infrastructure by 2028. Its chief warns of an “unprecedented rise” in incidents in the coming months.

The UK’s National Cyber Security Centre published its annual review and a new AI assessment on November 3. It judges it “highly likely” that Russia, China and other states will routinely use AI in cyber operations by 2028. Chief executive Richard Horne said AI has not invented new attacks. It exposes old weaknesses “at a scale that was once unimaginable”.

Chips & Infrastructure

AMD’s data centre sales nearly double to $8.5B

Total revenue rose 45% to $13.38 billion, and AMD forecasts about $17 billion for the next quarter. It says Anthropic, Meta, Microsoft, OpenAI and Oracle now run work on its new Helios AI racks.

AMD reported third-quarter results on November 3. Revenue rose 45% from a year earlier to $13.38 billion, above the $13.09 billion that analysts expected, according to LSEG. Data centre sales rose 97% to a record $8.5 billion, helped by server processors and new MI450 AI chips. AMD forecasts about $17 billion for the fourth quarter, and its shares rose after the report.

Business

Canva to cut, change or move 1,271 jobs as growth slows

That is 22% of its staff. Chief executive Melanie Perkins calls it the “hardest moment” in Canva’s 13-year history and says small changes won’t be enough. Reports say AI rivals and AI costs add to the pressure.

Canva told staff on November 4, Sydney time, that a restructure will affect 1,271 people, or 22% of the company. Some will leave, and others will change roles or be asked to move to central hubs. Canva will merge its product, growth and platform teams, and more product teams will be based in Sydney. Most hybrid staff will be in the office 2 to 3 days a week from January.

3 stories

Business

Anthropic’s IPO filing shows $17B revenue in one quarter

Anthropic’s updated prospectus, filed on November 2, shows revenue of $17.0 billion from July to September. That is about 12 times more than a year earlier. It also lists $895.4 billion of commitments, almost all for compute, and still no share price.

Anthropic filed an amended IPO prospectus with the US Securities and Exchange Commission on Monday, November 2. It adds results for the third quarter: revenue of $16.96 billion and a net loss of $23.29 billion. The filing says most of the loss comes from the higher value of its convertible notes, not from cash spent.

Chips & Infrastructure

Marvell agrees to buy Lumentum in a $115B bet on AI optics

The all-stock deal gives Lumentum an enterprise value of $115 billion, including debt. Marvell says copper links are reaching their limits in AI data centres. Both companies’ shares fell on Monday, and the deal is due to close in the first half of 2028.

Marvell and Lumentum announced on Monday, November 2, that they will combine. Each Lumentum share becomes 4.5 Marvell shares, worth about $1,132 at prices before reports of the talks. Marvell’s owners would hold about 67% of the combined company.

Business

Oracle plans to raise $20B, its whole limit through 2030

Oracle filed on Monday to sell $10 billion of new shares and $10 billion of bonds. Last week it promised to raise no more than $20 billion of new money through its 2030 financial year. Prices and interest rates are not set yet.

Oracle launched the two offerings in filings on Monday, November 2. It says the money is for general corporate purposes, which may include capital spending and paying back debt. Oracle is spending heavily on data centres for AI, and its shares fell on the news.

2 stories

Policy & Law

South Korea’s president asks for nearly double AI spending

Lee Jae Myung asked parliament on November 2 to pass a record 2027 budget. It puts 21.3 trillion won, about $16 billion, into AI, chips, physical AI and AI data centres. Lawmakers have until December 2.

President Lee Jae Myung gave his budget speech to the National Assembly on November 2. He said the 2027 budget puts 21.3 trillion won into AI and three “megaprojects”: chips, physical AI and AI data centres. He said that is nearly double this year’s amount. The whole budget is 820.9 trillion won, 12.8% more than this year.

Agents & Apps

Torvalds says AI-found fixes made Linux 7.4 extra big

The two weeks for new code brought almost 20,000 commits, against about 15,000 usually. Torvalds calls it a “much healthier model” that AI-found fixes now arrive early, not during testing.

Linus Torvalds released the first test version of Linux 7.4, called rc1, on Sunday, November 1. He wrote that this was a “BIG” merge window, with almost 20,000 commits, and he linked it to “AI triggered work”. He named a second reason too: many expect 7.4 to be a long-term support version.

4 stories

Agents & Apps

Apple Intelligence reaches China in December, with Qwen

Apple’s own models run the features, and Alibaba’s Qwen helps in Siri, Writing Tools and Visual Intelligence. Developers can test it from November 4, and only iPhones get it for now.

Apple said on October 31 that its first Apple Intelligence features are coming to iPhone users in China. Developers can test them in the latest iOS beta from November 4, and the public release follows in December with iOS 27.3. Through a partnership with Alibaba, its Qwen model adds answers in Siri, Writing Tools and Visual Intelligence.

Business

Aramco takes 49% of Saudi AI firm HUMAIN, adds $2.6B cash

Saudi Arabia’s oil giant closed the deal with the country’s wealth fund, which keeps 51%. HUMAIN says it plans 3 gigawatts of AI data centres by 2030.

Saudi Aramco and the Public Investment Fund (PIF), Saudi Arabia’s sovereign wealth fund, said on November 1 that they have closed a deal on HUMAIN, the country’s state-backed AI company. Aramco now owns 49% through a subsidiary, and PIF keeps 51%. Aramco puts in about 9.9 billion riyals ($2.6 billion) in cash during 2026, plus its own AI, computing and digital assets.

Agents & Apps

Anthropic lifts Claude’s 5-hour cap for verified Ultra users

Anthropic renamed its $200 Max 20x plan Ultra and its $100 Max 5x plan Max, with no price change. The weekly limits stay, and newer Ultra accounts must wait for an automatic check.

Anthropic renamed its two most expensive Claude plans on October 30. Max 5x is now Max, at $100 a month, and Max 20x is now Ultra, at $200. Verified Ultra accounts no longer have a usage limit that resets every five hours, starting with most existing subscribers. The weekly limit stays the same.

Chips & Infrastructure

Google’s satellite runs AI in orbit and sends back a haiku

Its Suncatcher prototype answered a first prompt with the Gemma 4 model, a month after launch. Google calls it the first AI response from a data-centre-class chip in orbit. Tests of heat and radiation come next.

Google said on October 30 that its Project Suncatcher prototype satellite has answered its first AI prompt in orbit. Engineers sent a prompt to Gemma 4, an open Google model on board, and asked for a haiku about being in space. The answer came down as the satellite passed over Svalbard, Norway.

3 stories

Safety & Security

Tampered Claude Code add-on in Cursor’s store had a backdoor

Versions 2.1.312 and 2.1.313 tried to give an unknown attacker remote access to developers’ computers. Cursor pulled them after about 11 hours, and Anthropic says it never built or published them.

For about 11 hours on October 29, Cursor’s extension store offered a tampered copy of Anthropic’s Claude Code extension. Cursor says the bad builds added an attacker’s SSH key, tried to open remote access and sent out details about each computer. Cursor shut the store down and rebuilt it, and it says removing the extension may not be enough.

Policy & Law

Google, Microsoft, OpenAI and Anthropic back one AI opt-out

A draft standard lets a website refuse AI training by all four companies with one robots.txt entry. Microsoft will honour it from February 1, 2027, and Google in early 2027. It does not stop crawling.

Microsoft, Google, OpenAI and Anthropic said on October 30 that they will support a shared signal for websites that do not want their pages used to train AI. The signal comes from a draft at the Internet Engineering Task Force (IETF), the body that writes many internet standards. A website adds one entry to its robots.txt file, instead of one for each AI company.

AI Models

Mistral releases the weights of its Large 4 model

Anyone can now download the 1-trillion-parameter model, under Mistral’s own licence. Mistral says it ties China’s GLM-5.3 at coding, but Artificial Analysis scores it one point lower there and further behind overall.

French company Mistral AI published the weights of Mistral Large 4 on Hugging Face on October 29, three weeks after a public preview. The model, nicknamed “Le Chonk”, has about 1 trillion parameters, and about 50 billion of them work on each token. Its API costs $0.70 per million input tokens and $2.10 per million output tokens.

4 stories

Chips & Infrastructure

Samsung says its 2027 HBM supply is already sold out

Samsung confirmed a record quarterly profit of 107.4 trillion won, almost all from chips. It plans to spend about 84.4 trillion won on factories and equipment this year, a record.

Samsung Electronics published its final results for July to September on October 29. Operating profit was a record 107.4 trillion won, about $80 billion, up 783% from a year earlier. On its earnings call, Samsung said it has finished selling its 2027 supply of HBM, the fast memory used in AI chips.

Business

US computer science enrollment drops 12% this fall

Early data from the National Student Clearinghouse show double-digit falls at four-year colleges, two-year colleges and graduate schools. Total college enrollment stayed about the same.

Fewer students in the United States are studying computer science, according to early fall 2026 data published on October 29 by the National Student Clearinghouse Research Center. Enrollment in computer and information science fell 12.4% at four-year colleges, 12.3% at two-year colleges and 10.4% at graduate schools. Total enrollment rose 0.2%.

Agents & Apps

Google takes ads in AI Mode to 10 more countries

Ads now appear in AI Mode answers in countries such as India, Canada and Australia, with Brazil, Mexico and the UK planned next. Google’s search sales grew 12%, more slowly than earlier this year.

Google said on October 28 that it is bringing ads in AI Mode, its chatbot-style search, to 10 countries outside the US. They include Australia, Canada, India, Nigeria and Singapore. In the US, ads will also start to appear in follow-up questions. The news came with Alphabet’s results, which showed search revenue up 12% to $63.35 billion.

AI Models

Nvidia’s open Nemotron model transcribes 40 languages

Nemotron 3.5 Transcribe turns recordings into text, translates and labels who is speaking. Anyone can download it for commercial use, and Nvidia also released labels for 463,062 hours of speech.

Nvidia released Nemotron 3.5 Transcribe on Hugging Face on October 28. It is an open-weight model that writes down speech in 40 languages, and Nvidia says it translates 31 of them to or from English. It can tell up to eight speakers apart. Nvidia also published Granary v2, a large set of labels for training speech models.