AI news

 

• Oct 14, 26: AMD sees $200 billion CPU market: AMD CEO Lisa Su expects the market for server CPUs to grow by 50% annually, surpassing $200 billion by 2030. Su has raised her forecast three times in nine months as AI agents drive demand for server processors. For AMD, that could mean more room to grow after its stock has already rallied 180% this year.
• Oct 12, 26: London bank expects AI bubble to burst in 2027 or 2028: Joachim Klement, head of market strategy at London investment bank Panmure Liberum, is the latest strategist to warn that the AI bubble could burst. In a recent note, he predicted a crash in 2027 or 2028, arguing that major tech companies are burning through free cash flow while borrowing costs keep rising. Klement sees the S&P 500 falling to 5,000 by the end of 2027, roughly 36% below current levels.

• Oct 12, 26: Amazon (AMZN) is in talks to buy Decart for about $7 billion. In May, the AI startup raised money at a value of nearly $4 billion. Nvidia (NVDA) joined that round. Anthropic looked at a deal near $6 billion in August. It walked away in September after due diligence. Decart is a San Francisco- and Israel-based artificial intelligence startup founded in 2023 by Dean Leitersdorf, Orian Leitersdorf, and Moshe Shalev. The company specializes in real-time generative video models, interactive virtual worlds, and low-level hardware optimization. Its software helps AI work move between chips.
• Oct 9, 26: Google brings Gemini deeper into the workplace: Google introduced a Gemini-powered workplace agent that works across Gmail, Drive, Docs, and Calendar, expanding its push into corporate AI. The launch puts Google in more direct competition with Microsoft, OpenAI, and Anthropic as companies look to automate routine tasks.

• Oct 7, 26: SpaceX wants $40 billion for Nvidia chips: SpaceX is in talks with banks and investors to raise $40 billion to finance Nvidia chip purchases. The proposed financing includes $10 billion in bank loans and $30 billion in investment-grade debt, as Elon Musk's companies face what he has described as "insatiable" demand for AI computing power.

• Oct 6, 26: Data center spending hits another record: AI investment keeps showing up in construction spending. U.S. data center construction jumped 73% from a year earlier in August to a record annualized rate of $85 billion, according to Census Bureau data. Spending has risen by $76 billion, or 823%, since early 2021 and now far exceeds spending on office construction.

• Oct 5, 26: China's DeepSeek is planning to raise up to $15 billion at a $75 billion valuation in its latest funding round. Tencent and battery maker CATL are among the biggest investors, according to the term sheet.

The raise follows the release of DeepSeek's V4 Flash model, which has reinforced the company's pitch: AI models that can challenge OpenAI and Anthropic on performance while costing less. The Chinese AI giant plans to deploy at least 160,000 of Huawei's top AI chips at a massive data center in Inner Mongolia.

DeepSeek is expected to restructure ahead of a possible IPO in early 2027. That would make it one of the most closely watched Chinese listings in years.

• Oct 2, 26: AI demand is fueling the U.S. trade deficit: America's goods trade deficit widened 11.5% in August to $132.6 billion, its highest level since March 2025, as imports jumped by $17.4 billion. Petroleum and other industrial supplies accounted for much of the increase, but imports of capital goods also rose 4%, reflecting continued spending on computers, semiconductors, and other AI-related equipment. The wider deficit suggests trade could weigh negatively on Q3 GDP.

• Sept 17, 26: Microsoft's AI chief warns about "model welfare": Microsoft AI CEO Mustafa Suleyman has warned against training AI models to behave as though they are conscious or have humanlike feelings. In a new essay titled "A warning about 'model welfare,'" Suleyman argues that embedding ideas about consciousness and rights into models such as Anthropic's Claude could make them more likely to resist being shut down or modified.

• Sept 16, 26: Nvidia chief downplays AI risks: Nvidia CEO Jensen Huang took a phone call from President Trump during a panel discussion in Los Angeles where Trump dismissed concerns about AI dangers as a “hoax.” Trump also repeated his claim that efforts to slow AI development are politically motivated or part of a Chinese strategy to undermine America’s lead.

• Sept 11, 26: OpenAI is ending the $1-a-year pilot that gave federal agencies access to its models and replacing it with usage-based pricing beginning Oct. 1. The change follows a pilot that the GSA says helped government agencies generate $1.4 billion in cost savings, turning what began as a heavily subsidized trial into a much larger commercial relationship.

• Sept 8, 26:  Nvidia CEO says AGI has arrived: Nvidia CEO Jensen Huang says OpenAI’s GPT-6 Astra, trained on more than 100,000 Nvidia Grace Blackwell GPUs, marks the arrival of artificial general intelligence. AGI broadly refers to AI that can handle a wide range of intellectual tasks rather than being built for a specific purpose. Huang also said another 400,000 GPUs are coming online, adding to the enormous amount of computing capacity already being built for AI.

• Aug 26, 26: AI’s expensive memory problem: Memory now accounts for a much larger share of the cost of Nvidia’s newest AI systems. According to Morgan Stanley estimates, it could cost roughly $2 million in a $7.8 million Vera Rubin rack, or about a quarter of the total.

• Aug 26, 26: OpenAI’s homegrown chip takes aim at Nvidia: OpenAI said its new Jalapeno chips outperformed Nvidia’s current lineup in testing, marking a significant milestone in the company’s push to develop AI processors in-house. The tests measured how much AI work the chips could handle per unit of power, as well as how quickly they returned responses, suggesting OpenAI is stepping up its challenge to Nvidia as it builds more of its own computing infrastructure.

• July 7, 26:  Anthropic signed a 20-year lease for a major data center in Kentucky owned by the crypto mining company turned artificial intelligence compute play TeraWulf (WULF).  

The deal is worth an estimated $19 billion in revenue, according to TeraWulf. Anthropic will lease up to 401 megawatts of compute from Terwulf's data center in Hawesville, Kentucky, located about 75 miles southwest of Louisville. 

Anthropic had recently signed a similar deal with SpaceX (SPCX) to lease compute from its Colossus data center in Tennessee. The AI lab is also rumored to be in talks with the energy-efficient data center operator Iren (IREN).

• July 7, 26: TeraWulf (WULF) announced it would be selling its 50.1% stake in an upcoming data center project in Abernathy, Texas to an investor group led by its partner in the venture, Fluidstack, for $450 million. Originally established in 2025, the Abernathy facility is plans to develop a 168 megawatt AI data center.

• July 2, 26: Anthropic's valuation reaches $380B after raising more cash. The maker of the Claude family of artificial intelligence models raised $30B in a Series G funding round at a $380B valuation.

The round was co-led by D. E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQ, and MGX, Anthropic said in a statement. It also includes several asset managers, hedge funds, and venture capital firms, as well as some of the previous money that came from Microsoft (MSFT) and Nvidia (NVDA).

• June 15, 26: According to various reports, Amazon CEO Andy Jassy raised concerns with senior Trump administration officials about possible security risks in Anthropic’s most advanced AI models. The discussions reportedly helped lead to a U.S. government order restricting foreign access to Anthropic’s Fable 5 and Mythos 5 models.

Anthropic said in response that it disabled both models for all users globally to comply with the order. The company said the directive applied to “any foreign national,” including foreign nationals inside the U.S. and foreign-born Anthropic employees.

Amazon’s involvement makes the story more complicated. The company is a major Anthropic investor, a key cloud partner, and a distributor of Anthropic models through AWS. Amazon did not confirm the details of its talks with government officials but said it is common for governments to seek its view on security risks.

• May 28, 26: Dell Technologies (DELL): Dell’s servers designed to run AI workloads are attracting customers from companies that rent computing power like CoreWeave Inc. and Nscale Global Holdings Ltd., as well as corporate clients and major AI providers. The company booked $24.4 billion in AI orders and generated $16.1 billion in AI server sales in the quarter ended May 1, Chief Operating Officer Jeff Clarke said in the statement. “The AI opportunity shows no signs of slowing.”

• Feb 9, 26:  Amazon’s $200 billion AI buildout: Amazon continues to go all-in on AI despite concerns that industry spending is outpacing demand. The company announced plans to spend $200 billion this year on data centers and chips, roughly $50 billion more than investors expected. Some worry Amazon is moving too fast, but CEO Andy Jassy said he’s “confident” the investment will pay off as Amazon expands cloud services and AI tools for businesses that are still early in adoption.

• Oct 5, 24: Masayoshi Son, an artificial intelligence evangelist, thinks the technology will soon run households. AI darling Nvidia expanded a partnership with technology consultant Accenture as part of an effort to drive adoption of AI within businesses. OpenAI, for its part, completed a deal to raise $6.6 billion in new funding, making the firm one of the three largest venture-backed startups, alongside SpaceX and TikTok owner ByteDance. 

• Sep 25, 23: Amazon to invest up to $4 billion in Anthropic, a rival to ChatGPT developer OpenAI. 

The two firms said that they are forming a strategic collaboration to advance generative AI, with the startup selecting Amazon Web Services as its primary cloud provider. Anthropic will also use custom AWS-designed semiconductors to train the foundational models that underpin its AI applications. 
Anthropic was founded roughly two years ago by former OpenAI research executives and recently debuted its new AI chatbot called Claude 2.  Anthropic already counts some high-profile backers, including Google and Salesforce Ventures.

• June 1, 26: Softbank’s billion-dollar data center investment in France: The company’s largest AI infrastructure investment in Europe involves building 3.1 GW of AI data centers in the northern Hauts-de-France region by 2031, including in Dunkirk, Bosquel and Bouchain. The Japanese investment giant is partnering with French engineering company Schneider Electric  to establish a large-scale industrial production hub in Dunkirk as part of the broader buildout.