AI industry marks financial maturity as chip rivalry deepens

The AI industry reported its first profitability milestones this week alongside stock-market records, even as a parallel race over chip architectures and model pricing intensified between Chinese and Western developers.

Anthropic closes $30 billion round and projects first quarterly profit

Anthropic is finalising a $30 billion funding round that would push its valuation to over $900 billion — above OpenAI's $852 billion private-market figure — co-led by Sequoia, Dragoneer, Altimeter, and Greenoaks. The company projects $10.9 billion in revenue and $559 million in operating income for the second quarter of 2026, its first quarterly profit ever. The figures mark a sharp change from the $380 billion valuation Anthropic carried just three months ago.

OpenAI prepares confidential IPO filing targeting $1 trillion valuation

OpenAI has retained Goldman Sachs and Morgan Stanley as advisers for a confidential SEC filing, with a public listing targeted as early as September 2026 at a valuation that could exceed $1 trillion. The move signals that the largest frontier AI developers are shifting from research-stage fundraising toward public-market finance.

DeepSeek makes 75% price cut permanent, escalating the global pricing war

Chinese AI developer DeepSeek has made its V4-Pro price cut permanent, bringing input pricing to $0.435 per million tokens — between 10 and 35 times cheaper than comparable offerings from OpenAI, Anthropic, and Google. The company attributes the reduction to efficiency gains from running its models on Huawei Ascend chips rather than Nvidia GPUs. China's National Development and Reform Commission is separately directing domestic AI developers to integrate homegrown chips into their full AI stacks as a matter of national policy.

Huawei proposes a 'Tau Scaling Law' to replace Moore's Law

Huawei has unveiled "LogicFolding," a chip design framework presented alongside a "Tau Scaling Law" as a successor to Moore's Law. The architecture promises a 55% increase in transistor density, has been used to mass-produce 381 chip types, and is set to reach Ascend AI processors by 2030, targeting 1.4-nanometre-class performance by 2031. This is Huawei's most detailed public claim yet about closing the gap with Western chipmakers under US export restrictions.

US states advance AI legislation as Congress hesitates

Colorado's governor signed SB 26-189 on 14 May, overhauling the state's AI Act to centre on automated decision-making in consequential domains such as hiring and credit. Connecticut is preparing to sign a bundle of AI rules that include regulations for AI companion bots, effective January 2027. Michigan introduced legislation to create a three-member AI governing board for state government. New York's Attorney General Letitia James joined a bipartisan coalition opposing a federal bill said to pre-empt state children's online-safety laws.

UK regulators issue joint AI warning; Britain and Australia deepen safety ties

The Bank of England, the Financial Conduct Authority, and HM Treasury issued a joint statement urging financial firms to treat frontier AI's ability to amplify cyberattacks as a material risk. The UK healthcare regulator MHRA separately called for "smarter regulation" to accelerate proven AI in clinical settings. The UK and Australia signed a memorandum of understanding on AI safety, agreeing to share capability assessments, conduct joint research, and support the International Network for Advanced AI Measurement, Evaluation and Science.

Micron Technology joins the trillion-dollar club on record AI memory demand

Micron's shares surged 18% on 26 May, pushing its market capitalisation above $1 trillion for the first time, driven by demand for high-bandwidth memory chips essential to AI model training. The company had previously announced that its entire 2026 HBM supply was already sold out. UBS raised its 12-month price target to an implied valuation of nearly $1.8 trillion.