On May 12, 2026, the financial news service Bloomberg reported that the American AI company Anthropic is in talks to raise about $30 billion in new investment. The deal would value the whole company at more than $900 billion — that is more than twice what Anthropic was worth only three months earlier, when it was valued at $380 billion.
If the round closes, Anthropic will be worth more than its main rival OpenAI. OpenAI, the maker of ChatGPT, was valued at $852 billion in its most recent funding round. The two companies are now in a tight race for the title of the world's most valuable AI start-up.
Anthropic is selling something called 'annualised revenue run-rate', or ARR. This is a way of saying how much money the company would earn in a full year at its current pace. The number was about $9 billion at the end of 2025, climbed to more than $30 billion by April 2026, and is on track to pass $45 billion by mid-summer.
On May 14, two days after the news of the funding round broke, Anthropic also signed a bigger deal with the global accounting firm PwC. PwC will give its workers around the world access to Anthropic's coding tool 'Claude Code' and a desktop tool called 'Cowork', and 30,000 US-based professionals will be trained and certified on the Claude family of models.
Bloomberg's Berber Jin and Riley Griffin first reported on the evening of Tuesday May 12, 2026 that Anthropic is in advanced discussions with a syndicate of investors led by Iconiq Capital and including Lightspeed Venture Partners, MGX of Abu Dhabi and Saudi Arabia's Public Investment Fund to raise approximately $30 billion at a post-money valuation in excess of $900 billion. The talks, which industry sources expect to close by the end of May, would represent the largest single private financing round in technology history and would lift Anthropic past OpenAI, whose most recent SoftBank-led round closed at $852 billion.
The pace of the markup is itself the story. Anthropic was valued at $61.5 billion as recently as March 2025; the Series G in February 2026 cleared $380 billion; the new round implies a roughly 2.4× jump in a single quarter and a roughly 14× jump in fourteen months. The justification investors are accepting hinges on annualised revenue run-rate (ARR), which company management showed to prospective backers climbing from $9 billion at the end of 2025 to north of $30 billion by April 2026, with internal projections of more than $45 billion by the time of an end-of-summer board update. Net retention is reportedly above 170 per cent, and the firm now claims more than 1,000 customers each spending over $1 million per year.
Investors are also being shown a thickening pipeline of enterprise wins. On Wednesday May 14, two days after the funding leak, Anthropic and PricewaterhouseCoopers announced the deepening of their alliance: PwC will roll out the Anthropic-built coding agent Claude Code and the new Cowork desktop product to its global workforce of nearly 400,000, certify 30,000 US-based professionals on the Claude model family, and stand up a joint Center of Excellence with Anthropic. The previous quarter had brought disclosed enterprise wins with Blackstone, Goldman Sachs, BBVA and Novo Nordisk.
The risks are commensurate with the price tag. Compute remains capital-intensive even after Anthropic's own DeployCo-style commitments from Brookfield-led infrastructure funds; the talent market for senior researchers has pushed median offers above $7 million per year of cash and equity; and the regulatory horizon is darkening, with the US Federal Trade Commission, the EU AI Office and the UK Competition & Markets Authority all running parallel reviews of the cloud-frontier-lab partnership model. Several limited partners contacted by the Financial Times have asked the round's lead to reduce their commitment, citing concentration limits in their private-markets sleeves now that two AI start-ups exceed the market capitalisation of every public stock except Apple, Microsoft, Nvidia, Aramco and Tesla.
The Bloomberg scoop that broke on the evening of Tuesday May 12, 2026 confirmed what San-Francisco insiders had been whispering for a fortnight: Anthropic is in advanced discussions to raise approximately US$30 billion at a post-money valuation north of US$900 billion in a syndicate led by Iconiq Capital and including Lightspeed, MGX, Saudi Arabia's Public Investment Fund and a renewed cheque from Google Cloud's strategic-investment arm. The talks, which insiders expect to clear final term-sheet on or before May 29, would represent the largest single private financing round in the history of the technology industry and would propel Anthropic past OpenAI, whose SoftBank-led tender closed at US$852 billion in late March.
The velocity of the markup is unprecedented. The Bessemer Cloud 100, which has tracked private-cloud valuations for over a decade, lists no precedent within an order of magnitude: Anthropic was last marked at US$61.5 billion in March 2025; the Series G of February 2026 cleared US$380 billion; the present round implies a 2.4× quarterly markup and a 14× year-on-year markup. Investors are accepting the price on the strength of an annualised revenue run-rate that has compounded from roughly US$9 billion at the end of 2025 to over US$30 billion in April 2026, with management projecting more than US$45 billion by the end-of-summer board meeting. Net dollar retention is reportedly above 170 per cent — territory previously claimed only by Snowflake at peak; more than 1,000 customers are spending in excess of US$1 million per year, and the average contract value among the top quintile clears US$11 million.
Enterprise momentum supplied the qualitative backdrop. On Wednesday May 14, two days after the funding-round leak, Anthropic and PricewaterhouseCoopers announced a substantial extension of their alliance: PwC will roll out Claude Code and the recently launched Cowork desktop product to its global workforce of nearly 400,000, certify 30,000 US-based professionals on the Claude model family, and stand up a joint Center of Excellence to industrialise reusable agentic-workflow templates. The Q1 logo list also included Blackstone, Goldman Sachs, BBVA, Novo Nordisk and a recently signed pilot at Saudi Aramco — the latter underwriting a US$3 billion three-year compute pre-commitment to Google Cloud's TPU v6 capacity that helps explain Google's renewed equity participation.
The risks are commensurate. Frontier-training capital intensity remains structurally elevated; Anthropic's DeployCo-style infrastructure vehicle, anchored by Brookfield and TPG, has raised a separate US$8 billion to underwrite a 1.4-gigawatt Texas data-centre build, but the firm is signalling pre-training spend approaching US$28 billion in 2027. The labour market for senior research leads has pushed median offers above US$7 million in cash-and-equity per year, and the legal-and-regulatory perimeter is darkening: the US Federal Trade Commission, the EU AI Office and the UK Competition & Markets Authority all have parallel investigations open on the model-provider–cloud-platform partnership structure that has come to define the sector. The Financial Times reports that several pension and sovereign limited partners contacted by the deal's lead have asked to reduce their pro-rata commitments, citing concentration limits now that two AI start-ups exceed the market capitalisation of every publicly listed company on earth except Apple, Microsoft, Nvidia, Aramco and Tesla.
Bloomberg reported on May 12, 2026 that Claude-maker Anthropic is negotiating a $30 billion funding round that would push its post-money valuation north of $900 billion — more than double the $380 billion mark it set only three months earlier and ahead of OpenAI's most recent $852 billion private valuation. The round, in talks for several weeks and now expected to close by month-end, is supported by an annualised revenue run-rate that has climbed from roughly $9 billion at the end of 2025 to over $30 billion in April 2026 and is on pace to clear $45 billion by mid-summer. The fundraise comes alongside an enlarged enterprise alliance with PwC announced on May 14, under which the Big-Four firm is rolling out Claude Code and Cowork to its global workforce and certifying 30,000 US professionals on Anthropic's models.
There is a company called Anthropic. Anthropic makes an artificial intelligence (AI) called Claude. Claude is like a smart chatbot.
Anthropic wants to get a lot of money from investors. The number is very big: 30 billion dollars. The full value of the company is more than 900 billion dollars.
Three months ago, the company was worth 380 billion dollars. Now it is more than two times bigger. It is also bigger than another famous AI company, OpenAI.
Last week, Anthropic also made a deal with a big accounting firm called PwC. PwC will teach 30,000 of its workers in the United States how to use Claude. The news is from May 12 to 14, 2026.
1What is Claude?
2How much money does Anthropic want from investors?
3How much is the company worth in total?
4Three months ago, how much was the company worth?
5Which accounting firm signed a deal with Anthropic?
6Anthropic makes Claude.
7The valuation went down in the last three months.
8Anthropic is now bigger than OpenAI.
9PwC will teach 30,000 workers how to use Claude.
10The news is from December 2025.
11Anthropic makes an AI chatbot called ___ .
12Anthropic wants to raise ___ billion dollars from investors.
13PwC will teach 30,000 workers in the ___ States.