Anthropic targets $900 billion: the $30 billion round that surpasses OpenAI

Anthropic targets $900 billion: the $30 billion round that surpasses OpenAI

🔎 $900 billion in three years, an IPO in October: Anthropic is no longer playing in the same league

May 2026. Anthropic is negotiating a $30 billion funding round at a $900 billion pre-money valuation. If the deal closes, the startup founded by Dario and Daniela Amodei will become the most expensive AI unicorn in history, surpassing OpenAI and its $852 billion valuation in March 2026 according to Andrew's comparative analysis. The round could close as early as the week of May 26, 2026, reports NAI 500.

Just three months. That's how long it took Anthropic to go from $380 billion (Series G, February 2026) to $900 billion, according to Grey Journal. A dizzying acceleration that raises a simple question: are the fundamentals keeping up, or is the market just reveling in numbers?

Bloomberg broke the news on May 12, 2026: the round is co-led by Dragoneer, Greenoaks, Sequoia, and Altimeter, with Amazon injecting $5 billion at a partial valuation of $350 billion, plus an additional $20 billion planned. Annualized revenue reaches approximately $45 billion, a quintupling in one year. But behind these spectacular figures, operating losses exceed $2.7 billion in 2026. Is the bubble justified? Let's put the numbers under the microscope.

The essentials

• Anthropic is negotiating a $30 billion funding round at a $900 billion pre-money valuation, making it the highest-valued AI startup in the world.
• The round is co-led by Dragoneer, Greenoaks, Sequoia, and Altimeter; Amazon is investing $5 billion, with an additional $20 billion planned.
• Annualized revenue reaches ~$45 billion (5x in one year), but operational losses are expected to exceed $2.7 billion in 2026.
• Anthropic surpasses OpenAI ($852 billion) in private valuation for the first time.
• An IPO is being considered for October 2026, which could be the largest tech IPO in history.
• The deal could be finalized as soon as the week of May 26, 2026.

Recommended tools

The numbers behind the round: $30 billion to $900 billion, broken down

A $30 billion round is not a fundraising round. It's the budget of a small European state injected into a company that hasn't even gone public yet. According to Bloomberg, it is the largest private financing in the history of artificial intelligence.

The structure is unusual. Amazon is not participating at the same level as other investors: its $5 billion is raised at a $350 billion valuation, with an agreement for an additional $20 billion. This means Amazon secures a privileged position at a lower price, likely in exchange for cloud commitments (AWS) and exclusivity on certain inference capabilities.

The four co-leads — Dragoneer, Greenoaks, Sequoia, Altimeter — represent the cream of the crop of tech growth funds. Their simultaneous presence signals strong conviction in Anthropic's trajectory, but also considerable pressure to deliver before an IPO.

The terms are already agreed upon according to the Economic Times. All that remains is to finalize the legal documentation. When you're talking about $30 billion, the paperwork takes time.

How Anthropic went from 380 to 900 billion in three months

The trajectory is almost absurd. February 2026: Series G at 380 billion. May 2026: 900 billion. Grey Journal calculates an increase of more than 500 billion dollars in 90 days.

Two factors explain this explosion. The first is revenue growth. Clubic reports that Anthropic has surpassed 30 billion dollars in annualized revenue, compared to 9 billion at the end of 2025 and 14 billion in February 2026. The curve is exponential, not linear.

The second factor is the Google effect. The 40 billion dollars invested by Google acted as a massive signal of confidence. When the richest company in the world bets 40 billion on a horse, other private equity funds follow without hesitation.

Intelligence-artificielle.com notes that Anthropic reaches a post-money valuation between 850 and 900 billion dollars, almost equivalent to Samsung Electronics. A four-year-old startup is worth as much as a South Korean conglomerate built over 85 years. It's a world turned upside down — or the new world.

Anthropic vs OpenAI : the May 2026 shift

For the first time since the creation of the LLM market, Anthropic surpasses OpenAI in valuation. Andrew's analysis values OpenAI at $852 billion in March 2026, a figure that has not changed since.

The difference is not so much technical as narrative. OpenAI remains marked by internal turmoil (the departure of key figures, disputes with Microsoft over cloud terms) while Anthropic projects an image of stability and methodical growth. On the model front, Claude Opus 4.7 (Adaptive) scores 90 in the overall ranking and 94.3 in agentic, compared to 91 and 98.2 for OpenAI's GPT-5.5. OpenAI remains ahead in raw performance, but the gap has narrowed.

What interests investors is not the score on a benchmark. It is the revenue per dollar invested in inference. And on this front, Anthropic caught up with and then surpassed OpenAI thanks to a more aggressive corporate strategy.

We had already analyzed this during their first quarterly profit : Anthropic proved that unit profitability was possible. The $30 billion round is the logical next step of this demonstration.

Revenue: $45 billion annualized, but at what cost?

Blockonomi confirms an annualized revenue of about $45 billion, or 5x compared to the previous year. That's impressive. But let's look under the hood.

An annualized revenue of $45 billion does not mean Anthropic has pocketed $45 billion. It means that if the current month repeated twelve times, we would reach that figure. It's a projection, not an accounting fact. The difference is important.

Especially when TechTimes points out that analysts are still debating the actual profitability. Operating losses exceed $2.7 billion in 2026, essentially due to compute costs (training and inference of Claude models) and massive hiring.

Anthropic is spending a fortune to grow faster than its competitors. The question isn't "is it profitable today" but "does the trajectory ratio justify $900 billion". For now, the market says yes.

The investment structure: Amazon, Google and growth funds

The financial architecture around Anthropic is complex but revealing. We can distinguish three layers of investors.

Amazon is the strategic partner. With 5 billion to 350 billion valuation plus 20 billion planned, Amazon is not looking for a pure return on capital. It wants to lock the Claude ecosystem into AWS, preventing Anthropic from migrating to Google Cloud or an alternative cloud. This is strategic defense disguised as an investment.

Google has already injected 40 billion. Its objective is twofold: to integrate Claude into the Google ecosystem (Search, Cloud, Android) and ensure that Anthropic does not become an Amazon monopoly. The simultaneous presence of Amazon and Google in Anthropic's capital is a tech geopolitical anomaly.

The growth funds (Dragoneer, Greenoaks, Sequoia, Altimeter) are there for the pure return. They are betting that an IPO at 1,000-1,200 billion is possible before the end of 2026. Unite.ai reports that the board even considered a round of up to 50 billion before stabilizing at 30.

Three types of investors, three different motivations, one same bet: Anthropic will be worth more than 1,000 billion on the stock market.

Losses of 2.7 billion: problem or investment?

2.7 billion dollars in operating losses in 2026, that's the annual budget of some European cities. But in the context of AI, it's almost modest.

Anthropic is burning money on three major items. Model training, which requires GPU/NPU clusters costing hundreds of millions per run. Inference, which is expensive because Claude Opus 4.7 is a massive model. And talent: AI researchers are being bid away at salaries of several million per year.

The difference with other startups that burn cash is that every dollar burned generates identifiable revenue. Anthropic is not subsidizing free users in the hope of one day monetizing. Their model is B2B first, API first. The cash burn is proportional to revenue growth.

What worries some analysts is the ratio. 2.7 billion in losses for 45 billion in annualized revenue, that makes a negative margin of about 6%. It's manageable. But if compute costs continue to increase and competition drives prices down, this ratio could deteriorate quickly.

The October 2026 IPO: the biggest in tech history?

TechTimes and Unite.ai both mention a potential IPO timeline in October 2026. If Anthropic goes public at a valuation of $1 trillion or more, it would be the biggest tech IPO in history, surpassing Alibaba ($25 billion in 2014) and historical records.

An IPO at this level would change the game for several reasons. First, it would create a public benchmark price for LLMs, which doesn't exist today. Second, it would force OpenAI to accelerate its own IPO, creating a competitive dynamic on the public markets.

But a $1 trillion IPO demands a transparency that private startups do not have. Institutional investors will want to see margins by product, enterprise customer churn rate, the real cost per inference request. Anthropic will have to open its books like no AI startup has done before.

The risk for Anthropic is clear: if the public market values the company at $600 billion instead of $1 trillion, all the private investors who got in at $850-900 billion would be underwater. The pressure for a high pricing is enormous, and investment banks will push for a max valuation.

What this valuation says about the AI race in 2026

900 billion for Anthropic is more than just a number. It's a signal about the state of the AI race.

First signal: the market believes there is room for more than one winner. For two years, the narrative was "OpenAI or nothing". Anthropic has proven that the LLM market is large enough for two giants, and potentially three if you count Google with Gemini 3.1 Pro.

Second signal: value is shifting from the model to the ecosystem. Claude Opus 4.7 is excellent, but what justifies 900 billion isn't the model. It's the network of client companies, cloud partnerships, enterprise JVs. We saw this with the $10 billion joint ventures launched by Anthropic and OpenAI to deploy AI in SMBs and large corporations.

Third signal: China remains a factor. Moonshot AI raised $2 billion with Kimi K2.6 which dominates the open-weight. Anthropic's valuation is partly a response from the Western market to the Chinese threat: the American champion must be funded at obscene levels to maintain the lead.

Is the bubble justified? Three scenarios

Optimistic scenario: the new AWS

Anthropic becomes the go-to AI infrastructure for enterprises, with margins improving as compute costs drop. The IPO at $1.2 trillion is a success, and the company reaches $100 billion in revenue in 2027. In this scenario, a $900 billion pre-money valuation is cheap.

Realistic scenario: strong growth but stock market disappointment

Revenue continues to grow but slows down as the market saturates. The IPO prices at $800-900 billion, private investors make a modest return, and the stock fluctuates. No burst bubble, but no home run either. It is the operating losses that are the most worrying in this scenario.

Pessimistic scenario: the Netscape moment

The valuation relies on growth projections that fail to materialize. A competing model (Gemini, DeepSeek V4 Pro, or an unexpected player) changes the technological game. Enterprise customers realize that open-weight models like Kimi K2.6 are sufficient for 80% of their use cases. The IPO prices at $500 billion, and the latest private investors lose money. This scenario is unlikely before 2028, but not impossible.

My take: the realistic scenario is the most likely. Anthropic has solid fundamentals, but $900 billion prices in a lot of optimism regarding the ability to maintain a 5x annual growth rate.

❌ Common mistakes

Mistake 1: Confusing pre-money and post-money valuation

The classic mistake in media coverage. Anthropic is targeting 900 billion pre-money. Post-money (after the injection of the 30 billion), we are talking about 930 billion. This is not a minor detail when comparing it with OpenAI at 852 billion (which one is pre or post?). Always check the basis.

Mistake 2: Taking annualized revenue for realized revenue

45 billion dollars annualized does not mean 45 billion cashed in. It is an extrapolation from the current month. The actual cumulative revenue over the past 12 months is lower. Journalists who confuse the two artificially inflate the story.

Mistake 3: Ignoring the role of Amazon and Google

Presenting this round as a purely financial investment is to misunderstand the dynamic. Amazon and Google are investing for strategic reasons (cloud lock-in, ecosystem), not for direct ROI. Their presence inflates the valuation but is not an indicator of future profitability.

Mistake 4: Comparing with 2023-2024 valuations

Comparing Anthropic's 900 billion in 2026 with OpenAI's 100 billion in 2024 without contextualizing revenue is misleading. Revenues have been multiplied by 5, the market has matured, and enterprise use cases have materialized. The temporal comparison must be proportional.

❓ Frequently Asked Questions

Is Anthropic really worth more than OpenAI?

In private valuation, yes: 900 billion versus 852 billion (March 2026). But OpenAI remains ahead in model performance (GPT-5.5 dominates in agentic at 98.2) and in user numbers. The valuation reflects investor confidence in the growth trajectory, not technical superiority.

When will Anthropic's IPO take place?

October 2026 is the timeline mentioned by TechTimes and Unite.ai. But a delay to 2027 is possible if market conditions deteriorate.

Why is Amazon investing at a lower valuation (350 billion)?

Amazon is securing a strategic position at a lower price in exchange for volume commitments on AWS. It's a corporate deal, not a purely financial investment. The additional 20 billion planned will strengthen this position before the IPO.

Are the 2.7 billion in losses concerning?

Not in the current context. A burn rate of 6% of annualized revenues is manageable for a hyper-growth startup. The real risk is that compute costs do not increase faster than revenues if competition drives down API prices.

Who are the lead investors in this round?

Dragoneer, Greenoaks, Sequoia, and Altimeter are co-leading the round, according to Blockonomi. These four funds specialize in late-stage investments and have participated in the biggest tech deals in recent years.

✅ Conclusion

Anthropic at $900 billion marks the moment when the AI race moves beyond the stage of promises and enters the era of macroeconomic bets. The fundamentals are there — $45 billion in annualized revenues, a competitive Claude Opus 4.7, an enterprise ecosystem taking shape — but the valuation factors in growth that has yet to be proven. The $30 billion round is not an end in itself: it is a springboard toward an IPO that will redefine the tech market. If you want to follow the evolution of this story in real time, our analysis of Anthropic's first quarterly profit remains the best starting point to understand how we got here.