Anthropic in talks to raise $30B+ at ~$900-950B valuation

The single most consequential thing about this round is not the absolute number — it is the slope. In March 2025 Anthropic was valued at $61.5 billion at its Series E close ^[7]. In September 2025 the Series F priced the company at $183 billion ^[7]. By February 12, 2026, Series G landed at $380 billion post-money, becoming the second-largest venture funding deal of all time behind only OpenAI's 2025 $40 billion round ^[8]. Now, less than ninety days after that Series G, Bloomberg reports talks at a $900 billion-plus valuation ^[1]. That is a roughly 15x re-rating in twelve months on the same underlying company, and it has become a recurring talking point across the investor community.

The re-rating is being underwritten by a revenue line that is itself moving at an unusual cadence. Anthropic's annualized revenue run rate has surpassed $30 billion ^[12], and PYMNTS reports it is 'expected to soon surpass $45 billion, up from $9 billion at the end of 2025' ^[6]. That implies roughly an 80x revenue increase in three years and pulls the implied forward multiple down from headline-shocking to merely aggressive. The pricing question is therefore less 'is Anthropic worth $900 billion today?' and more 'is the $45 billion run rate durable enough to amortize a $900 billion entry price for late-stage allocators who need a 3-5x outcome from here?' That is a different — and harder — question than the one the Series E investors had to answer.

Earlier Anthropic rounds were growth capital. This one is positioning capital. TechCrunch characterizes it as 'likely its last private round before IPO' ^[2], and Yahoo Finance describes the May board meeting as the decision point on whether to proceed and on what terms ^[10]. That framing matters because the pricing of a final private round sets the implicit floor for any subsequent S-1 — public-market underwriters are reluctant to price an IPO below the last private mark, and crossover funds expect a step-up. A $900 billion private valuation effectively commits Anthropic to a trillion-dollar-class IPO range, which is why CoinDesk has reported unauthorized secondary-market tokens already implying trillion-dollar valuations ^[11].

The second structural feature is the OpenAI leapfrog. CNBC notes that at a $900 billion-plus mark Anthropic would surpass OpenAI's roughly $852 billion valuation, making it the world's most valuable AI startup ^[4]. Yahoo Finance describes a May board meeting as the formal decision point for the round ^[10], which turns this private financing into a pricing benchmark fight as much as a fundraise. Anthropic's institutional pitch — articulated by GIC's Choo Yong Cheen as 'the clear category leader in enterprise AI' ^[7]— leans hard into enterprise revenue mix rather than consumer attention, an explicit differentiator from OpenAI's ChatGPT-driven story.

The bull case is loud. The bear case is more interesting because it is technical rather than vibes-based. AI CERTs flags that the round is being priced at roughly '27x forward sales multiples' and warns that 'private markdowns may spread across interconnected VC portfolios if sentiment shifts' ^[9]. Jefferies trader Jeffrey Favuzza has separately characterized the recent SaaS selloff tied to AI valuations as 'dramatic and fear driven' ^[9], suggesting that public-market participants are already pricing in a re-rating risk that private allocators have not yet absorbed.

The sharpest contrarian framing is the unit-economics critique surfacing in investor communities: that for foundation-model providers, selling more tokens means burning more cash, so revenue growth and gross-margin compression are coupled rather than decoupled. AI CERTs also flags the risk that 'compute costs rise faster than projected efficiency gains' ^[9], which is the cleanest way to articulate why a 27x sales multiple is fragile: the multiple compresses if gross margin compresses, and gross margin compresses if Nvidia and the hyperscalers capture more of each marginal token's economics than model vendors do.

Reading the cap table tells you who actually controls Anthropic's trajectory. Google has committed up to $40 billion — $10 billion immediate and $30 billion contingent ^[6]. Amazon has committed up to $25 billion plus 5 gigawatts of compute capacity ^[4]. Those two strategic checks dwarf the financial allocations reportedly in talks from Dragoneer, General Catalyst, and Lightspeed for the new round ^[4]. The reason is mechanical: 5 GW of compute is not something a financial sponsor can write a check for, and at Anthropic's projected scale compute is the binding constraint, not capital.

This is why the $30 billion financial round and the hyperscaler commitments are best read as two halves of the same deal. The financial round buys time, talent, and an IPO-ready balance sheet; the hyperscaler commitments buy the silicon and the power. Yahoo Finance's reporting that 'at least one institution willing to write a check of up to $5 billion' ^[10]is striking precisely because it is a financial actor underwriting a slice of compute spend that historically only Microsoft-for-OpenAI or AWS-for-Anthropic could absorb. The structural risk is that Anthropic's gross margin becomes a negotiated outcome between Google and Amazon rather than a market outcome — a dependency that public-market investors will scrutinize harder than private ones have.

Community reaction across the major platforms has converged on a specific frame: the pace, not the price, is what people cannot metabolize. The dominant social trope is the valuation ladder — $61.5B to $183B to $380B to $900B — recited as a punchline rather than a financial fact. The 'read that again' format applied to a $61.5B-to-$900B move in roughly twelve months captures something the headline numbers do not: this is happening faster than any post-2000 venture cycle, including the 2021 SaaS peak.

The more substantive reaction is bifurcated. Enterprise-leaning communities have focused on whether compute capacity and the Amazon 5 GW commitment justify the price, with secondary debate about defense linkages and enterprise pricing power. Career-focused communities have framed the question as personal exposure: whether to take equity at a $900 billion mark when the IPO step-up math is now structurally compressed. And contrarian investing communities have anchored on the unit-economics critique — that token-based revenue growth may not decouple from compute COGS in the way SaaS revenue decoupled from server COGS in the 2010s. None of these reactions are dispositive, but together they suggest the market is no longer debating whether Anthropic is a top-tier AI company. It is debating whether top-tier is worth $900 billion ninety days after it was worth $380 billion ^[8].