A year ago, Anthropic was generating about $2 billion in annualized revenue. Today it's approaching $20 billion. By any normal business standard, that's a staggering trajectory. But Anthropic isn't operating by normal standards - and a Reuters Breakingviews column published today makes the case that the company's own projections might be its biggest vulnerability.
The core tension: Anthropic's internal target for 2026 is $55 billion in revenue. It's currently at roughly $19-20 billion in annualized run rate (ARR). That means the company needs to nearly triple its revenue in the remaining nine months of the year. The column borrows the AI industry's own terminology, calling this gap a "revenue hallucination" - projections that look convincing on paper but may not survive contact with reality.
The Growth Is Real, but So Are the Costs
Credit where it's due: Anthropic's growth numbers are genuinely remarkable. The company went from $1 billion ARR in December 2024 to $4 billion by July 2025 to $9 billion by December 2025. Claude Code alone generates roughly $2.5 billion in annualized revenue. Eight of the Fortune 10 use Claude models, and Anthropic now claims 40% of enterprise AI spending - up from a distant second place behind OpenAI a year ago.
But the cost side of the ledger tells a different story. The company projects spending around $12 billion on model training and $7 billion on inference (the computing power needed to actually run Claude for users) in 2026 - roughly $19 billion in compute costs alone. Gross margins have dropped to about 40%, down 10 percentage points from internal estimates, because inference costs came in 23% higher than expected. Anthropic posted an estimated $3 billion net loss in 2025 and pushed its target for reaching positive cash flow from 2027 to 2028.
The Valuation Math
In February, Anthropic closed a $30 billion Series G round at a $380 billion valuation - the second-largest venture deal in history, behind only OpenAI's $40 billion raise. At the time of that round, the valuation implied a 27x revenue multiple, which analysts noted was actually more disciplined than OpenAI's 42.5x multiple.
The comparison with OpenAI is instructive. OpenAI sits at roughly $25 billion ARR to Anthropic's $20 billion - a $6 billion gap that's been shrinking fast. Anthropic's growth rate (roughly 10x annually, now slowing to about 7x) far outpaces OpenAI's 3.4x. Some projections suggest Anthropic could overtake OpenAI in revenue by mid-2026 if current trends hold.
But CEO Dario Amodei himself has acknowledged the risk. He's publicly warned that if AI growth forecasts are off by even one year, "then you go bankrupt." That's unusually candid for a CEO sitting on $64 billion in total funding. The Pentagon blacklisting of Anthropic, reported earlier this month, adds another variable - executives have said the ban could cost billions in projected government sales and harm the company's reputation.
The AI industry's biggest companies are all making the same bet: that growth will continue at a pace that justifies their current spending. Anthropic is executing that bet better than almost anyone. The question Reuters is raising isn't whether Anthropic is a real business - it clearly is - but whether the distance between $20 billion today and $55 billion by December is a projection or a prayer.