Field noteNov 11, 20265 min read1 source

Goldman Sachs expects Anthropic revenue to hit $600 billion by 2030

In one minute

Goldman Sachs forecasts that Anthropic’s revenue will exceed $600 billion in 2030, while the company remains deeply investment-intensive. The Financial Times says the projections were shared with a potential investor during Anthropic’s IPO roadshow; for AI builders, they frame the commercial scale and infrastructure commitments behind a leading model provider, not a guarantee of future demand.

What changed: Goldman Sachs now projects Anthropic will exceed $600 billion in annual revenue by 2030 and become cash-flow positive.

What happened

The Financial Times reports that Goldman Sachs, Anthropic’s lead investment bank, expects the AI company to generate about $52 billion in total revenue in 2026, $136 billion in 2027 and $442 billion in 2029, before exceeding $600 billion in 2030. Anthropic reported revenue of $4.6 billion in 2025. The forecast implies sales more than 130 times that 2025 figure by the end of the decade. Goldman discussed its model with a potential investor as Anthropic began its initial public offering roadshow.

The projections are not the same as Anthropic’s preferred annualised revenue run rate, or ARR. Goldman’s estimates combine enterprise and non-enterprise revenue projections. ARR extrapolates the result from a particular month. Anthropic said in filings released on Tuesday that its annualised run rate reached $81 billion in October, $10 billion higher than the previous month. Those figures describe different measures, and the FT cautions against treating them as interchangeable.

Goldman’s model also projects substantial cash needs. It expects Anthropic to remain free-cash-flow negative until 2030, when it estimates the company would generate $33.6 billion. The bank forecasts free-cash-flow outflows of $12.8 billion in the fourth quarter of 2026 and $65.7 billion in 2027. The FT says Anthropic has committed to spend about $900 billion on computing capacity over the coming years, citing filings.

The operating-profit forecast follows a different path: Goldman estimates operating losses will peak at $40.4 billion in 2027, turn to operating profit in 2029, and reach $65.2 billion in 2030. Anthropic has never made an operating profit, according to its IPO filing’s financial reporting back to 2023. Its filings report an $8.1 billion operating loss for 2025 and $1.3 billion for the first nine months of 2026. After adjustments for stock-based compensation and non-cash charitable contributions, the company recorded $4.1 billion in operating income for those nine months.

The roadshow was seeking to sell money managers on a deal that could raise as much as $64 billion. At the high end of its indicated price range, the IPO would value Anthropic at $1.4 trillion, according to the FT. Anthropic did not immediately respond to the newspaper’s request for comment; Goldman declined to comment.

What the archive adds

New Runtime’s earlier coverage on October 21 covered Anthropic’s reported quarterly revenue growth, adjusted profitability and compute efficiency, while noting that those figures did not establish a durable unit-economics trend. The Goldman projections are a new fact on the same financial story, not confirmation that the forecast will be achieved.

Trend signal

This is a new data point, but it does not yet support a pattern New Runtime tracks. Goldman’s projections describe one company’s possible financial trajectory; they do not establish that the economically meaningful unit for agent systems is shifting to verified task completion or that AI providers have solved unit economics.

Why now

Goldman’s projections surfaced during Anthropic’s IPO roadshow, when bankers are seeking investor interest in a possible offering of up to $64 billion. The timing puts a long-range growth case alongside nearer-term scrutiny of the company’s losses and the infrastructure spending required to support expansion.

Connections to other events

New Runtime’s October 21 coverage described Anthropic’s reported quarterly revenue growth and compute efficiency, while distinguishing those indicators from proof of durable unit economics. The new projections make that distinction more consequential: very high sales estimates coexist with a forecast of heavy cash outflows through 2027 and positive free cash flow only in 2030.

Why it matters

For teams building with AI, vendor revenue projections are not a substitute for reliability, pricing, capacity or product-roadmap evidence. But the forecast illustrates the scale of investment Anthropic and its financiers believe may be needed to expand model services: Goldman’s estimate of $900 billion in planned computing capacity spending sits beside years of expected cash outflows. That combination matters to buyers assessing supplier durability and to builders whose products depend on sustained access to model infrastructure.

The numbers also show why financial comparisons need careful definitions. Anthropic’s reported $81 billion October ARR and Goldman’s $600 billion-plus 2030 total-revenue estimate are different measures with different time horizons. Treating either as a direct measure of current customer usage or profitability would overstate what the figures show. Procurement and architecture decisions should continue to rest on observed service performance, costs and contractual terms, not an investment bank’s scenario.

Limits and uncertainty

The figures are Goldman Sachs estimates relayed by the FT, not company guidance or realized results. Anthropic did not immediately comment, and Goldman declined to comment. The FT reports no detailed assumptions behind the bank’s revenue model, so its sensitivity to pricing, competition, customer retention, model costs or compute availability cannot be assessed here. The $900 billion commitment is reported by the newspaper as based on filings, but its timing and composition are not detailed in this account. Future operating and cash-flow results remain uncertain.

Reader takeaway

Goldman Sachs sees Anthropic growing into a $600 billion-plus revenue business by 2030, but its own model pairs that ambition with years of heavy spending and cash outflows. Treat the forecast as an investor’s scenario, not evidence that the economics or demand are already settled.

Evidence / sources

  1. 1ft.com/content/fd209548-d2df-470c-ab9a-caadbec08b58
All Field Notes

Next reads

Read next

Open the graph
  1. 01topicAnthropicExplore the anthropic topic hub.
  2. 02topicClaudeExplore the claude topic hub.
  3. 03archiveField NotesOpen the latest editorial analysis.
  4. 04sourcesSourcesSee the original sources this archive cites.
Who read this page?Machine requests, hidden until opened

Loading the privacy-safe route aggregate...

Open the JSON contract