What has Anthropic’s prospectus changed in the outlook?

  • Anthropic generated 11.6 billion dollars in revenue in the second quarter of 2026, with an adjusted operating margin of 19%. DBS believes this strengthens confidence that model businesses can be profitable at a certain scale and cover compute spending.
  • As of the third quarter, net revenue retention was about 900%, indicating that spending by existing customers was expanding rapidly. Revenue outside the top three customers is not the measure used in the report; the report discloses the share of revenue from outside the top five customers, which rose from 64% in the first quarter to 79% in the third quarter, indicating lower customer concentration.
  • Operating cash flow in the first half of 2026 was 3.93 billion dollars, above capital expenditures of 482 million dollars; cash and short-term investments totaled 131.7 billion dollars, providing a financial cushion for future spending.

What should investors watch in October after growth slowed?

  • Anthropic’s third-quarter revenue grew about 45% quarter over quarter, down from more than 140% in the second quarter. DBS attributes the slowdown to the easing of the initial adoption surge for Claude Code, intensifying competition, and an estimated 5%–7% drag on revenue from tighter restrictions on model distillation.
  • DBS believes that new models and lower prices have improved value for money, while token usage on OpenRouter and spending indicators from Ramp also point to a recovery in momentum. The prospectus leaves a placeholder for annual recurring revenue data as of 31 October, and DBS expects the figure to be stronger.
  • The bank believes annual recurring revenue of 80 billion dollars or more by the end of October would support the case for a growth rebound; it was 65 billion dollars at the end of July, while the market expects 100 billion–110 billion dollars by year-end.

How could U.S. cloud earnings affect China’s AI sector?

  • DBS views new backlog, cloud revenue growth and cloud business margins as key indicators of AI monetization. If capital expenditure plans remain stable while these indicators meet or exceed expectations, investor confidence in compute demand and the earnings outlook for the industry’s supply chain may strengthen.
  • The bank believes that any further increase in capital expenditure in 2027 would need to be matched by corresponding demand to provide stronger support for the investment thesis. Sustained cloud growth and orders could also improve revenue visibility for upstream suppliers of chips, memory, optical components and power equipment, helping sentiment recover in China’s AI infrastructure sector.

Why does DBS prefer cloud providers but remain cautious on model companies?

  • DBS believes that independent Chinese model companies face domestic price competition, high R&D spending and limited access to advanced AI chips. Improved value for money among U.S. models could also weaken the software-efficiency advantage of Chinese models and their opportunities to earn revenue overseas.
  • The bank expects around ten Chinese companies actively developing large models to undergo consolidation. It prefers large cloud providers with profitable core businesses that can support investment and have stronger access to compute.
  • Alibaba is DBS’s top pick. The bank expects its cloud revenue for the quarter ending September 2026 to grow more than 50% year over year, up from 45% in the quarter ending June; its adjusted EBITA margin is expected to rise to 12.5%, above 11.6% in the previous quarter, reflecting how stronger demand is gradually translating into earnings.