GIC invested in Anthropic 3 times in a year, AI behemoth valued at S$1.2 tril...
Source: Mothership
Singapore's sovereign wealth fund GIC has been quietly building one of the most concentrated AI investment portfolios among global institutional investors, and its latest annual report offers an unusually detailed look at how it picks winners in the space. The fund backed Anthropic three times in under a...

Singapore's sovereign wealth fund GIC has been quietly building one of the most concentrated AI investment portfolios among global institutional investors, and its latest annual report offers an unusually detailed look at how it picks winners in the space. The fund backed Anthropic three times in under a year — from a US3 billion Series F in September 2025 to a staggering US5 billion Series H in May 2026 — valuing the Claude AI developer at roughly S.2 trillion by the final round. But the numbers only tell part of the story; what's more revealing is the investment framework GIC has developed to navigate a sector where hyperscaler capital expenditure is now approaching 2 per cent of US GDP.
The three Anthropic rounds tell a clear narrative of escalating conviction. GIC first participated in Anthropic's US3 billion Series F at a US83 billion valuation in September 2025, then led a US0 billion Series G in February 2026 at US80 billion, and finally co-led the US5 billion Series H in May 2026 at US65 billion — roughly S.2 trillion. Over the same period, Anthropic's run-rate revenue surged from around US billion at the end of 2025 to roughly US7 billion today, a more than fivefold increase in under a year. GIC declined to disclose its current stake size, but the escalating round sizes suggest significant allocation growth.
GIC's approach to AI investing goes beyond simply writing cheques. The fund uses two frameworks to guide deployment: a value-chain map that categorises AI exposures into Enablers (infrastructure like computing hardware and energy), Monetisers (AI-powered products and services), and Adopters (companies integrating AI into existing operations), and a company-level filter called the three Ms — Moat (structural advantages like proprietary data), Management (leaders who translate AI into real business outcomes), and Momentum (the self-reinforcing cycle where early wins generate better data and performance). This granularity matters because AI investing today spans everything from chip fab power contracts to healthcare SaaS revenue cycle management.
Beyond Anthropic, GIC's portfolio includes Databricks (US.4 billion revenue run-rate, growing 65 per cent year-on-year), Ramp (the corporate expense platform backed eight times, now valued at US4 billion), athenahealth (where its AI medical coding helped practices recover 30 per cent more revenue on denied claims), and Eli Lilly, which operates its own NVIDIA supercomputer. The fund is also deploying US0 billion into hedge funds over the next three years and expanding into digital infrastructure, utilities, and venture capital across the AI value chain. GIC's Enterprise AI Survey of nearly 200 North American companies found that more than half ranked AI among their top three priorities, with one in five putting it first.
Why it matters for Singapore: GIC manages over S00 billion in assets, and its growing conviction in AI as a long-term thematic bet signals something important for Singapore's broader positioning. The fund is not just an equity investor but also an internal user of Claude, GPT, and its own enterprise AI platform — making it simultaneously a backer and a consumer of the technology it helps fund. As GIC increases its allocation to digital infrastructure, energy resilience, and AI-native companies, the spillover effects for Singapore's tech ecosystem — from data centre demand to talent development — will be material. GIC's 20-year annualised real return of 3.4 per cent, the lowest in six years, also underscores why the fund is leaning harder into high-conviction bets like AI rather than broad market exposure.


