Anthropic closes $30B Series G at $380B valuation as AI capital concentrates

What happened

Anthropic announced on February 12, 2026 that it had closed a $30 billion Series G led by GIC and Coatue, valuing the company at $380 billion post-money ¹. Co-leads included D. E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQ, and MGX ². Crunchbase ranked the deal as the second-largest venture financing in history, behind only OpenAI's $40 billion round in 2025 ³. Anthropic disclosed that its run-rate revenue had reached $14 billion, growing more than tenfold in each of the past three years, with Claude Code alone above a $2.5 billion run rate ⁴.

Why it matters

The round confirms a structural shift: frontier-model capital is no longer being spread across a long tail of labs. February 2026 saw $189 billion in global venture funding, with AI absorbing $171 billion, or roughly 90 percent of the month's total ⁵. For founders outside the US-China duopoly, this concentration raises the cost of competing on raw model capability and pushes the centre of gravity toward applied layers, vertical agents, and regulated workflows where distribution and data still matter more than parameter count.

Anthropic's enterprise positioning, anchored on Claude Code and on regulated-industry contracts, signals where the next round of margin will sit. Capital is now flowing into companies that can show paying enterprise revenue, not laboratory benchmarks. International entrepreneurs reading the round should note the implicit message to limited partners: the venture asset class is being re-priced around AI infrastructure, and non-AI rounds will face tighter scrutiny and lower multiples for the rest of the year.

Historical parallel

The closest analogue is the late-1990s telecom-equipment buildout, where Lucent, Nortel, and Cisco absorbed an outsized share of equity and debt capital between 1998 and 2000 on the basis of forward-looking infrastructure demand ⁶. That cycle ended with severe overcapacity and a multi-year drawdown for survivors. A more recent comparison is Uber's 2016 $3.5 billion raise from Saudi Arabia's PIF, which similarly signalled that one or two ride-hailing platforms would absorb the bulk of mobility venture capital and squeeze regional players into acquisition or exit ⁷. In both cases, scale begat more scale until either policy or unit economics intervened.

What to watch next

• Whether Anthropic's revenue growth holds at greater than 5x annualised through Q3 2026, or compresses as enterprise pricing pressure rises.
• The trajectory of Claude Code versus GitHub Copilot and Cursor in enterprise procurement cycles.
• Any move by GIC or MGX to seek board observer rights, which would signal a shift from purely financial to strategic sovereign capital.
• US export-control updates affecting Anthropic's chip supply via Amazon's Trainium and Nvidia's Blackwell families.
• Whether secondary-market valuations (reportedly approaching $1 trillion implied) compress back toward primary marks if AI multiples reset ⁸.

Implications for global entrepreneurs

Cross-border founders building on top of Claude or comparable frontier models should treat model-layer pricing as a strategic variable, not a fixed cost. Anthropic's enterprise focus and pricing discipline suggest API costs will hold or rise for premium tiers even as commodity models cheapen. Founders in Toronto, London, Singapore, and Shenzhen should design their products to be model-portable from day one, both for cost arbitrage and for regulatory resilience as data-residency rules tighten.

For Canadian and UK founders specifically, the round is a reminder that domestic capital pools remain shallow relative to the US. A Series A in Toronto rarely exceeds C$20 million; a comparable AI-applications Series A in San Francisco now routinely clears US$50 million. Founders building for global markets should consider US incorporation or a dual-headquartered structure earlier than they would have two years ago, while preserving R&D presence in jurisdictions with favourable tax credits such as Canada's SR&ED programme.

Risks and caveats

The headline valuation rests on private-market signalling, not on a public-equity test. If a major frontier lab posts a disappointing model release, or if a regulator imposes structural separation between model providers and cloud distribution, the entire sub-sector could re-rate sharply. Anthropic's revenue figures are self-reported and not audited to public-company standards. The 90-percent AI share of February venture funding may also be partially an artefact of a small number of mega-rounds; median early-stage rounds have not seen the same expansion.

---

Sources