Anthropic has closed a $65 billion Series H at a $965 billion post-money valuation, likely its last private raise before going public. The round was co-led by Altimeter, Dragoneer, Greenoaks, Sequoia, Capital Group, Coatue, and D1, with institutional money from Baillie Gifford, Blackstone, Brookfield, DST Global, and Fidelity. Strategic infrastructure partners Samsung, SK Hynix, and Micron also took allocations, and $15 billion of the round carries over previously committed hyperscaler capital, including the $5 billion from Amazon announced in April alongside a $100 billion cloud-spend pledge.
The financial story is loud, but the operating story is louder. Anthropic's run-rate revenue crossed $47 billion earlier this month, and the Wall Street Journal reported the company expects a 130% revenue surge into its first operating profit. The article is explicit about the engine: growth has come "particularly among enterprise customers that rely on Claude Code."
The round landed the same day as Claude Opus 4.8, positioned on agentic tasks, advanced coding, and self-correction. For context on the arms race, OpenAI raised $122 billion in March at an $852 billion valuation. The numbers are a scoreboard. What sits underneath them is a textbook multi-motion migration, executed in public.
Buyers did not fund this round, but they paid for it. A $47 billion run rate built on enterprise Claude Code adoption tells you exactly how the world's most demanding organizations now buy AI: a developer swipes in, a team standardizes, and procurement formalizes the spend it can no longer ignore. The enterprise is no longer waiting for a sales motion to reach it. It is self-selecting through the product and pulling the contract upward. That a hyperscaler committed $5 billion plus a $100 billion cloud pledge tells you procurement is happy to route the spend through cloud commitments it already holds. The buyer behavior here is not "evaluate the vendor." It is "expand the deployment we already trust."
The mistake is reading this as a capital story you cannot replicate. The replicable part is the sequence. Anthropic did not pick one motion and scale it. It ran a developer wedge, layered consumer-grade trust, and converted both into Fortune 500 procurement through a hyperscaler. The $65 billion is the receipt for getting the migration right, not the cause of it. If you are a Series B founder reading this and your takeaway is "raise more," you have misread it. The takeaway is that your wedge motion and your enterprise motion need to be the same growth curve, separated by stage, not two disconnected bets.
[52]. Land-and-Expand Seeding · Cross-Quadrant AmplifiersClaude Code is the clearest land-and-expand engine in AI right now. Individual developers adopt it, usage spreads across engineering teams, and the enterprise SKU formalizes what was already happening bottom-up. The article naming Claude Code as the specific driver of enterprise growth is the motion working exactly as designed: the first seat is the foothold, not the goal.
[2]. API-first / docs-as-funnel · The Wedge (Q1)The Claude API remains the entry surface that lets technical buyers integrate before anyone talks to sales. This is the wedge that splits open accounts at $0 ACV and self-qualifies the funnel into commit-level revenue. It fits exactly the developer-and-engineering-leader buyer who needs product evidence before institutional cover.
[38]. Hyperscaler Co-sell · The Cathedral (Q4)The Amazon relationship, $5 billion in capital plus a $100 billion cloud-spend pledge, is the Cathedral motion made literal. F500 buyers route Claude spend through pre-committed AWS budgets, and the hyperscaler field team becomes a distribution channel. This is what migration looks like when the developer wedge graduates into procurement-grade enterprise sales.
[6]. Eval / benchmark as marketing · The Wedge (Q1)Shipping Opus 4.8 on the same day as the raise, positioned on agentic and coding benchmarks, is not a coincidence of timing. The benchmark is the trust signal that keeps the developer wedge sharp while the enterprise motion matures above it. The model release is the marketing artifact that re-arms the bottom of the funnel.
The combination, a Wedge entry, a cross-quadrant land-and-expand amplifier, and a Cathedral co-sell running simultaneously, is the entire multi-motion thesis in one company: you do not pick a motion, you sequence three and let each fund the next.
Across the 700+ enterprise AI transformations and 88 insurance AI vendor profiles we have mapped, the same pattern keeps showing up. Three layers worth naming:
Layer one. The buyer enters through the cheapest, lowest-friction surface available, almost always a developer or a single team using a product-led wedge. Nobody in procurement approved Claude Code's first deployment. It arrived as individual usage and became a line item only after the organization could no longer pretend it was not depending on it. The land happened before the deal did.
Layer two. Trust gets manufactured in parallel, not after. Anthropic's safety positioning, model benchmarks, and consumer Claude brand built institutional cover at the same time the developer wedge was spreading. By the time the enterprise procurement conversation started, the CISO and the economic buyer already trusted the name. The vendors who lose run these sequentially, building trust only once sales asks for it. By then the cycle is six months longer.
Layer three. The migration to the Cathedral is funded by the wedge, not separate from it. The hyperscaler co-sell, the $100 billion cloud pledge, the procurement-grade contracts, all of it sits on top of a usage curve that was already compounding. Vendors who try to jump straight to the Cathedral, hiring a CRO and chasing F500 logos before the wedge produces self-serve pull, build the institutional motion on sand. The Cathedral only holds when the Wedge is paying for the foundation.
This round is the most expensive proof point yet that the multi-motion migration, Wedge to Beacon to Cathedral, is not a theory. It is the operating model that produces a $47 billion run rate and a path to operating profit. The capital is downstream of the GTM discipline, not a substitute for it.
Three questions every founder and CRO should sit with this week. First: is your wedge motion and your enterprise motion the same growth curve, or two disconnected bets you are funding in parallel? Second: are you building institutional trust now, while the wedge compounds, or are you waiting for a security review to force it? Third: when your developer adoption hits critical mass inside an account, do you have the enterprise SKU, the admin controls, and the co-sell relationship ready, or will expansion require re-selling from zero?
By Q4 2026, every enterprise AI vendor at Series B and above will be measured by whether their bottom-up product adoption and their top-down enterprise contract sit on a single revenue curve. The ones running them as separate motions will stall at the ARR ceiling of whichever one they started with.
What's your experience with land-and-expand into hyperscaler co-sell? Drop a note or reach out directly.
.svg)
