Anthropic raises $65B, ships Opus 4.8, confirms Mythos rollout, and opens two offices — all in one week
On May 28, Anthropic closed a $65 billion Series H at a $965 billion post-money valuation — surpassing OpenAI — while simultaneously releasing Claude Opus 4.8 with improved honesty and a 2.5× faster mode. The company also confirmed a Mythos-class model public rollout within weeks, appointed a new Korea Representative Director ahead of the Seoul office opening, and opened a Milan office.
Anthropic spent the week of May 26–31, 2026 doing several years' worth of things at once: a historic fundraise, a new flagship model, a confirmed timeline for its most capable (and most closely held) model, and two international office announcements. This brief covers all four events.
$65 billion Series H at a $965 billion valuation
On May 28, Anthropic announced it had closed a $65 billion Series H round, valuing the company at $965 billion post-money — the highest valuation ever assigned to a private AI company and ahead of OpenAI's March 2026 valuation of $852 billion.1
The round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, with co-lead participation from Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ, and XN. Strategic investors include Baillie Gifford, Blackstone, Brookfield, Fidelity, General Catalyst, Jane Street, Lightspeed, and Temasek. Chip companies Micron, Samsung, and SK hynix joined as infrastructure partners.1
The $65 billion figure includes $15 billion in previously committed hyperscaler investments — $5 billion of that from Amazon, which remains Anthropic's primary cloud and training partner.1 In recent weeks, Anthropic also signed compute agreements with Amazon (up to five gigawatts of new capacity), Google and Broadcom (five gigawatts of next-generation TPU capacity), and SpaceX (GPU access in Colossus 1 and Colossus 2).
CFO Krishna Rao said the company's run-rate revenue crossed $47 billion earlier in May. That figure puts the valuation at roughly 20× annualized revenue — a multiple that signals investor belief that enterprise AI adoption is still early, not plateauing.
"Startups and Global 5000 companies alike are deploying Claude to handle complex workflows, and in doing so, Claude is learning how businesses actually operate: the context, the processes, the judgment." — Alfred Lin, Sequoia Capital
The Guardian noted that Anthropic's rise also stems from its positioning as the "safe" enterprise AI choice, a contrast it has cultivated through public safety commitments, interpretability research, and a visible refusal to remove guardrails for government surveillance or autonomous weapons — a stance currently being litigated with the Pentagon.2
Claude Opus 4.8: faster, cheaper fast mode, and better honesty
On the same day as the funding announcement, Anthropic released Claude Opus 4.8 as a direct upgrade to Opus 4.7.3
The headline improvement is honesty and self-awareness: Opus 4.8 is 4× less likely to let flaws in generated code pass without flagging them, and shows substantially lower rates of deception and cooperation with misuse — the company says its alignment profile is now close to Claude Mythos Preview.4
Key capability additions:
- Dynamic workflows (Claude Code) — Research preview for Enterprise, Team, and Max plans. A single session can now spawn hundreds of parallel sub-agents and handle codebase-scale migrations across hundreds of thousands of lines of code.
- Effort control — Available on all plans via claude.ai and Cowork. Users can dial effort up for deeper reasoning or down for faster responses at slower rate-limit consumption.
- Fast mode — 2.5× speed of prior models, 3× cheaper than the previous fast-mode pricing: $10/M input tokens, $50/M output tokens.
- Mid-task system updates — The Messages API now accepts system entries inside the message array, letting developers update instructions (permissions, token budgets, environment context) mid-run without breaking prompt cache.
Standard pricing is unchanged from Opus 4.7: $5/M input tokens, $25/M output tokens.4
Mythos-class models confirmed for public rollout
Separately, Anthropic confirmed that its Mythos-class models — currently held back due to cybersecurity concerns over vulnerabilities to financial systems and critical infrastructure — will be made available to all customers "within the next few weeks."5
The Mythos models are described as substantially more capable than Opus 4.8 in code reasoning and autonomous operation. The original April rollout was pulled after the company determined that safety guardrails were not sufficient to prevent misuse. A "Mythos-preview" variant appeared briefly in Claude Code before being taken offline.5
A limited number of organizations are already using Mythos Preview for cybersecurity research. The company has not specified whether the public version will be the same model or a more restricted variant.
Leadership and international expansion
Two organizational moves round out the week:
Korea office, new Representative Director. On May 26, Anthropic appointed KiYoung Choi as Representative Director of Korea, ahead of the Seoul office opening. Choi joins from Snowflake (Korea GM) and previously held country leadership at Google Cloud, Adobe, Autodesk, and Microsoft.6 South Korea's usage rate is 3.5× what its population size would predict, per Anthropic's Economic Index. Existing Korean enterprise customers include SK Telecom and Law&Company.
Milan office. Anthropic also announced the opening of its Milan office — its sixth in Europe — to support Italian enterprise, research, and developers.7
Ongoing: Pentagon litigation
The Guardian's funding piece flagged that Anthropic remains in litigation with the Pentagon after refusing to remove safeguards that would allow Claude to be used for mass domestic surveillance or lethal autonomous weapons.2 The company originally filed suit when the Pentagon labeled Anthropic a "supply chain threat," a designation that would have triggered restrictions on major cloud partners including Amazon and Google. That case is ongoing; no court dates have been publicly announced this week.
Anthropic Corporate Intelligence tracks product launches, funding, leadership changes, major customer announcements, and lawsuits. Next scheduled: weekly roundup Monday, June 2.
参考ソース
- 1Anthropic raises $65B in Series H funding at $965B post-money valuation
- 2Anthropic reaches valuation of $965bn, beating OpenAI to become world's most valuable AI firm
- 3Introducing Claude Opus 4.8
- 4Anthropic releases Claude Opus 4.8
- 5Anthropic confirms Claude Mythos-class models will roll out to the public
- 6Anthropic appoints KiYoung Choi as Representative Director of Korea ahead of Seoul office opening
- 7Anthropic opens Milan office to support Italian enterprise, research, and developers
このコンテンツについて、さらに観点や背景を補足しましょう。