Anthropic has done what many once thought improbable. The AI company founded by former OpenAI executives has raised $65 billion in a Series H round that values it at $965 billion post-money. That figure pushes it ahead of OpenAI, whose most recent valuation stood at $852 billion. The announcement landed on May 28 alongside the release of Claude Opus 4.8. Momentum no longer feels gradual. It feels decisive.
Revenue tells part of the story. Anthropic’s run-rate crossed $47 billion earlier in May. That compares with OpenAI’s $25 billion reported earlier this year. Growth like this explains why investors lined up. Altimeter Capital, Dragoneer, Greenoaks and Sequoia Capital led the round. Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ and XN co-led. The list of participants reads like a who’s who of technology and finance. It includes Amazon, Google, Broadcom, SpaceX and memory giants Micron, Samsung and SK hynix. Hyperscalers alone committed $15 billion of previously pledged capital.
The money will fund three priorities. Safety and interpretability research come first. Expansion of compute capacity follows to handle surging demand for Claude. Finally the company will scale products and partnerships that customers now treat as essential. “Claude is increasingly indispensable to our growing global community of customers, and we work tirelessly to make tools like Claude Code and Cowork more helpful, more powerful, and more adaptable to their needs,” said Krishna Rao, Anthropic’s chief financial officer, in the official announcement (Anthropic).
Brad Gerstner of Altimeter Capital captured the investor mood. “Claude’s latest advancements have driven large-scale adoption among the world’s most demanding organizations. This momentum positions Anthropic to lead the next phase of AI innovation and capture the enormous opportunity ahead.” Similar endorsements came from Marc Stad at Dragoneer, Neil Mehta at Greenoaks and Alfred Lin at Sequoia. Each emphasized culture, technical progress and commercial traction. Rarely do funding announcements feel this unanimous.
Yet capital alone does not win markets. Models do. Opus 4.8 arrives just 41 days after Opus 4.7. That pace is unusual for Anthropic. Sonnet and Haiku models have lingered longer in recent cycles. The haste may reflect feedback on 4.7. Some users called it disappointing. Early testers wanted sharper judgment and fewer hallucinations. Opus 4.8 attempts to deliver exactly that.
The new model improves across coding, agentic tasks and professional work. It shows greater consistency on long-running projects. Dynamic workflows let it plan and orchestrate up to 100 parallel subagents for complex jobs. Effort control gives users a dial. Higher effort yields deeper reasoning at greater cost and latency. Lower effort delivers faster, cheaper responses. Fast mode runs 2.5 times quicker and costs one-third as much as standard Opus usage. Pricing stays at $5 per million input tokens and $25 per million output for regular mode. Fast mode doubles those figures but still undercuts prior economics.
Benchmarks back the claims. Opus 4.8 leads on several agentic and professional evaluations. It became the first model to break 10 percent all-pass on the Legal Agent Benchmark. It scores 84 percent on Online-Mind2Web for browser-based tasks. Testers note it flags uncertainties more often and makes fewer unsupported claims. One early user from Bridgewater Associates said the biggest difference is “Opus 4.8’s tendency to proactively flag issues with the inputs and outputs of an analysis, something other models routinely missed and left to the users to catch” (TechCrunch).
Honesty metrics stand out. The model is roughly four times less likely to let unremarked flaws slip through in code. Alignment scores reached new highs on prosocial traits. Rates of misaligned behavior dropped below Opus 4.7 and sit close to those of the still-restricted Mythos preview. Anthropic has held Mythos back because of its exceptional ability to discover and exploit cybersecurity vulnerabilities. The company tested it with trusted partners and governments. “We’re making swift progress on developing these safeguards and expect to be able to bring Mythos-class models to all our customers in the coming weeks,” the announcement stated. Opus 4.8 remains substantially behind Mythos on cyber capabilities. It can still be tricked into claiming it is human in about 3 percent of tests.
These details matter to enterprise buyers. Companies now deploy Claude inside core operations. They run migrations spanning hundreds of thousands of lines of code. They automate legal analysis, financial workflows and data engineering. Reliability over long sessions counts more than raw benchmark scores. Opus 4.8 appears tuned for exactly that environment. It asks better questions, catches its own mistakes and evaluates plans before executing. That reduces the supervision burden that has frustrated many AI pilots.
The valuation jump from $380 billion in February to $965 billion now raises familiar questions. How sustainable is this pace? Compute costs remain enormous. Inference expenses for paid users could reach $7 billion this year and $16 billion next, according to some estimates. Anthropic’s partnerships with cloud providers and chipmakers aim to secure supply. Agreements for multiple gigawatts of capacity with Amazon, Google and SpaceX signal serious scale. Still, margins will face pressure until efficiency gains catch up.
OpenAI faces parallel challenges. Its GPT-5.5 release came in April. Revenue trails. Valuation sits lower. Competition has intensified on every front. Google’s integrations and Meta’s open-source efforts add further pressure. Yet Anthropic’s lead in enterprise adoption and its tighter focus on safety have won over boardrooms and investors alike. CNBC named it the top Disruptor for 2026, citing exactly this combination of growth and discipline (CNBC).
So what happens next? More models will follow. Mythos-class systems could reach general availability soon. Compute expansion will accelerate. Enterprise features will deepen. The funding gives Anthropic breathing room that few competitors enjoy. But breathing room does not guarantee victory. Execution on safety, cost and product velocity will decide the next leg of the race.
For now the numbers speak loudly. A company that did not exist five years ago sits at nearly a trillion-dollar valuation. Its latest model pushes the frontier forward while addressing practical complaints. Customers vote with usage. Investors vote with capital. Both appear to be choosing Anthropic. The AI industry has a new leader by market value. Whether that lead holds depends on what arrives after Opus 4.8. The calendar suggests the answer will come quickly.
WebProNews is an iEntry Publication