Anthropic Did Two Things in One Day: Shipped Claude Opus 4.8, Then Passed OpenAI With a $965 Billion Valuation
On May 28, Anthropic announced a $65 billion Series H at a $965 billion post-money valuation, making it the world's most valuable AI startup. The same day it released Claude Opus 4.8, built around honesty and parallel agents, at the same price as before. Here are the funding numbers, the model changes, and the pushback.
On Thursday, May 28, Anthropic dropped two pieces of news in a single day.
First: a $65 billion Series H round at a $965 billion post-money valuation. Second: Claude Opus 4.8, a flagship model upgrade at the same price as its predecessor.
Announcing both on the same day isn’t a coincidence. The funding says “this is what we’re worth.” The model is the argument for why.
The funding, taken apart
TechCrunch covered the key facts: $65 billion is one of the largest private rounds ever, the round was led by Altimeter, Dragoneer, Greenoaks, and Sequoia, and it’s likely Anthropic’s last private fundraise before going public. Crunchbase added the comparison that this round more than doubled Anthropic’s post-money valuation from the previous one.
What does $965 billion mean in context? CNBC’s framing: Anthropic has formally passed OpenAI as the most valuable AI company in Silicon Valley. Reports differ on OpenAI’s prior valuation — Sherwood says $850 billion, the New York Times says $730 billion — the gap comes down to methodology and timing. Either way, the ranking flipped.
Worth remembering: two years ago Anthropic was valued at roughly half of OpenAI and had always been the second name in the sentence. What reversed the order wasn’t chatbots. It was the enterprise side — Claude’s revenue growth in coding and business workflows is the story investors bought. The use of funds is equally plain: Anthropic says the money mostly goes to computing capacity to keep up with Claude demand.
What changed in Opus 4.8
Claude Opus 4.8, released the same day, is a regular flagship upgrade with better coding and knowledge work, at unchanged pricing. The specifics:
- Honesty is the headline feature. The Next Web’s writeup notes that in internal evals the model admits mistakes it previously wouldn’t, and misses fewer of its own flaws when writing code
- Dynamic Workflows and agent swarms: Memeburn reports a new dynamic workflow system that can orchestrate up to 1,000 parallel subagents, plus a fast mode about 3x cheaper
- Mythos is still coming: the next-generation model Anthropic previewed in April will get a wide release “in weeks,” according to Fortune. Opus 4.8 reads like the opening act
We use Claude heavily in our own workflow, and we’ll publish a separate hands-on review of Opus 4.8 after running it for a while. First impression: in long agentic tasks it’s more willing to stop midway and say “this approach won’t work” instead of grinding out a wrong solution to the end. If that behavior holds up, it’s a real improvement for anyone who codes with it.
More detail from the announcements
Fortune’s report adds several numbers that help explain the valuation and the model:
- Valuation path: Anthropic was valued at $380 billion after its Series G in February, so the valuation grew about 2.5 times in roughly three months.
- Revenue: Anthropic said annualised run-rate revenue had passed $47 billion by May, up from about $10 billion at the end of 2025, driven largely by enterprise sales and Claude Code.
- Round structure: besides the four lead investors, Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ and XN co-led. Of the $65 billion, $15 billion was previously committed money from cloud providers, including $5 billion from Amazon.
- OpenAI comparison: Fortune puts OpenAI’s latest valuation at $852 billion, set in its March round.
On the model side, Anthropic says Opus 4.8 shows “substantially lower” rates of misaligned behaviour, such as deception and cooperation with misuse, than Opus 4.7, and is about four times less likely to let flaws in its own code pass without comment. List pricing stays at $5 per million input tokens and $25 per million output tokens. The faster “fast mode” configuration is now three times cheaper than it was for 4.7, which matters mainly to high-volume users.
What this means if you use Claude
For developers and teams, the practical points are modest but real. Same pricing means no budget change to switch from 4.7, and cheaper fast mode lowers the cost of latency-sensitive workloads. The honesty changes are most useful in long coding sessions, where a model that flags its own mistakes saves review time. Teams that rely on Claude for routine business tasks should test their own prompts before switching, because independent testers reported more refusals in some cases.
The pushback
Not all the feedback is applause. The same Fortune piece collected user complaints that Opus 4.8 is “too cautious” — too much hedging where a direct answer would do. The dial between honesty and decisiveness apparently moved one notch toward honesty this generation.
The valuation drew sharper questions. Investor’s Business Daily raised the point that matters most right now: with both OpenAI and Anthropic preparing IPOs, private-round pricing and public-market pricing are different animals, and revenue recognition deserves a close read. $965 billion is what investors bid. It’s not a market verdict yet.
Evaluators were also split. Fortune cites Andon Labs, which found Opus 4.8 did worse than 4.7 on some business-task and adversarial tests and described it as more cautious. Anthropic’s own and Cursor’s benchmarks showed gains on agentic coding and long-horizon work. Mixed results like these are common after alignment-focused updates, and they usually narrow as the vendor tunes behaviour.
What’s next in the ranking contest
| Item | Status |
|---|---|
| $65B Series H | Closed (May 28) |
| Opus 4.8 | Live, same pricing |
| Mythos wide release | ”In weeks,” per Anthropic |
| IPO | Not announced; TechCrunch calls this round “likely the last private one” |
Two things to watch from here. One is what Mythos actually delivers when it opens up — Opus 4.8 is an incremental release, and Anthropic is clearly holding its big card. The other is IPO timing: once either Anthropic or OpenAI lists, AI lab valuations face the public market for the first time, and “who’s worth more” finally gets an answer that isn’t defined by a funding round.
Reader Dispatches
No reader responses yet. Share your technical perspective below.
Related Posts
Claude Fable 5 Is Out: Anthropic Opens Up the Mythos That 'Spooked the Government,' With a Gate in Front
On June 9, Anthropic released Claude Fable 5, the first Mythos-class model available to the public. It shares its underlying model with the restricted Mythos 5; the difference is Fable 5's classifier gate, which reroutes high-risk requests in cybersecurity and bio-chem to Opus 4.8. Pricing is double Opus 4.8. Here's the capability, the limits, and the controversy.
Anthropic Ships Opus 4.7 and a Design Tool in the Same Week — The Model Isn't the Story
The Information broke on April 15, 2026 that Anthropic would launch Claude Opus 4.7 alongside a new AI design tool this same week. The 4.6-to-4.7 bump is incremental. The design tool is Anthropic's first serious move from 'model API vendor' toward a full-stack AI studio — and that's the shift worth watching.
Microsoft Build 2026: Seven In-House MAI Models at Once, Because Microsoft Is Done Depending on OpenAI
At Build 2026 on June 2, Microsoft launched seven in-house MAI models covering coding, reasoning, image, and voice, plus Scout autonomous agents and the Nvidia RTX Spark dev kit, and quietly dropped the Copilot+ PC hardware requirements. The subtext of the whole event: reduce the dependence on OpenAI.