Anthropic Is Going Public. Does the Safety-First AI Lab Actually Score Safer?

Anthropic Is Going Public. Does the Safety-First AI Lab Actually Score Safer?

Anthropic filed a confidential S-1 with the SEC on June 1, 2026, weeks after closing a $65 billion Series H that valued the company at roughly $965 billion. The filing was reported by CNBC and Fortune, and confirmed by the company itself. With a revenue run-rate near $47 billion, a debut above $1 trillion is widely treated as the base case.

The pitch to public investors will lean heavily on one word: safety. Anthropic has built its brand on being the responsible AI lab, the one founded by researchers who left OpenAI over exactly these concerns.

That is a testable claim. Mashinii scores companies on documented evidence from court filings, regulatory actions, and investigative reporting rather than self-reported sustainability decks. So we checked: when you strip away the branding, does Anthropic actually score safer than its rivals?

The short answer is yes, with two caveats investors should price in before the roadshow.

The 2026 AI IPO wave

Anthropic is not arriving alone. OpenAI confirmed its own IPO filing on June 8, and SpaceX debuts on June 12. Within a fortnight, public markets are absorbing the largest cluster of private-market giants since the dot-com era.

That timing matters for anyone searching "anthropic vs openai stock." For the first time, retail and institutional investors can choose between the two leading AI labs directly. Financial comparisons will dominate the coverage, as TechCrunch's filing report already shows.

But the financials are converging. The ethics records are not. If you intend to hold one of these names for a decade, the conduct evidence is the differentiator the prospectuses will not give you.

What the independent evidence shows

Mashinii scores companies across seven values dimensions. Scores are built from adversarial sources: litigation records, regulatory enforcement, NGO investigations, and investigative journalism. A score of 0 means no significant documented evidence either way. Positive scores require documented positive conduct; negative scores require documented harm.

Here is how the three major US AI labs compare.

The headline number is the Safe & Smart Tech row. Anthropic scores +40. OpenAI scores 0. xAI scores -10.

A +40 is unusual in our coverage. Most companies cluster at zero on this dimension because positive evidence of responsible technology development is rare and self-published safety frameworks do not count. Anthropic's score reflects documented conduct, not marketing. The safety-first branding, in other words, survives contact with independent evidence.

Where the prospectus will be quiet

The caveats are real, and they sit in two rows of that table.

First, Honest & Fair Business: Anthropic scores -20, worse than OpenAI's -10 and level with xAI. Documented evidence on business conduct cuts against the clean-hands narrative. A company can be careful with its models and still attract credible claims about how it competes and contracts.

Second, Planet-Friendly Business: -20, driven by Anthropic's data-center footprint. Training and serving frontier models consumes power and water at industrial scale, and the evidence base reflects that. Anthropic still scores meaningfully better here than OpenAI (-40) and xAI (-50), but better-than-worse is not the same as good.

Investors screening on environmental grounds should read all three scores together. The AI buildout has an energy problem across the board; Anthropic is the least exposed of the three, not exempt.

Anthropic vs OpenAI: the stock question

If both labs list this year, the conduct comparison becomes a portfolio decision rather than an academic one.

On our evidence, Anthropic leads or ties OpenAI on every dimension except Honest & Fair Business, where it trails by 10 points. OpenAI carries negative scores across six of seven dimensions, including a -20 on No War No Weapons where Anthropic sits at zero.

That pattern is worth dwelling on. OpenAI's negatives are broad but mostly shallow. Anthropic's record is narrow: clean across most dimensions, strongly positive on safety, with two concentrated weak spots. For a values-screened portfolio, those are different risk shapes, and the difference will not appear in either S-1.

We unpack the head-to-head record in more depth in OpenAI vs Anthropic: the ethics evidence.

What this means for IPO investors

Three practical takeaways.

The safety premium has evidence behind it. Anthropic will ask public investors to pay up for responsibility. Unusually for an IPO narrative, the independent record supports the core claim. The +40 Safe & Smart Tech score is the strongest in our AI coverage.

The weak spots are specific and checkable. Business-conduct and environmental evidence are where Anthropic's record diverges from its brand. Both are documented on the Anthropic score page, with sources, so you can weigh them yourself rather than take our aggregate on faith.

Relative scoring matters in a wave. With OpenAI filing the same month and xAI inevitably discussed alongside both, the question is rarely "is Anthropic ethical" in isolation. It is which exposure you take. On the current evidence, Anthropic is the cleanest of the three by a clear margin, and the only one with a documented positive.

None of this is a buy recommendation. Valuation, lock-ups, and governance structure will do most of the work in determining returns. But conduct risk compounds quietly, and it is the one input you can audit before the first trade prints.

Check your own exposure

If you already hold AI through funds or index products, you may own all three of these companies at once without knowing your blended exposure. Audit your portfolio to see how your actual holdings score across all seven values, or search any company to read the evidence behind its scores before the IPO window opens.