Anthropic is reportedly telling investors the truth about the AI backlash

The document that’s supposed to sell investors on Anthropic will reportedly include a section about how much the public hates what companies like Anthropic do.

That’s according to CNBC, which cites anonymous “people familiar with [the] matter.” The outlet first reported on a series of meetings in San Francisco where Anthropic Chief Financial Officer Krishna Rao has been taking questions from would-be investors ahead of the public release of the company’s IPO prospectus.

Per that report, public antipathy toward Anthropic’s product and the data centers that power it will show up in the prospectus as a “key risk factor.”

What a risk factor actually means here

The prospectus is the most important formal document disclosing what investors need to know before they put money into Anthropic. Listing public outcry there isn’t a marketing choice. It’s an admission.

And the outcry in question isn’t a comment section full of grumbling. Politicians on both sides of the aisle increasingly campaign against AI during elections.

Depending on how deep the prospectus goes, it might have to acknowledge disruptive protests against AI and data centers, reportedly including multiple incidents involving gunshots, and at least one molotov cocktail.

Compare it to how SpaceX handled the same question

SpaceX’s IPO earlier this year didn’t focus on public sentiment. That’s notable, because AI was arguably the main source of future revenue cited in the SpaceX prospectus.

The document did mention a reliance on natural gas. But it put that in the section on regulatory considerations, not in anything resembling a section on public outcry over pollution.

It also flagged Grok. Specifically, that the chatbot’s brashness and its ability to create “potential nonconsensual or exploitative imagery” might trigger “reputational damage” or “user or advertiser backlash.”

Read that carefully and you’ll notice what’s missing. Those are risks about a product behaving badly and advertisers walking away. Philosophical opposition to AI itself doesn’t appear to be in there at all.

The difference is who the enemy is

One filing treats backlash as a customer-relations problem. The other, if the reporting holds, treats it as a structural fact about the business.

That’s a meaningfully harder thing to write down in a document lawyers comb through line by line, because you can’t fix it with a content policy update or a better moderation model.

We reached out to Anthropic to confirm that investor meetings were taking place in San Francisco, and to ask for comment on the claim about the company’s IPO prospectus. We did not immediately hear back.

If you want to know how seriously to take any of this, watch for one thing when the prospectus goes public: whether the data center language sits in the regulatory section, where it’s a compliance cost, or in the risk factors, where it’s a threat to the business.