Anthropic Overtakes OpenAI: $30B Revenue, Vatican Ethics, Liability Fight

Welcome to The Aigency Works Dispatch, your backstage pass to what's fresh, fascinating, and flying off the innovation shelves in the world of AI. From breakthrough tools to bold new use cases, we're serving up bite-sized updates to keep you (and your Aigents) ahead of the curve. Let's dive into what made waves this week

AI + Religion?: How Anthropic has called upon the Vatican for counsel

Anthropic has done something no other AI lab has thought, or dared, to do: ring up a Catholic priest. Father Brendan McGuire, Silicon Valley's go-to chaplain, is now helping shape the ethical rules that govern Claude's behaviour. Rather than sticking to the usual circuit of policy academics and philosophy professors, Anthropic went looking for a tradition that's been seriously stress-testing moral questions for actual centuries. What does harm by omission look like? When does following orders become cowardice? Theology has notes. It fits neatly into Anthropic's whole "we're the responsible ones" identity, and McGuire isn't a random pick either - he's spent years counselling tech workers and executives in the Valley. This is less "PR stunt" and more "someone at Anthropic genuinely did their homework."

Look, the jokes write themselves - we get it. But think about what most AI ethics actually looks like: slide decks, red-teaming sessions, policy memos. Bringing in a religious tradition forces genuinely different questions onto the table, the kind that don't get asked in a three-year-old startup. The real question isn't why Anthropic did this. It's why nobody else has.

Source: The Wall Street Journal — 'Anthropic Turns to a Priest to Help Figure Out If Its AI Has Good Values'

Who Ends up on the right side of history?

A proposed law - with OpenAI's fingerprints on it - would have shielded AI companies from liability if their systems caused mass casualties or over a billion dollars in damage. The logic: companies need legal breathing room to innovate without the threat of civilisation-ending lawsuits. Anthropic took one look at it and said no. That's a big deal, because the industry instinct is almost always to close ranks on anything that reduces regulatory heat. The counter-argument is simple: if you can't be sued when your AI causes catastrophic harm, your incentive to prevent it gets a lot weaker. Liability is blunt, but it works. Taking it off the table is essentially asking the industry to pinky-promise on existential risk.

The cynical read is obvious - Anthropic loves being the responsible adult in the room, and this is great branding. But it's also just... consistent with everything they've ever said. And there's a telling detail buried in the story: if you're genuinely confident your AI won't cause mass harm, why do you need immunity from that exact scenario in the first place?

Source: Reuters — 'Anthropic Opposes Bill That Would Shield AI Companies From Liability for Catastrophic Harm'

A Tipping of the Scales in the AI Race?

Anthropic is now reportedly sitting on a $30 billion annualised revenue run rate. A few months ago it was $9 billion. Yes, that's the same Anthropic with no ChatGPT, no Super Bowl ad, and no viral moment to speak of. OpenAI - the company that essentially invented this market - is reportedly at around $25 billion. Somehow the quieter, nerdier lab that went all-in on enterprise has run straight past it. The secret? Boring, reliable B2B adoption. Businesses embedding Claude into workflows, legal tools, and customer service stacks don't make headlines. They do, however, compound.

This completely reframes the AI race, which was always told as an OpenAI vs Google story with everyone else as supporting cast. Anthropic just skipped past one of the leads without anyone really noticing. The enterprise-first strategy - the one that felt cautious and unsexy while OpenAI was getting all the glory - turns out to have been the right call. Who knew that "actually works reliably in a business context" was a product strategy?

Source: Bloomberg — 'Anthropic Revenue Surges to $30 Billion Annualized Rate, Surpassing OpenAI'