Anthropic to Stand Beside Pope at Artificial Intelligence Encyclical Unveiling
When Pope Leo XIV unveils his first encyclical on artificial intelligence on Monday, the most consequential question will not be what the document says. It will be whether Anthropic helped write it. Cofounder Chris Olah will stand beside the pontiff at the unveiling — a visible marker of a year-long Anthropic program of ethics dialogues with theologians, philosophers, and religious leaders, including sessions at its San Francisco headquarters with the programmers building the models. Brian Patrick Green, director of technology ethics at Santa Clara University, said Anthropic's engagement with the Vatican predates Pope Leo XIV's election and reflects years of relationship-building. "The Vatican has been cultivating relationships with the tech community for about 10 years," Green told Religion News Service.
The document, titled Magnifica Humanitas, is expected to address the protection of human dignity in the age of AI, the use of AI in warfare, and challenges to workers' rights. Whether it names Anthropic as a partner in building more ethical AI, or names it as one of the companies that needs to be held accountable, will be the first real signal of whether the Vatican's engagement changed anything — or whether it was always going to end the way the AI labs needed it to.
What is knowable is that Anthropic is not simply a technology company that stumbled into ethics. It has a chief legal officer who was formerly at the State Department. It has an in-house philosopher. It has donated $20 million to political candidates who support AI regulation — a fact that would be unremarkable from a company lobbying for favorable rules, but reads differently from a company that presents itself as the industry's conscience. Sources say David Sacks has publicly accused Anthropic of running what he calls a sophisticated regulatory capture strategy based on fear.
That accusation landed in a specific context. Last Thursday, Sacks called Donald Trump directly and killed an executive order that would have created a voluntary pre-release review system for frontier AI models. The order, planned for a Thursday afternoon signing ceremony with CEOs from OpenAI, Anthropic, and Meta, was postponed indefinitely. Trump told reporters he didn't like certain aspects of it. "We're leading China," he said. According to a senior White House official, Sacks worried that the voluntary framework could one day become mandatory. Mark Zuckerberg, Elon Musk, and Sacks all called the White House between Wednesday night and Thursday morning. A photograph op with tech CEOs became a cancellation.
The executive order Sacks killed would have created a voluntary system for pre-release review of frontier AI models. It was not mandatory. It was not even particularly burdensome — companies could have submitted their models and continued operating normally. What it would have done is establish a precedent: that frontier AI models get reviewed before release, and that the federal government has a role in determining what "safe" looks like. Anthropic opposed even that.
The revenue picture shows why the institutional positioning matters. Anthropic announced its run-rate revenue had surpassed $30 billion, up from $9 billion at the end of 2025. The company's $1 million-plus customer count had doubled to more than 1,000 in less than two months. CEO Dario Amodei told a developer conference the company had planned for tenfold annual growth and saw eightyfold instead. What Anthropic has built in the past eighteen months is a case study in the relationship between scale and political leverage. The company that did not exist seven years ago is now large enough that its preferences carry weight in presidential scheduling. The $30 billion run-rate, then reportedly $44 billion in annual recurring revenue by May, represents a customer base of enterprises, governments, and financial institutions that have made Claude part of their operating infrastructure. You do not regulate infrastructure that your own regulators depend on without a different kind of fight.
The IPO race is the financial layer of the same gambit. OpenAI and Anthropic are both preparing public offerings, with OpenAI targeting Q4 2026. Anthropic's reported $900 billion valuation in its latest funding round, and its Series G of approximately $50 billion, suggest institutional investors are pricing in the same theory: the company that owns the infrastructure of the AI economy will eventually own the leverage in the AI governance debate.
Compute is the third layer. Anthropic's deal with Google and Broadcom for multiple gigawatts of next-generation TPU capacity — coming online starting in 2027 — and its separate arrangement with SpaceX for 300 megawatts from the Colossus 1 data center in Memphis are not just infrastructure investments. They are evidence of a company that has identified its constraint and is buying everything available before competitors can.
Amodei told the developer conference that the company's internal pull requests have inflected upward for the first time because of work done by Claude Code on its own codebase. The company that sells the tool is now being built by the tool. That feedback loop is not in the encyclical. It is not in any executive order. It is the actual story of what changed.
The executive order is still postponed. The IPO paperwork is still being prepared. The compute deals are still being signed. None of this requires Monday's event to say anything in particular.
It helps if it does.