Transcript
SPEAKER_1: This is the Weekly Download, number forty-nine. We're covering Anthropic's massive new valuation, OpenAI's platform consolidation, AI's cultural backlash, and the growing energy crisis tied to data centers. Let's start with the biggest headline: Anthropic just closed a thirty billion dollar raise at a nine hundred billion dollar valuation. SPEAKER_2: And that number is staggering when you put it in context. Just six months ago, Anthropic was raising at around a hundred and fifty billion. So this is roughly a six times jump in valuation in half a year. SPEAKER_1: Who led the round? SPEAKER_2: Sequoia, Dragoneer, Greenoaks, and Altimeter co-led it. And investor documents show the revenue run rate is on track to hit fifty billion dollars by the end of June. That would make Anthropic one of the fastest-growing enterprise software companies in history. SPEAKER_1: Fifty billion. How does that compare to the broader AI startup landscape? SPEAKER_2: The Information reported that thirty-four leading AI startups are generating roughly eighty billion in annualized revenue combined. Anthropic and OpenAI together capture eighty-nine percent of that. So two companies are dominating the entire space. SPEAKER_1: And there's a signal from Ramp that caught my attention. What's happening there? SPEAKER_2: Ramp told investors that more of its customers now use Anthropic than OpenAI. That's a first. And it matters because Ramp's customer base skews toward fast-growing startups and mid-market companies, buyers who switch tools based on performance, not habit. SPEAKER_1: So it's a performance signal, not just a brand preference. What else is Anthropic doing on the business side? SPEAKER_2: Two things. First, Anthropic and the Gates Foundation announced a two hundred million dollar partnership to deploy AI in health and education. The Gates Foundation had signed a smaller fifty million deal with OpenAI back in January, so this is a bigger commitment. SPEAKER_1: And the second thing? SPEAKER_2: Anthropic launched Claude for Small Business, adding automated bookkeeping, business insights, and ad campaign tools. So they're pushing downmarket while the enterprise flywheel keeps accelerating at the same time. SPEAKER_1: Their CFO, Krishna Rao, gave an interesting interview. What was his framing? SPEAKER_2: He described a cone of uncertainty in AI, where returns to frontier intelligence remain steep but nobody can predict when they flatten. He said Anthropic intentionally raises less money than is available and stays conservative on revenue projections. At a nine hundred billion valuation, conservative is doing a lot of work. SPEAKER_1: Right. And beyond the business side, Anthropic is starting to look like something more than a company. What's happening on that front? SPEAKER_2: The Financial Stability Board, which is a central banking body, directly asked Anthropic to brief them on global financial system vulnerabilities discovered by Mythos. That's a central bank asking an AI company for a security briefing. SPEAKER_1: That's remarkable. And there's a Vatican connection too? SPEAKER_2: Yes. The Vatican announced that Anthropic co-founder Christopher Olah will join Pope Leo on May twenty-fifth to launch the pope's first encyclical, setting out his views on the AI age. SPEAKER_1: And on policy? SPEAKER_2: Anthropic published a policy paper urging the US and allies to enforce export controls, curb distillation attacks, and export US AI to hold the frontier lead over China through twenty twenty-eight. So in eighteen months, they've gone from research lab to briefing the FSB, advising the Pope, and writing geopolitical policy papers. SPEAKER_1: At what point does an AI company stop being a startup and become something closer to a sovereign capability? That's the real question. Let's shift to OpenAI. What happened there this week? SPEAKER_2: OpenAI spent the week consolidating product, expanding into finance, and watching threats emerge from its own alumni and its biggest partner. Greg Brockman is taking over product strategy, folding ChatGPT, Codex, and the developer-facing API into a single core product team. SPEAKER_1: So the message is that OpenAI is no longer a collection of research projects with a chat interface. It's one platform now. SPEAKER_2: Exactly. And the same week, OpenAI debuted personal finance tools for ChatGPT Pro users through a Plaid integration covering twelve thousand plus financial institutions. Users can connect bank accounts, analyze spending, and get AI-powered financial guidance inside the same app they use for coding and search. SPEAKER_1: So they're expanding the platform horizontally at the same time they're consolidating internally. There's also a legal threat brewing with Apple. What's that about? SPEAKER_2: OpenAI may sue Apple over Siri revenue. That's a significant escalation given how intertwined those two companies have been. SPEAKER_1: And then there's the alumni threat. Mira Murati's new model? SPEAKER_2: Yes. Mira Murati, OpenAI's former CTO, just shipped a model called TML-Interaction-Small, and it crushed GPT-Realtime on every benchmark. So OpenAI is now competing with its own former leadership. SPEAKER_1: That's a striking development. Now let's talk about the energy story, because the infrastructure costs of all this AI are becoming impossible to ignore. SPEAKER_2: Right. NextEra is buying Dominion for sixty-seven billion dollars. That's the largest utility merger in US history, and it's driven entirely by AI power demand. SPEAKER_1: And grid prices are reflecting that pressure too? SPEAKER_2: PJM grid prices jumped seventy-six percent year over year. And the backlash is starting to show up locally. A Texas county passed the first county-wide data center ban. And a Gallup poll says seventy-one percent of Americans oppose data centers in their local communities. SPEAKER_1: So there's a political dimension emerging as well. And that connects to the Bannon story? SPEAKER_2: Yes. Steve Bannon and sixty-plus Trump allies signed a Humans First-led letter urging mandatory government testing of AI models before release. They're calling it a kill switch proposal. SPEAKER_1: So to recap this week: Anthropic raised thirty billion at a nine hundred billion valuation with a fifty billion revenue run rate on the horizon, and it's now briefing central banks and advising the Vatican. OpenAI consolidated its products under Brockman, launched finance tools, and faces threats from both Apple and its own alumni. The energy bill for AI is coming due, with a record utility merger, spiking grid prices, a county-wide data center ban, and growing public opposition. And political pressure for AI oversight is building from an unexpected coalition. Next time, we'll go deeper into the cultural and regulatory forces shaping where all of this is headed.