“Anthropic will be a $3 trillion company, SpaceX $2 trillion, and OpenAI $1 to $1.5 trillion by Q2 of next year.” — Dave McClure
Yesterday, Keith Teare and I debated the circularity of the AI economy. Today, two of Silicon Valley’s most experienced investors, Dave McClure and Aman Verjee, not only straighten out this supposed “circularity” but also burst the pessimism bubble that envelops so many conversations about AI.
Verjee is not only McClure’s partner at Practical Venture Capital, but also the author of the newly published A Brief History of Financial Bubbles. According to him, today’s AI-stoked economy is not an unusually large bubble. It may not even be a bubble, given that AI revenue — from Anthropic’s $70 billion to OpenAI’s $50 billion — is real. The irrational exuberance lives elsewhere — in companies “draping themselves in AI magic sauce” and in the “SaaSpocalypse” that is decimating software-as-a-service companies.
They are both bullish about our AI future. McClure predicts that by the first half of next year, Anthropic and OpenAI will have joined SpaceX as public companies. Together, these three AI darlings will be worth $7 trillion. That’s seven thousand billion reasons to be optimistic about 2027.
Five Takeaways
• Not a Bubble — a Repricing. Both partners reject the bubble call, on the numbers: Anthropic at roughly $70 billion in revenue on under two gigawatts of compute, OpenAI at $40–50 billion, SpaceX guiding to $100 billion with more than half from AI — real revenue, increasingly real profits. The froth is specific: companies “draping themselves in AI magic sauce” without the substance, and the SaaSpocalypse — cloud-software companies whose cash flows are suddenly perceived as far less durable as AI encroaches on design, legal, and medical verticals. Michael Burry’s warning gets Aman’s definitive treatment (“he’s called nine of the last two bubbles”), and the Aschenbrenner blowup was leverage — running four-x in volatile chip stocks — not AI: he kept his Anthropic position, is married to Dario’s chief of staff, and “will be just fine.”
• The $2 Trillion Filing. The week’s news, baked into the episode: Anthropic has filed to go public, with a very intentionally leaked $2 trillion valuation hinging on a $190–200 billion 2028 revenue forecast — which Dave suspects is conservative. Eight months ago, when these two last visited, the show was about Elon and Sam and Dario was the bit player; then came the weeks when decades happen: Anthropic’s bet on coding agents — reportedly inspired by watching Cursor — captured the revenue engine of the entire application layer. Aman’s sequencing: SpaceX is absorbing $75–85 billion of IPO capital, Anthropic goes next, and if both trade well, 2026 breaks every record for money raised — leaving 2027 for OpenAI at a $100 billion revenue guide. Google, he reminds us, went fourth after Yahoo, Lycos, and Excite: better to do it right than to do it first.
• The Fastest Pivot in Corporate History. Dave’s account of SpaceX’s transformation: the $250 billion xAI merger (a largely private transaction Elon approved with himself), the acquisition of Cursor that closed Friday, Colossus data centers scaling from two gigawatts toward ten, and compute deals renting capacity to Anthropic and Google — former competitors — all executed in roughly six months. The S-1, with unprecedented forward projections of $300 billion in annual revenue, mentions artificial intelligence over 1,100 times (“I used AI to count it,” Aman admits). The result is an economy Andrew calls incestuous: SpaceX’s valuation now rests on Anthropic’s progress. On Elon himself, Dave separates the art from the artist — terrific products, dubious politics — and on OpenAI: more board changes than Spinal Tap had drummers, a team still storming and norming, but Sam is savvy and the IPO lands by Q2 next year at $1–1.5 trillion.
• Circularity as Asset Class. The New York Times sees a vulnerability in tech giants funding their own customers; Aman, a former CFO at eBay and Sonos, sees asset-backed finance. His analogy: buying a Corvette with GMAC financing isn’t a conspiracy as long as the terms are commercially reasonable — and NVIDIA’s $500 billion backstop, syndicated with Goldman Sachs, Apollo, Brookfield, and KKR, brings third-party money that validates the asset. GPUs, he argues, are cars rather than smartphones: financeable over eight to ten years, not obsolete in three. The red flags to watch are rebates and self-dealing on non-commercial terms; the current evidence looks more like aircraft leasing than Enron. Dave’s deeper worry isn’t the AI economy at all — it’s the national deficit, whose interest payments are now the largest single line item in the federal budget.
• The Luddite Summer Meets the Long Boom. Aman’s sharpest historical observation: this may be the first technological revolution whose leaders are the doomers — Sam prophesying idleness, Dario predicting half of entry-level white-collar jobs destroyed within five years (already wrong at eighteen months, with no 10–20 percent unemployment in sight). Against the WSJ’s jobless-boom and nation-of-Luddites anxieties, the book offers the long view: of ten historical bubbles, the two positive ones — Britain’s 1845 railway mania and America’s 1997–2000 internet boom — overbuilt, crashed, and left the world a valuable technology. Buy every stock founded in the boom and hold, and you’d have owned NVIDIA, Amazon, Google, and PayPal. The 1970s wiped out four to six million secretarial jobs in a decade; women’s participation rose from 52 to 77 percent. And on China, the free-trader’s answer: partners in progress — there’s more to gain than lose if we do this right.
About the Guests
Dave McClure and Aman Verjee are the co-founders and managing partners of Practical Venture Capital, a Silicon Valley firm specializing in venture secondaries. Dave founded 500 Startups, invested at Founders Fund, and ran marketing at PayPal; Aman was COO of 500 Startups, led strategy at PayPal and eBay, served as CFO of Sonos and of eBay’s North American marketplace — and wrote the first draft of PayPal’s S-1. Aman’s new book, A Brief History of Financial Bubbles (out this week), is available at bigbubbletrouble.com.
References:
• A Brief History of Financial Bubbles by Aman Verjee — ten manias from the tulips to the subprime crash, out this week at bigbubbletrouble.com.
• Reuters on Anthropic’s IPO filing — the $2 trillion valuation and the $190–200 billion 2028 revenue forecast it hinges on.
• “The Summer That America Became a Nation of Luddites” and the “jobless boom” — the Wall Street Journal pieces threading this week’s episodes.
• The New York Times on tech giants’ circular AI economy — the piece that framed yesterday’s TWTW debate and today’s rebuttal.
• The SpaceX S-1 — forward projections of $300 billion in ann...
