Emergency Pod: Anthropic's S-1 IPO Filing — With Dan Primack

Anthropic

Across the shows: 49 speakers in other episodes, 29 bullish and 38 bearish calls, past 30 days

• Dan Primack · AxiosUncertain

  Circular financial ties among Anthropic, SpaceX and Broadcom remain a real concern, though the economy should hold.

  Primack says the economy will hold up, but the circularity of the AI deal web is still the issue. He points to Anthropic's contracts with SpaceX and Broadcom and says these companies are so tightly connected to one another that the concern persists. The S-1 confirms the leaked financials, with revenue soaring and spending rising too, so the core risk he cares about is the interlocking commitments rather than the headline numbers.

  “the circularity of it is still, like, the issue. Right? ... All of these companies are so incredibly connected to one another. I think those concerns are still there.”at 0:57

• Dan Primack · AxiosBullish

  Anthropic will be able to raise its roughly $100B IPO because capital markets are eager to back the AI story.

  Primack says the S-1's contents are largely irrelevant to whether Anthropic can raise the expected ~$100B at a ~$2T valuation. He argues the market will give it what it wants, as it did SpaceX's ~$40B GPU financing. Capital markets have large AI exposure, and Anthropic and OpenAI are the headlines of that trade, so he expects the deal to get done. He adds that the offering is a company event, not a referendum that a few hundred billion dollars of valuation difference would change much.

  “the capital markets have a lot invested in the AI story, and Anthropic and OpenAI are kind of the headlines of that. I think they're going to be able to get it.”at 2:18

• Dan Primack · AxiosBullish

  Anthropic's revenue growth is extraordinary, and recent quarters matter more than trailing financials.

  Primack calls Anthropic's growth extraordinary and says investors are buying it. He argues that trailing numbers through June 30 matter less than the last three months in a market changing faster than any industry he has seen. The year-over-year figures from 2025 are, in his words, irrelevant, since the company is already past Q4. Revenue per employee of about $6.6M, versus about $400K in September 2025, shows how fast it is scaling at the top line.

  “everybody is buying growth and Anthropic has shown it's got extraordinary growth.”at 6:23

• Dan Primack · AxiosBullish

  Claude CoWork is growing faster than Claude Code did, Primack believes.

  Primack agrees with Alex that Claude Code and Claude CoWork have been a big part of Anthropic's ability to scale revenue. He thinks CoWork itself is growing faster than Code did, though he doesn't have the numbers in front of him. The view is tentative and rests on his recollection of the trajectory.

  “I think that [Claude] CoWork itself is actually growing faster than Code did. I think you're right about that.”at 13:13

• Dan Primack · AxiosUncertain

  Anthropic's biology push could become its core business in three or four years, though it earns little now.

  Primack notes the S-1 spends a lot of space on biology and bio efforts, which are not making money yet. He reads the document as suggesting that in three or four years the company's core revenue may come from this area, and he says Dario thinks so. The claim is a speculative reading of the filing's emphasis rather than established revenue.

  “You almost read this and think man in three or four years Dario kind of thinks that might be the core of their business in terms of where their money is coming from.”at 14:00

• Dan Primack · AxiosBullish

  Anthropic's bet on B2B has clearly paid off, and its customer base is diversified.

  Primack says Anthropic made a bet, that most people weren't making a couple of years ago, to start with business customers or hit that market hard through sales, and it has clearly paid off. No single customer exceeds 10% of revenue, and about 79% comes from outside the top five customers, a long tail he says the company needs to be foundational. He downplays the 48% of revenue tied to Anthropic, Google and Microsoft marketplaces, since the end users are essentially everyone. He cautions that the hyperscalers' investments in the company add a circularity question.

  “they made a bet that most people weren't making a couple of years ago to start there, or if not start there, to really hit there hard from a sales perspective. And it has clearly paid off.”at 15:50

• Dan Primack · AxiosBullish

  Compute efficiency gains are Anthropic's main bull case, since its problem is spend rather than revenue.

  Primack says Anthropic's problem is the spend side, not the revenue side. Revenue per dollar of compute rose from about 40 cents in Q1 2025 to about $2.08 today, per the chart Alex cites. If models become much more efficient and need less compute, its spending could drop sharply as a share of revenue. He argues that efficiency is the ballgame and would make these businesses viable, though the company is still unprofitable.

  “It's not the revenue side, it's the spend side. The efficiency is kind of the ballgame for them. And if they can point to increased efficiency going forward, that's going to make these actually viable businesses.”at 17:41

• Dan Primack · AxiosUncertain

  Anthropic needs IPO cash to fund itself; it is still unprofitable and must prove viable businesses.

  Primack says Anthropic is going public because it needs money, not just for headlines. It is still unprofitable as a whole, even if it had a few profitable quarters, and its large compute commitments require ongoing spending. He compares it with Amazon, which ran losses for years on a reinvestment story, but notes the capital involved now is far larger and Anthropic will have to prove these can be viable businesses. Dan frames this as a risk rather than a clear call.

  “Anthropic isn't just going public because it wants headlines. Anthropic is going public because it needs money.”at 17:50

• Dan Primack · AxiosNeutral

  Anthropic will spend heavily on acquisitions, using its new public stock as currency.

  Primack says that for Anthropic to be more than a utility, it has to build applications on top of the models. It is building a lot in house and will also buy a lot of that, likely using cash and, after going public, stock. He notes that OpenAI has done more of this so far. The acquisitions will consume cash that the company will need.

  “They're gonna use a lot of cash to buy stuff, and they're gonna need cash to buy stuff.”at 18:53

• Dan Primack · AxiosBullish

  Anthropic will have no trouble buying compute, as one of the largest buyers at the front of the line.

  Primack has no concern that Anthropic will be unable to buy the compute it needs. He says it is one of the biggest buyers on the block and will be at the front of the line, not a new startup. The IPO and the extra cash on hand would reinforce that position. He calls the compute-access risk factor the most practical one in the filing, even though it isn't the most interesting.

  “I have no concerns that Anthropix is gonna be able to buy the compute. Right? Because for starters, it's one of the biggest dogs on the block. They are going to be at the front of the line.”at 20:13

Space Exploration Technologies

Across the shows: 32 speakers in other episodes, 26 bullish and 15 bearish calls, past 30 days

• Dan Primack · AxiosUncertain

  SpaceX's valuation rested on its AI compute story despite no demonstrated AI compute business.

  Primack says SpaceX has a diversified business in launch and Starlink, but the valuation and the money it raised were really based on its AI compute piece. The market was happy to give it that even though SpaceX had no demonstrated business in AI compute. He uses this to argue that the market is paying for growth stories, and that Anthropic's case is similar.

  “the valuation that SpaceX got, the money SpaceX got, was really based on the AI compute piece... even though they didn't really have a demonstrated business on AI compute.”at 6:23

OpenAI

Across the shows: 22 speakers in other episodes, 10 bullish and 17 bearish calls, past 30 days

• Dan Primack · AxiosUncertain

  Anthropic's IPO trading multiple will directly shape OpenAI's valuation when it goes public.

  Primack calls Anthropic and OpenAI essentially the same company, making Anthropic the most direct comparison. He expects whatever Anthropic trades at, including on a multiples basis, to affect OpenAI's IPO valuation, unless some major innovation or change happens in between. SpaceX is a looser comparison. He sees the read-through as direct but conditional on what happens to Anthropic.

  “whatever Anthropic is trading at, even from multiples perspective, is going to impact what happens to OpenAI if and when it goes public, unless there's some major innovation or change in the interim.”at 22:41

AI infrastructure & data centers

• Dan Primack · Axios

  Wall Street can absorb the parallel AI mega-raises because they draw on different pools of capital.

  Primack reads the 'capacity test' framing and says Wall Street will pass it, perhaps with a 'gentleman's B'. He points out that SpaceX's investment-grade debt, Meta's equity raise and Anthropic's IPO each draw on different investor pools, even within the same institutions. He calls the combined sum extraordinary, equal to about 0.7% of US GDP, but says it would be more questionable if roughly $300B of other IPOs hit at the same time.

  “I think it's capacity test, and I think Wall Street is going to pass said test. I think it's gonna pass it with a a or maybe a gentleman's b.”at 3:34

• Dan Primack · Axios

  Data centers may become far more efficient, meaning much less buildout spend than current plans imply.

  Primack notes the large buildout of data centers and the money and energy going into them. He asks what happens if data centers, or the models running on them, become far more efficient, so that much of that infrastructure isn't needed. That scenario is a bull case for Anthropic, since its spending would fall as a percentage of revenue, but it implies a bearish read for data center capex. He frames it as a possibility rather than a forecast.

  “What if data centers become just so much more efficient, not incrementally, but becomes so much more efficient that we, or GPUs do, that we don't need all the stuff or if the models do.”at 17:09

AI bubble & tech valuations

• Dan Primack · Axios

  Investors are buying the AI story and will accept one basket member failing.

  Primack argues that institutional investors cannot afford to miss the hottest IPO, so they will buy regardless of how the numbers look. He says the purchase is driven by narrative more than balance sheets, and compares it to the SpaceX story. Investors will treat the AI names as a basket, expecting one to fail, 'like a tomato' splatting, while believing the basket as a whole will be extraordinary. He expects the same pattern with OpenAI next year.

  “I think investors will look at this and say, all right, we're going to be wrong about one of these, but we need to take the basket because the basket as a whole is going to be extraordinary for us.”at 10:16

AI: overall outlook

• Dan Primack · Axios

  Frontier AI labs will be foundational, unlike fiber-laying CLECs, since they also build applications.

  Primack says the story pitched by Anthropic and OpenAI is that foundation models will be the foundation for business and society. He compares this with the dot-com era, when CLECs laid fiber and did not capture value on top, which ended badly for most. He argues the difference is that these companies build many applications themselves and are trying to be both the utility and the application layer. The comparison is a sober risk framing as well as a bull case.

  “these companies themselves are actually building a lot of applications, whereas the CLEC's of the day and the fiber layers didn't build anything on top of it. In this case, Anthropic is trying to be both and so is OpenAI.”at 11:16