DX Today | No-Hype Podcast & News About AI & DX

Sixty Five Billion Dollars and the Question of What It Counts: Inside the Anthropic IPO Filing - August 23, 2026

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 13:49
Sixty Five Billion Dollars and the Question of What It Counts: Inside the Anthropic IPO Filing Anthropic is preparing to file its initial public offering prospectus publicly, following a confidential submission to regulators in June, carrying a headline annualized revenue run rate above sixty five billion dollars and a post money valuation of nine hundred sixty five billion. Rick Spair and Laura dig into the gross versus net revenue question hiding inside that headline, why the principal versus agent classification is about to face real scrutiny, what a projected first profitable quarter does and does not prove against eighty billion in committed infrastructure spending, and the four things worth reading first when the document finally lands. Hosted by Rick Spair and Laura. The DX Today Podcast brings you daily deep dives into the most consequential stories in the AI ecosystem. #AI #Anthropic #IPO #AIInvesting #TechFinance
SPEAKER_00

Welcome to the DX Today Podcast, your weekly deep dive into the AI ecosystem. I'm Rick Spare, and joining me as always is Laura.

SPEAKER_01

Thank you, Rick, and I have been waiting all week to talk about this one because there is a very large number sitting at the center of today's story, and I'm fairly convinced that almost nobody repeating it is reading it correctly.

SPEAKER_00

Let's set the table for people who have been living their actual lives this week instead of uh refreshing financial news feeds. What happened and why should anyone outside of finance care about it?

SPEAKER_01

Anthropic is preparing to publicly file its initial public offering prospectus, potentially as soon as the end of this month, after quietly submitting a confidential draft to regulators back at the beginning of June.

SPEAKER_00

So the confidential version has been sitting with the securities regulator for roughly three months, and now the version the rest of us actually get to read is about to land in public view for the first time.

SPEAKER_01

That is exactly the situation. And the headline number everyone is repeating is an annualized revenue run rate that surpassed $65 billion as of the end of July, which is a genuinely staggering figure by any historical standard you want to apply. Here is the trajectory, and I want you to feel the slope of it rather than just hear the endpoints. In January of 2024, the run rate was $87 million. That is million with an M.

SPEAKER_00

87 million is a perfectly respectable software business. Plenty of founders would take that outcome, ring the bell, and retire completely happy. So where does the curve travel from that fairly ordinary starting point?

SPEAKER_01

By December of 2024, it reached $1 billion. By the end of 2025, it was nine billion. By February of this year, $14 billion. By March, $19 billion. $30 billion by April, $47 billion by late May.

SPEAKER_00

And it keeps climbing from there, I assume, because otherwise you would not have that particular expression on your face right now as you read those figures off your notes to me.

SPEAKER_01

$65 billion by the end of July. That is roughly a seven-fold increase across seven months of calendar time, which is a slope that does not really have a comparison point in enterprise software history.

SPEAKER_00

And uh I have honestly never seen a curve shaped like that outside of a consumer social product in a viral moment. But you said the number is being misread.

SPEAKER_01

I did say that, and this is the part I think matters far more than the headline itself. The question is not how big the number is, the question is what precisely the number is actually counting.

SPEAKER_00

All right, walk me through it slowly because accounting is exactly where most people's eyes glaze over. And I have a strong suspicion this is precisely where the genuinely interesting part of the story lives.

SPEAKER_01

Imagine you are an enterprise customer and you buy one dollar of Claude tokens through a cloud marketplace like AWS or Azure. That entire dollar does not stay with the model company, and it never did.

SPEAKER_00

Right. The cloud provider takes a cut for hosting the transaction, handling the billing relationship, and uh carrying the customer contract on its own paper. That is a completely standard marketplace arrangement across enterprise software.

SPEAKER_01

Correct. And here's the crux of the whole thing. Anthropic books that entire dollar as revenue and then lists the cloud provider's share as a marketing expense sitting further down the income statement.

SPEAKER_00

So the top line shows a full dollar, even though a meaningful slice of that dollar walks straight out the door to a partner before anyone at the company ever really touches it.

SPEAKER_01

That is the gross method, and it is a legitimate choice. Now compare it to how the main competitor handles the identical economic transaction. On a comparable cloud marketplace sale, the reported figure is roughly the 20 cent share.

SPEAKER_00

Hold on a moment. So two companies can process an economically identical dollar of business, and one reports a full dollar of revenue while the other reports 20 cents. That is a serious problem for anyone trying to compare them.

SPEAKER_01

It is a serious problem for comparison, and yet both treatments can be perfectly legitimate under generally accepted accounting principles. This is the classic principle versus agent question, and the cleanest way to picture it is a travel agent selling you a flight. If they control the seat, they book the full ticket price. And if they are merely the agent, they book only their commission.

SPEAKER_00

So the entire argument comes down to who um genuinely controls the thing being delivered to the customer. And uh reasonable, uh, well-credentialed accountants can look at the very same contract and disagree with each other about that.

SPEAKER_01

They absolutely can. And that is exactly why regulator scrutiny of this specific classification is one of the most consequential open questions hanging over this entire public offering process right now.

SPEAKER_00

Because if regulators or auditors push the presentation toward the net method, that $65 billion headline could arrive in the public filing looking materially smaller than everyone in the market has been assuming all summer.

SPEAKER_01

Precisely that. And I want to be very clear here because this is where commentary usually goes off the rails within about a day. Nothing about any of this suggests anyone did anything improper or misleading.

SPEAKER_00

That is an important distinction uh to draw carefully. There is an enormous difference between an aggressive but entirely defensible presentation choice and something that anybody should responsibly describe as deceptive or fraudulent behavior.

SPEAKER_01

The difference matters enormously. Private companies choose a presentation and stick with it consistently. The moment audited financials hit a public prospectus, that choice gets stress tested by people whose entire professional job is skepticism.

SPEAKER_00

Let's talk about what else that prospectus is going to reveal because I imagine the revenue line is only one of several numbers that could meaningfully reset what the market thinks it already knows.

SPEAKER_01

The one I'm watching hardest is gross margin, along with the compute cost line item that sits underneath it. Those two figures have never been disclosed with audited precision by any frontier laboratory ever, anywhere.

SPEAKER_00

And those are precisely the numbers that tell you whether this is a genuinely durable business or an extremely impressive mechanism for converting um investor capital into computing bills at truly enormous scale.

SPEAKER_01

That is the whole question compressed into a single sentence, and I could not have phrased it better myself. Everything else is narrative. Gross margin either supports the narrative or quietly demolishes it.

SPEAKER_00

So what do the projections say? Um, at least according to the reporting um circulating ahead of the public filing, I assume the internal picture that investors have been shown is fairly optimistic on that particular front.

SPEAKER_01

The second quarter of this year is projected to show operating profit of $559 million on roughly $11 billion in revenue, which would be a genuine first for this entire industry.

SPEAKER_00

A Frontier Laboratory um posting an actual quarterly operating profit. I want to sit with that for a second because the conventional wisdom of this entire sector has been that real profitability was still several years away.

SPEAKER_01

It has been the settled assumption, and this would be the first visible crack in it. Preliminary second quarter revenue reportedly exceeded 11.5 billion, which is around 14 times the figure from a year earlier.

SPEAKER_00

14 times year over year, um, while uh simultaneously flipping to operating profit is a very unusual combination. Normally you get hypergrowth or you get profitability, and the market simply accepts that you have to pick one of them.

SPEAKER_01

You almost always do have to pick, which is exactly why this filing is going to be picked apart line by line by every analyst on the street. And there is a very large asterisk sitting right next to that profit figure.

SPEAKER_00

I had a feeling there would be. What is the asterisk? Um, and um how much should it temper the enthusiasm uh that this number is obviously going to generate across financial media over the next several days?

SPEAKER_01

Infrastructure spending. The estimated commitment runs to roughly $80 billion through the end of the decade, including five gigawatts of new capacity from one cloud partner and five gigawatts of next generation accelerators from another.

SPEAKER_00

10 gigawatts of uh committed capacity is not a software company commitment. That is much closer to the capital profile of a regulated utility or a heavy industrial manufacturer building out physical plant over many years.

SPEAKER_01

That comparison is far more apt than most people realize. And it means a single profitable quarter can very easily be followed by several deeply unprofitable ones as that committed spending actually lands on the books.

SPEAKER_00

Which um raises the governance question I have been chewing on all week. Two of the largest cloud partners here are also significant investors in the company itself. How on earth does that get disclosed cleanly?

SPEAKER_01

It has to be disclosed as a related party consideration, and it is genuinely tangled. The same entities are simultaneously shareholders, infrastructure vendors, and distribution channels for the underlying product being sold.

SPEAKER_00

So money flows in as investment, flows back out again as compute spending, and then flows around once more as marketplace revenue. Untangling the real economics of that from the outside sounds close to impossible.

SPEAKER_01

It is genuinely difficult. And that circularity is precisely what a rigorous prospectus is supposed to make legible to ordinary investors. Whether it actually succeeds at doing that is one of the first things I will be reading for.

SPEAKER_00

Let's zoom out to the strategic picture for a moment. Why file now in this particular narrow window rather than simply staying private for another year with what is obviously abundant access to capital?

SPEAKER_01

The funding round that closed at the end of May raised $65 billion at a post-money valuation of $965 billion, which pushed past the valuation of the primary competitor.

SPEAKER_00

So there is no capital necessity driving any of this. Nobody's running out of money, which means the timing is purely a strategic choice. And strategic choices always have a competitive logic hiding behind them.

SPEAKER_01

The logic appears to be sequencing. Being the first frontier laboratory to trade publicly sets the valuation framework, the disclosure norms, and the comparison baseline that every single subsequent listing then has to answer to.

SPEAKER_00

That is a real and durable advantage, and it is exactly the kind of advantage that looks invisible in a press release, but turns out to be enormously valuable 18 months later when a competitor files.

SPEAKER_01

And it works in the other direction too. Whoever goes second gets measured against whatever the first company disclosed on metrics that the first company effectively got to define on its own terms.

SPEAKER_00

What about the customer concentration question? My understanding is that the revenue mix here leans very heavily toward large corporate customers rather than the individual consumer subscriptions that dominate the public conversation.

SPEAKER_01

Roughly 80% enterprise, with eight of the ten largest American corporations as paying users, and the count of customers spending more than a million dollars annually doubling in about two months this spring.

SPEAKER_00

That is a far stickier revenue base than consumer subscriptions, but it also concentrates risk considerably. Losing a handful of very large accounts moves the needle in a way that a consumer business simply never experiences.

SPEAKER_01

Both of those things are true simultaneously. Enterprise revenue is remarkably durable right up until the moment a procurement committee decides the underlying capability has become a commodity that three different vendors can supply equally well.

SPEAKER_00

And commoditization is the long-term risk that absolutely everyone waves at and nobody can really price with any confidence. So what should our listeners actually watch for when this document finally becomes public?

SPEAKER_01

Three things specifically. First, whether revenue is presented on a gross or a net basis. Second, the audited gross margin figure. Third, the exact language describing the related party relationships with those cloud partners.

SPEAKER_00

That is a clear and genuinely useful checklist. And I would add a fourth item, which is simply how honestly the risk factor section describes the sustainability of that one single profitable quarter.

SPEAKER_01

That is a very good addition. Risk factor sections are usually pure boilerplate, but in a first of its kind filing, they occasionally contain the most honest writing in the entire several hundred-page document.

SPEAKER_00

Which is a wonderfully uh strange thing uh about public offerings generally. The section legally required to explain why the company might fail is often the only place where the real business gets described plainly.

SPEAKER_01

It really is, and that is exactly why I read them first and read them twice. The narrative section tells you what a company wants to be. The risk factors tell you what keeps the executives awake at night.

SPEAKER_00

That is a perfect note to close on. Um Thank you, Laura. This was one of the more genuinely useful conversations we have had about a financial story in quite a long while.

SPEAKER_01

Thank you, Rick. And to everyone listening, when that document finally lands, resist the headline number and go read the accounting policy note instead. That is where the actual story of this company is going to live.

SPEAKER_00

That's all for today's episode of the DX Today podcast. Thanks for listening, and we'll see you next time.