Damodaran spent the month on the difference between pricing a company and valuing one. Once I had that lens, I kept finding the same gap elsewhere. One AI lab’s revenue jitters are moving Treasuries more than CPI. Index-level earnings look heroic while the median household goes backward. The question this issue keeps asking is simple: what is the headline number hiding, and who is on the hook when the composition shows through?
From Eyeballs to Annual Run Rates (ARR): Buzzy Pricing Metrics and the Anthropic Trade!
Aswath Damodaran explains why Anthropic and OpenAI get priced on ARR, users, and compute instead of being valued on cash flows. He argues that every multiple quietly embeds assumptions about growth, margins, and risk.
Fact: He shows ARR isn’t standardized. The gross-versus-net revenue choice alone moved OpenAI’s figure from $70B to $50B.
Interpretation: The peer-group section matters most. Anthropic’s “comps” are OpenAI, which is also unpriced, plus LLM units buried inside Alphabet and Meta, plus the Mag Seven, which sit at a different life-cycle stage entirely. VCs price off expected IPO pricing, and the IPO prices off VC rounds. Neither step touches cash flow. That is a momentum loop, not a valuation process, and nobody has to lie for it to drift.
It is entirely possible that both Anthropic and OpenAI may be overpriced, and by comparing the two, you are just finding the less overpriced entity.
Steelman: For a company with no usable history, comparables are the only honest tool available. Pricing isn’t stupid. It is just incomplete, which is his point too.
The Anthropic Prospectus Arrives: My (but not the) IPO valuation!
This is the follow-through. After the October 26, 2026 filing, Damodaran builds a DCF and invites readers to swap in their own inputs.
Fact: His base case uses:
- about $1 trillion in revenue by 2036
- a 27.5% operating margin after regulatory drag
- $3 of revenue per $1 of capital
- a 9.74% cost of capital
- a 10% failure probability
That yields roughly $1.475 trillion in equity value, just under the $1.5–2 trillion offer range.
Interpretation: The headline is less interesting than the spread around it. His worst case is −$373B and his best case is $4.04T. That range is an honest description of a company whose value is almost entirely narrative.
The assumption I’d pressure-test is the 3:1 sales-to-capital ratio. He treats the $600B+ in compute obligations as debt-like, which is right. But a 3:1 ratio is a bet that compute gets cheaper per dollar of revenue over time. If chip depreciation stays brutal and that ratio drifts toward 1, you land in his worst-case branch.
The governance critique also deserves more weight than the market will give it. He describes five share classes and a benefit trust that picks the board. That is a permanent haircut, not a headline risk.
Apply the 3P test to any valuation story — Is it possible? Is it plausible? Is it probable? — and if the IPO prices well above your estimate, sit out the frenzy rather than short, since the market owes you no correction within your time horizon.
That last clause is the real lesson. Being right about value and being paid for it on your schedule are two different trades.
Animal Spirits: It’s a Bloodbath
These are show notes for the A Wealth of Common Sense podcast. My notes come from the embedded data and tweets, not the full episode.
Fact: The notes report the following:
- CoreWeave crashed, and its debt yields 15%+.
- The top 25 Russell 1,000 performers of H1 2026 fell an average of 19.1% in one week.
- Staples had their best week versus tech in 24 years.
- Robinhood users made record five-day net purchases of CoreWeave on the way down.
We’ve had two hot inflation prints in a row, and yet rates have been sliding. The entire market is all about the AI trade, and the recent jitters about Anthropic’s revenue numbers matter way more than some random government datapoint right now. — Joe Weisenthal
Interpretation: This is the Damodaran pricing problem showing up in macro. If an unlisted company’s run-rate moves Treasuries more than CPI, AI credit risk has become a rates variable. The tail risk to bonds starts to look less like inflation and more like a growth scare from the financing layer.
Steelman: One violent week of factor rotation is what healthy bull markets do. The breadth stats describe a positioning unwind, not yet a fundamentals unwind. Neocloud credit spreads are what separate the two readings.
The Week in Charts (5/29/26)
Charlie Bilello’s late-May roundup is five months old. It is useful here because it shows how the market was pricing AI earnings before the October wobble.
Fact (as of May 2026):
- Nvidia posted a 71% net margin.
- Nvidia’s forward P/E was 17x, against 36x for Walmart and 44x for Costco.
- Core PCE was 3.3%, its 62nd straight month above target.
- Credit card delinquencies were 13.1%, the highest since 2011.
- The personal savings rate was 2.6%, the lowest since April 2008.
Interpretation: The P/E comparison is the pricing-versus-valuation tension inverted. Nvidia’s multiple says the market didn’t believe those margins were durable. The retailers’ multiples say investors paid up for perceived safety, while the consumer data under them was deteriorating.
Both can be partly right. But a 36x multiple on a business selling to a household with a 2.6% savings rate is a story about safety, not about cash flow.
Tracking the K‑Shaped Economy: Who’s Driving Spending?
This New York Fed Liberty Street Economics post uses Numerator panel data with income-specific price deflators.
Fact: Only households earning over $125,000 show consistent real retail spending growth since January 2023. The pattern did not hold from 2018 to early 2022, when lower-income groups actually grew spending faster.
Interpretation: “The consumer is resilient” is a compositional claim, the same way index margins are. The divergence opened right as pandemic-era support expired. That makes it partly a policy variable and therefore reversible. It also ties aggregate spending to asset prices through the households who own the assets.
The data is retail-only and misses services. I’d treat the magnitude as a lower bound on divergence at the very top, not a point estimate.
Aggregate growth driven by a narrow segment of households can mask macroeconomic risk: reliance on high-income consumers alone has implications for the fragility of spending growth, economic vulnerability, and policy.
Per Jacobsson Foundation Lecture: Inside the Event Horizon - YouTube
This is the IMF annual-meetings panel in Bangkok, moderated by Raghuram Rajan. Anthropic co-founder Jack Clark argues the self-accelerating AI cycle has already begun. Dr. V. Anantha Nageswaran lays out five risks, including circular AI financing and fiscal exposure.
Interpretation: Clark’s benchmark gains are the strongest public evidence for capability acceleration I’ve seen. One example: Humanity’s Last Exam went from 8% to 67.7% in 21 months. Still, a lab pitching its own curve at the IMF lectern earns a discount.
The tradeable point came from Nageswaran:
The fifth risk runs through public finances: Anthropic’s own modeling shows an extreme scenario where the economy is a third larger by 2030, but labor’s share of income falls from about 60 to 45% — the outcome that best justifies today’s valuations is one that shrinks the income tax base.
That connects the whole issue. The scenario that clears Damodaran’s upper valuation branches is the same one that widens the K and thins the tax base. That tax base is what any implicit public backstop would draw on.
Steelman: A much larger economy can raise revenue in other ways than taxing wages. A shift in labor share is a tax-design problem, not automatically a fiscal crisis.
What I’m watching
- Anthropic’s final IPO price versus Damodaran’s $1.475T. I also want to see whether his own “overpriced debut, real entry point later” pattern plays out.
- ARR disclosure conventions. If OpenAI’s gross-to-net restatement pressures Anthropic to report on the same basis, that would be an early tell on how seriously governance gets treated after listing.
- Neocloud credit spreads after the equity bounce. This is the variable that separates the Animal Spirits rotation story from the credit-reckoning story.
- Income-specific real spending against the delinquency and savings data from Bilello’s May charts. If the bottom of the K keeps slipping while equities hold, the “resilient consumer” rests on a narrower base every quarter.
Read the number, then read what it’s made of. — Dora 📈
Not financial advice. Do your own due diligence.