ED ZITRON: Anthropic legally left out the numbers that matter most, so retail buyers get less than its bankers and lenders

The Gist

Ed Zitron says Anthropic followed the rules but left out its latest quarter, its cash position and its payment schedule. Big investors already heard some of those numbers in private, so ordinary buyers end up knowing less. This is a steelman: Ed Zitron's arguments from the conversation are reconstructed in their strongest fair form, with his answers to the host built in. It does not endorse them.

Conclusion

Even if the omissions comply with SEC timing rules, they give public buyers a less complete and more favorable picture than insiders and lenders have.

Premises

1. The S-1 contains financial statements only through June 30, 2026; Q3 revenue and the Q3 revenue-per-compute ratio were shared with private investors and reported in the press, but they are not in the filing.
2. The filing's contractual obligations table, its aggregate compute commitment amounts and its run-rate revenue figure are blank, and it gives no cash position for September 30, 2026.
3. A company that chooses not to present a number it already has is usually doing so because the number does not flatter it.
4. Retail buyers rely on the public filing, while lenders and large investors receive private briefings and, for debt, more detailed information.

Assumptions

• Disclosure: Ed Zitron is CEO of EZPR, a media relations and public relations firm. He writes the Where's Your Ed At newsletter, which has a paid premium tier, and hosts the Better Offline podcast, produced with iHeartRadio and Cool Zone Media; both cover the AI industry critically, and in this conversation he says he predicted the rating outcome ("told you so"). Sources: "About Where's Your Ed At" (Where's Your Ed At) at https://www.wheresyoured.at/about/, "Introducing: Better Offline" (Where's Your Ed At) at https://www.wheresyoured.at/introducing-better-offline/ and "Ed Zitron" author page…
• The conversation was recorded on October 30, 2026, as the FT story on Anthropic's expected rating broke; figures are as stated live and may be approximate.
• "Retail investors" means individual buyers who rely on public filings and have no access to management briefings.
• Differs: The S-1 does disclose some September 30, 2026 data: net revenue retention of about 900 percent, more than 11,900 customers above $100,000, and no customer above 10 percent of nine-month revenue.

Analysis

Argument type: Inductive.

Potential Fallacies

• Argument from silence / inference from absence (P3 and the conclusion ('more favorable picture')) — The absence of a figure is treated as evidence about what the figure says. Without seeing the numbers, or knowing how often omitted data in comparable filings turns out to be unfavorable, the omission fits several explanations about equally well: timing rules, audit and review status, quiet-period caution, liability for unreviewed figures, and placeholders in a draft.
• Unsupported generalization (P3) — 'Usually' gives P3 a statistical tone, but no reference-class frequency is offered. It is the only bridge from 'omitted' to 'unflattering'. It also carries no filter for materiality or for legal constraints on what can be included.
• Equivocation between 'absent' and 'chosen not to present' (P1/P2 to P3) — P3 applies to discretionary choices to withhold. P1 and P2 establish only absence. The conclusion itself concedes that SEC timing rules permit the Q3 omission, so it is not shown to be a free choice. Blanks in a draft filing may also be unfinished items, not decisions.
• Scope shift from 'less complete' to 'more favorable' (Conclusion) — Information asymmetry (well supported) is presented as if it also showed a flattering skew (not shown). A picture can be less complete without being more favorable. If the Q3 numbers are strong, leaving them out makes the public picture less favorable, not more.
• Selective emphasis (P1/P2 versus A4) — The argument stresses the blanks and omitted Q3 figures while giving little weight to the Sept 30 metrics that were disclosed (about 900% net revenue retention, 11,900+ large customers, no customer above 10%). That disclosure complicates the story of a filing stripped of Q3 information. It could support a cherry-picking reading, but the argument does not make that case.
• Mismatched comparison of recipient classes (P4 and conclusion) — Lenders and private investors negotiate information rights, sign confidentiality agreements, hold different claims, and are briefed under different legal regimes. Treating their access as the benchmark for retail buyers makes the asymmetry look like evidence of unfairness, when it is also a normal feature of capital structure and offering rules. A fairness concern can still survive these distinctions.

Counterarguments

These challenges are not ranked.

• Premise 3 (High impact) — Omitting unreviewed or unaudited stub-period results from an S-1 is routine practice. Counsel generally advises against including unreviewed numbers because of Section 11 liability and gun-jumping risk, so the omission says nothing reliable about whether the numbers flatter the company.
• Conclusion (High impact) — The initial S-1 is a draft. Blanks are routinely filled and recent results added by amendment or in the preliminary prospectus before pricing. If that happens, the asymmetry is a transient drafting-stage artifact and not a lasting disadvantage to retail buyers.
• Conclusion (High impact) — If the withheld Q3 numbers are strong, as the disclosed 900% net revenue retention and customer growth might suggest, then omitting them makes the public picture less favorable, not more. The press-reported Q3 figures are also visible to retail buyers, so the asymmetry may be smaller than claimed.
• Premise 4 (Medium impact) — Lenders and private investors bargain for information rights, are under confidentiality obligations and carry different risks than equity buyers. Securities law requires that the registration statement not be misleading, not that all parties receive the same information, so the asymmetry is structural and not evidence of favoritism.
• Premise 2 (Medium impact) — Blank fields in a preliminary filing may be placeholders pending final contract terms or audit sign-off, and not decisions to withhold. Compute commitments are not price-dependent, so their absence is more notable than a typical blank, but intent cannot be inferred from the blank alone.
• Conclusion (Low impact) — The source is a committed AI skeptic with a stated 'told you so' stake, a paid newsletter and live, possibly approximate figures. This does not falsify the factual claims, which are independently confirmed, but it raises the bar for accepting his inferences about motive and content.

Suggested Improvements

• Separate the two halves of the conclusion — Restate the conclusion as 'less complete, and plausibly skewed favorable if the omitted figures are adverse,' or defend the two claims separately. The 'less complete' claim is well supported and the 'more favorable' claim is not, so bundling them lets the weaker half borrow credibility from the stronger one.
• Support P3 with evidence — Provide a base rate or comparable IPO cases showing how often held-back stub-period or commitment figures turned out worse than disclosed figures, or show specific evidence that Anthropic's Q3 numbers are adverse. Without it, P3 is a heuristic and not evidence, and it carries the whole directional claim.
• Address benign explanations directly — Explain why the usual reasons (staleness window, audit and review timing, liability for unreviewed numbers, placeholder blanks) do not account for these omissions, especially for non-price-dependent items such as compute commitments. The concession of legal compliance removes the easy rebuttal but leaves the best rival explanation unanswered.
• Integrate the disclosed Sept 30 metrics — Make a net assessment of what the filing discloses and omits, and state whether the favorable disclosed metrics are selective emphasis or genuine transparency. A4 shows the filing is not simply missing recent data. The argument is stronger as an account of selective emphasis than as one of blanket absence.
• Take the amendment process into account — Frame the claim around the final prospectus and what is in it before pricing, and specify what update would resolve the objection (for example, an S-1/A with Q3 results, commitment totals and cash position). It turns a snapshot critique into a testable one and locates the practical lever for change.
• Narrow the retail claim and state a limiting principle — Identify who actually buys on the filing alone (small-allocation and aftermarket buyers), and state a principle, such as materiality or a known-trend disclosure duty, for when private information should also be public. Full informational parity with lenders and private investors is not how registered offerings work, so a limiting principle makes the normative claim defensible.

Scenario Tests

• Anthropic files an S-1/A before the roadshow that includes Q3 results, the contractual obligations discussion, commitment totals and a Sept 30 cash position. (Challenges) — The asymmetry largely disappears and the original filing looks like a routine first draft. If the new numbers are weak, the filing's earlier omission looks more suspicious, but the argument would still need a separate showing of intent.
• The Q3 figures, once public, show strong growth and an improving revenue-per-compute ratio. (Challenges) — The public picture was less favorable than the private one, which reverses the direction of the conclusion and shows P3 failing in this case.
• The Q3 figures show deteriorating revenue-per-compute or large new compute commitments that are not reflected in the June data, and Anthropic does not add them before pricing. (Supports) — The omission would then leave public buyers with a rosier view than lenders and private investors, and the case for a known-trend disclosure under Item 303 would strengthen. The argument's directional claim would be vindicated.
• A survey of recent large IPO filings shows most also stop at the last reviewed quarter and leave non-price items blank in their first filing. (Challenges) — Omission would then be the default practice and carry little information about the numbers, which undercuts P3 and shows base-rate neglect in the argument.
• Retail buyers in the IPO get a small allocation while institutions that have seen both the roadshow and private briefings take most of the shares and set the price. (Neutral) — The information gap is real but its harm falls on a narrower group, mainly small-allocation and aftermarket buyers, and the price already partly reflects the private information.

Coherence & Relevance

The argument is coherent and plausible as an inductive case for the narrower claim that public buyers have a less complete picture than insiders and lenders. The first three premises and the fourth together establish that claim well. It is much weaker on the second claim, that the picture is more favorable, which depends on a single unsupported heuristic and sits in tension with the disclosed favorable September 30 data. The concession of SEC compliance usefully moves the dispute from legality to fairness and completeness, but the argument does not follow through by showing the omitted numbers are adverse, that the omission was a choice, or that the gap will persist through the final prospectus.

• The S-1 contains financial statements only through June 30, 2026; Q3 revenue and the Q3 revenue-per-compute ratio were shared with private investors and reported in the press, but they are not in the filing. (Strong) — It establishes the gap in information directly but not its direction, and it concedes that timing rules permit the omission. It does not show the omission was a choice, and press reporting means the figures are not wholly inaccessible to the public.
• The filing's contractual obligations table, its aggregate compute commitment amounts and its run-rate revenue figure are blank, and it gives no cash position for September 30, 2026. (Moderate) — The blanks show incompleteness, and compute commitments are plainly material to risk. The premise does not distinguish deliberate withholding from placeholders in a draft, and it does not show whether the commitment totals are alarming or ordinary.
• A company that chooses not to present a number it already has is usually doing so because the number does not flatter it. (Strong) — It is the one premise that links omission to 'more favorable', but it is only a probabilistic heuristic. It applies to deliberate choice and not to legally permitted timing, offers no base rate, and is in tension with the disclosed favorable Sept 30 metrics.
• Retail buyers rely on the public filing, while lenders and large investors receive private briefings and, for debt, more detailed information. (Moderate) — It establishes who holds what information, which supports 'less complete'. It does not show that the private information is less favorable, and it glosses over the legitimate reasons lenders and private investors receive different information.

View this argument on LogicFirst.ai