Anthropic IPO Valuation Multi
Full research prompt
Calculate the valuation as a multiple of run-rate revenue and of forecast 2027 revenue, citing the source of each revenue figure (prospectus, company statements, or "publicly estimated" analyst figures). Compare with the OpenAI private round at $1.25 trillion (its implied multiples, terms, and date) and with Nvidia (market cap, revenue multiple, margins, growth). Include a comparison table and note differences in growth, gross margin, and compute costs.
From Anthropic IPO Valuation, Share Price and First-Day Trading Explained
Anthropic’s public valuation, set by its November 18, 2026 IPO at $83 and its first-day close, prices the company at roughly 14–19× its own October run-rate and about 8–11× bank forecasts for 2027 calendar revenue—materially cheaper on current run-rate than OpenAI’s contemporaneous private mark, and in the same neighborhood as Nvidia once Nvidia’s much higher margins and free-cash-flow conversion are ignored. The gap that matters is not the headline multiple; it is that Anthropic’s revenue is still a fraction of the compute it has already contracted to buy.
Anthropic’s valuation multiples
Anthropic priced 980 million Class A shares at $83 on November 18, 2026 (above the $70–$78 range), raising $81.3 billion in the IPO plus a concurrent $26 billion private placement from chip and cloud suppliers. On the roughly 13.22 billion Class A and Class N shares outstanding after the offering and placement, that price implied about $1.10 trillion of equity value; on a fully diluted basis (options, RSUs, and convertibles) the figure cited by Reuters and others was about $1.5 trillion. Shares closed the first session (November 19) at $97.17, a market capitalization of about $1.27–$1.28 trillion on the same outstanding-share count. [1] [2] [3] [4]
Run-rate revenue is a company figure from the prospectus, not an analyst construct. The amended S-1 states that Run-Rate Revenue—last 28 days of usage fees annualized (×13) plus a month of subscription fees annualized (×12)—was approximately $81 billion as of October 31, 2026, up $10 billion from September 30. Nine-month 2026 GAAP revenue was $33.2 billion and third-quarter revenue was about $17.0 billion. The prospectus contains no forward guidance. [5] [6] [7]
Publicly estimated 2027 revenue comes from the banks, not the company. Goldman Sachs (a lead underwriter) projected about $136 billion of 2027 revenue in materials discussed with investors; other sell-side work clustered at $135–137 billion for 2027 and $270–280 billion for 2028. Those are not prospectus forecasts. [8] [9] [10]
| Valuation basis | Equity value | ÷ $81B Oct run-rate (prospectus) | ÷ ~$136B 2027E (Goldman / sell-side) |
|---|---|---|---|
| IPO price, shares outstanding | ~$1.10T | ~13.6× | ~8.1× |
| IPO price, fully diluted (reported) | ~$1.5T | ~18.5× | ~11.0× |
| Nov 19 close, shares outstanding | ~$1.28T | ~15.8× | ~9.4× |
Pro forma cash, cash equivalents, and short-term investments after the IPO and placement were about $232 billion in the final prospectus. Subtracting that cash (and ignoring modest debt) drops enterprise value at the close to roughly $1.05 trillion, or about 13× the October run-rate and about 7.7× the 2027 bank forecast. Multiples above are equity values unless noted; cash-rich balance sheets make EV the cleaner comparison to Nvidia. [11]
OpenAI’s $1.25 trillion private round
OpenAI’s latest private mark is a partial close, not a completed round and not a public filing. Bloomberg reported that roughly $11 billion had been wired by November 6, 2026, as the first tranche of a round valuing the company at $1.25 trillion, with the total potentially reaching about $20 billion. That was a step down from earlier talk of $30 billion or more at $1.4 trillion. Reported participants included Coatue, T. Rowe Price, SoftBank, Sequoia, Founders Fund, Andreessen Horowitz, and UAE-linked capital. OpenAI did not confirm the figures. Sources differ on whether $1.25 trillion is strictly pre- or post-money; the number attached to the round in the November reporting is $1.25 trillion. [12] [13] [14]
Revenue is company communication to investors, not a prospectus. OpenAI told investors its annualized revenue was approaching $50 billion at the end of September 2026 on a net basis (after partner share on some cloud sales), and that it expected to reach or exceed $70 billion annualized by year-end 2026, driven by enterprise. Earlier “grossed-up” figures near $70 billion mixed methods with Anthropic’s. There is no audited public revenue line. [15] [16]
Implied multiples on the $1.25 trillion mark:
- About 25× the ~$50 billion end-September run-rate (company statements to investors).
- About 18× the company’s own year-end 2026 target of ≥$70 billion annualized.
- No equivalent bank 2027 calendar forecast is in the same public circulation as Anthropic’s $136 billion figure. Independent trackers (for example 06.ai’s base case of a ~$200 billion end-2027 run-rate) are model estimates, not company or underwriter numbers, and should not be treated as comparable to Goldman’s Anthropic case.
On the September figure OpenAI is priced richer than Anthropic’s outstanding-share IPO and close. On a year-end $70 billion target the gap narrows toward Anthropic’s fully diluted IPO multiple. The accounting difference matters: OpenAI’s disclosed run-rate is net of partner share; Anthropic’s prospectus run-rate is its own gross measure.
Nvidia
Nvidia’s market capitalization was about $3.5–3.65 trillion at the November 19, 2026 close (share price near $146, roughly 24.1 billion shares outstanding; quote services differ by a few percent on the exact cap). [17] [18]
Trailing revenue to the quarter ended October 25, 2026 is on the order of $362 billion. The just-reported Q3 FY2027 quarter was $115.8 billion, up 103% year over year, with data-center revenue of $107.6 billion. GAAP gross margin was 74.8% and operating margin 66.9%. Q4 revenue was guided to $130 billion (±3%), with gross margin guided to 71–72%. Management raised the following fiscal year’s revenue-growth outlook to about 75%. [19] [20] [21]
At ~$3.5–3.65 trillion, Nvidia trades at roughly 10× trailing twelve-month sales and in the high-single-digits on the fiscal year that includes the $130 billion quarter (management’s path implies fiscal 2027 revenue near $420 billion if the guide is met). That is a lower revenue multiple than either lab on current run-rate, earned with mid-70s gross margins, mid-60s operating margins, and tens of billions of quarterly operating cash flow—none of which the model labs currently produce.
Comparison
| Anthropic | OpenAI | Nvidia | |
|---|---|---|---|
| Valuation | ~$1.10T IPO outstanding / ~$1.5T fully diluted IPO / ~$1.28T Nov 19 close | $1.25T private mark (partial close ~Nov 6, 2026; round up to ~$20B) | ~$3.5–3.65T market cap (Nov 19, 2026) |
| Revenue base used | $81B Oct 31 run-rate (prospectus) | ~$50B end-Sep annualized, net (company to investors); ≥$70B targeted by end-2026 | ~$362B TTM; Q3 $115.8B |
| Multiple on that base | ~14–19× equity on run-rate (lower on EV after ~$232B cash) | ~25× on $50B; ~18× on $70B target | ~10× TTM sales |
| 2027 revenue | ~$136B publicly estimated (Goldman / sell-side); prospectus silent | No comparable public bank forecast cited | Path consistent with ~75% growth into the next fiscal year (company outlook) |
| Multiple on 2027E | ~8–11× equity (~8× EV at the close) | Not cleanly computable from public bank figures | High-single-digit on near-term forward revenue |
| Recent growth | Q3 revenue ~12× year earlier; run-rate +$10B in October alone | Company cited ~70% quarterly growth in annualized revenue (not a filed number) | Q3 revenue +103% YoY from a far larger base |
| Gross / operating margin | ~85% gross margin on cost of revenue (9M 2026); total compute & infrastructure ~57% of 9M revenue; Q3 GAAP operating loss after a Q2 profit | Not publicly filed; still scaling losses against compute | GAAP gross margin 74.8%; operating margin 66.9% (Q3 FY2027) |
| Compute cost structure | ~$901B aggregate commitments, substantially all compute and infrastructure; ~$71B due in the 12 months after Sept 2026 | Large multi-year compute contracts; dollar total not in a public filing comparable to the S-1 | Supplier: captures high margin on the labs’ spend; customer concentration and custom-silicon risk are the offset |
Sources for the table cells are the same as the narrative above: Anthropic prospectus and pricing reports, OpenAI funding and investor-revenue reporting, Nvidia’s November 2026 results and market quotes.
What the differences actually mean
Growth is not the scarce input. Anthropic’s third-quarter revenue was about twelve times the year-earlier quarter, and October alone added roughly $10 billion of run-rate—more than twice its entire 2025 revenue. OpenAI’s investor materials also describe very rapid quarterly expansion. Nvidia is still doubling from a base already larger than either lab’s run-rate. A revenue multiple that looks “cheap” versus OpenAI can still be expensive if the growth rate that justifies it has to be sustained while a fixed compute bill arrives.
Gross margin is the misleading comparison. Anthropic’s prospectus supports an ~85% gross margin for the first nine months of 2026 because cost of revenue is mostly customer-serving compute. Training, research, and free-tier compute sit elsewhere. Total compute and infrastructure expense was $18.8 billion on $33.2 billion of nine-month revenue (about 57%), and it rose to about 58% of revenue in the third quarter, coinciding with a swing back to an operating loss after a profitable second quarter. Short-term compute, which the company says costs significantly more than capacity under multi-year deals, was a material part of that quarter. Nvidia’s mid-70s gross margin and mid-60s operating margin are earned after the full cost of designing and selling the chips; they convert into cash. The labs’ reported gross margins do not.
Compute commitments are the binding constraint on the multiple. Anthropic has disclosed about $900.9 billion of aggregate purchase commitments and leases, substantially all compute and infrastructure, with roughly $70.7 billion falling due in the twelve months after September 30, 2026, and several hundred billion more before the end of 2030. That stack does not shrink if demand slows. The IPO and placement left the company with on the order of $230 billion of cash and short-term investments, which covers the near-term bill several times over and is why EV multiples are lower than equity multiples—but the commitments extend well past the cash. Nvidia’s economics are the mirror image: it is paid, at high margin, for a large share of that capacity. OpenAI faces the same structural issue (revenue must grow into contracted compute) without a public commitment schedule as detailed as Anthropic’s S-1.
For anyone underwriting the stock or a competing lab, the relevant question is not whether 9–16× a fast-growing run-rate is lower than OpenAI’s 25×. It is whether 2027–2028 revenue, at sustainable margins after the capacity lands, covers a commitment schedule already signed at a scale that rivals the equity value itself. Bank cases that get Anthropic to the mid-$100 billions in 2027 and toward $200 billion-plus in 2028 are what make the current price look like a mid-single-digit to high-single-digit forward revenue multiple. Those cases are publicly estimated analyst figures, not company guidance, and they assume the usage growth that has to keep filling capacity whose cost is already largely fixed.
Recent Findings Supplement (November 2026)
xAI (now integrated as SpaceXAI within SpaceX following its February 2026 all-stock merger) carries a $250 billion standalone valuation from the merger (implied value in the deal that valued the combined entity at ~$1.25 trillion). This sits at roughly 8.3× its most recent reported revenue run-rate of $30.2 billion (annualized from Q3 2026 Super Intelligence segment revenue of $7.56 billion per SpaceX's SEC 8-K filing as of September 30, 2026).[1][1]
This multiple reflects a blend of core Grok/API revenue (earlier standalone estimates around $500 million ARR mid-2026) plus significant compute-rental and infrastructure income (including deals with Anthropic, Google, and others), which has scaled the segment rapidly but introduces different economics than pure model labs.[2][3] No specific company or analyst 2027 revenue forecast for xAI/SpaceXAI appears in recent post-May 2026 sources; prior standalone targets (e.g., "north of $2 billion" for 2026) predate the merger and compute-leasing expansion.[4]
OpenAI's November 2026 funding round closed an initial ~$11 billion tranche at a $1.25 trillion pre-money valuation (with the round potentially reaching ~$20 billion total; first close wired by November 6, 2026).[5][6] This values it at ~18× on a $70 billion claimed end-2026 run-rate target or ~25× on the $50 billion net annualized run-rate reported for end-September 2026 (net of partner shares; the higher figure reflects grossing up for comparability with peers like Anthropic).[7][8][9] The round (involving SoftBank, MGX, Coatue, T. Rowe Price, etc.) came after scaling back from higher targets (~$1.4 trillion at up to $30 billion) amid AI stock volatility.[10]
Nvidia's market capitalization stood at approximately $3.5–3.75 trillion as of mid-to-late November 2026 (share price ~$146–$156). Its TTM revenue reached ~$362 billion (through October 2026), implying a revenue multiple of roughly 10×.[11][12] Q3 FY2027 (ended October 25, 2026) revenue hit $115.8 billion (+103% YoY), with data center at $107.6 billion; GAAP gross margin was 74.8%.[13][14] The company guided Q4 FY2027 revenue to $130 billion (±3%) and raised its FY2028 (fiscal year ending January 2028) revenue growth outlook to ~75% YoY (from prior ~70%).[15]
Comparison table (using latest consistent post-May 2026 figures; multiples are valuation/market cap divided by the cited revenue run-rate or TTM):
- xAI/SpaceXAI: Valuation $250B (Feb 2026 merger-implied); Run-rate revenue $30.2B (Q3 2026 annualized, SpaceX 8-K); Multiple ~8.3×; 2027 forecast: N/A in sources.
- OpenAI: Valuation $1.25T (Nov 2026 round); Run-rate revenue $50B net (end-Sep 2026) or $70B target (end-2026); Multiple ~25× or ~18×; 2027 forecast: Limited data; trajectory from $50B run-rate suggests potential further growth but no explicit figure.
- Nvidia: Market cap ~$3.5–3.75T (Nov 2026); TTM revenue ~$362B; Multiple ~10×; 2027/ FY2028 outlook: ~75% YoY growth (raised guidance).
Key differences in growth, gross margins, and compute costs: xAI's lower multiple stems partly from its hybrid model (product + large-scale compute rentals/infrastructure, now embedded in public SpaceX disclosures), contrasting with pure-play labs like OpenAI that command higher multiples on consumer/enterprise model revenue despite heavy losses.[16] OpenAI and Anthropic show explosive growth (e.g., OpenAI Q3 quarterly revenue reportedly >$10B, up ~50% QoQ in some reports) but face high compute commitments and negative or low margins; Anthropic's S-1 highlighted improving unit economics (revenue per compute dollar rising sharply) yet massive future infrastructure spend.[17][18] Nvidia enjoys superior ~75% gross margins as a chip seller (vs. the compute buyers' high costs), with sustained high growth from AI demand and strong cash generation.[13]
For competitors or entrants: xAI's integration into SpaceX provides capital and infrastructure advantages (e.g., compute leasing revenue stream) that pure startups lack, but it trades at a discount to OpenAI's valuation multiple. Nvidia's position as the margin-rich enabler of the ecosystem highlights why hardware moats differ from application/model layers. New entrants should note the sensitivity of multiples to accounting (net vs. gross revenue) and the capital intensity of scaling, where compute rental or efficiency gains can alter economics faster than model performance alone. Data is drawn from SEC filings, Bloomberg/Reuters reporting, and company disclosures post-May 2026; 2027 forecasts remain sparse outside Nvidia's guidance.