AI Research

Anthropic IPO Prices at $70–$78, Targeting Up to $1.4 Trillion Valuation — Pricing Nov 18, Trading Nov 19

Anthropic IPO pricing announcement with Nasdaq ticker ANT and $1.4 trillion valuation

Anthropic IPO: The $1.4 Trillion Valuation Test for Pure-Play AI Labs

Anthropic filed its amended S-1 on November 10, setting an IPO price range of $70–$78 per share for 825 million Class A shares — targeting a $57.75B to $64.35B raise and an implied fully diluted valuation of up to $1.404 trillion. Pricing is expected the evening of November 18, with Nasdaq trading under ticker ANT starting November 19. This is the first live market pricing for a frontier AI lab — a reference point for OpenAI's eventual IPO and every private AI valuation.

The company also secured a $26B private placement of non-voting Class N shares at the IPO price with seven strategic investors: Google ($10B), NVIDIA ($10B), AMD, Broadcom, Microsoft, Micron, and TSMC. Combined with the IPO and underwriter overallotment (up to 123.75M additional shares), total proceeds could reach $84.6B–$93.7B — one of the largest capital raises in history.

The Numbers: Revenue, Losses, and Compute Obligations

Metric Value Context
Q3 2026 Revenue $16.96B 47% QoQ growth
Annualized Run-Rate ~$81B Consumption-based × 13 factor
Q3 Operating Loss Not disclosed Adjusted margin improving
Compute Obligations (Total) $901B $71B due by Sep 2027
Cash & Equivalents (Pre-IPO) ~$8.2B Per S-1
IPO Proceeds (Mid-Range) ~$61B Before $26B private placement
Total Capital After IPO ~$87–$94B Including private placement

The Compute Obligation Mismatch

The most striking figure in the S-1: $901 billion in committed compute spend — primarily to Google Cloud (TPUs), AWS (Trainium), and NVIDIA (GPUs) — with $71 billion due by September 2027. At the mid-range IPO proceeds (~$87B total with private placement), Anthropic covers roughly one year of compute commitments. The runway math requires:

  1. Revenue scaling from $81B run-rate to $200B+ by 2027
  2. Margin improvement from current negative territory to positive
  3. Continued investor appetite for follow-on offerings or debt

Anthropic's S-1 discloses adjusted gross margins of ~51% (improving from ~38% a year ago) but operating margins remain deeply negative due to R&D and compute depreciation. The path to profitability hinges on inference revenue outpacing training compute amortization — a bet on the "consumption model" where every API call generates margin.

The Supplier-Investor Circular Financing

Investor Private Placement Commercial Relationship
Google $10B Primary TPU cloud provider; TPU 8th-gen customer
NVIDIA $10B GPU supplier (H100/H200); neocloud partner
Microsoft Undisclosed Azure OpenAI alternative; enterprise channel
AMD Undisclosed MI300X supplier; ROCm ecosystem
Broadcom Undisclosed Networking ASICs for TPU/GPU clusters
TSMC Undisclosed Foundry for Google TPUs, NVIDIA GPUs, AMD MI300
Micron Undisclosed HBM supplier for all of the above

This creates a circular financing loop: suppliers invest in Anthropic at IPO valuation → Anthropic pays suppliers for compute → suppliers recognize revenue → suppliers' stock benefits. The S-1 flags this as a related-party risk: "Our cornerstone investors are also our primary vendors. Their investment decisions may be influenced by commercial interests rather than pure financial returns."

Valuation Context: Anthropic vs. OpenAI vs. Public Comps

Company Valuation Revenue (Annualized) Revenue Multiple Status
Anthropic (IPO top) $1.404T $81B 17.3× Pricing Nov 18
OpenAI (Oct 2026) $1.25T ~$50B 25× Private
NVIDIA (Nov 16 close) $2.8T $109B (Q3) 6.4× Public
Microsoft $3.1T $245B 12.7× Public
Google (Alphabet) $2.0T $340B 5.9× Public
Meta $1.3T $156B 8.3× Public

Anthropic's 17.3× multiple sits between pure-play SaaS (10–15×) and hype-phase AI (20–30×). The market is pricing in:

  • Consumption model scalability — API revenue grows with customer usage, not seats
  • Model efficiency gains — Fable 5.2 / Mythos 5.2 (launched Nov 9) deliver 30% faster inference at 20% lower cost
  • Enterprise adoption curve — 50%+ of revenue now from enterprise (vs. 30% a year ago)

The PBC Structure: Founders Control, Retail Gets 20%

Anthropic is a Public Benefit Corporation (PBC) with a dual-class structure:

  • Class F shares (founders + Long-Term Benefit Trust): 50.1% voting power
  • Class A shares (IPO buyers): Economic rights only, limited voting
  • Long-Term Benefit Trust: Elects board majority; mission-aligned governance
  • Retail allocation: Up to 20% of offering (vs. typical 5–10%)

The seven founders (Dario Amodei, Daniela Amodei, Jack Clark, Sam McCandlish, Tom Brown, Jared Kaplan, Chris Olah) each sell ~$250M in the offering. Post-IPO, they retain Class F control — retail and institutional Class A holders cannot outvote the Trust.

Risk Factors: 90 Pages of "What Could Go Wrong"

The S-1 devotes ~90 of 300 pages to risk factors — unusually high. Key themes:

  1. Compute dependency — "We cannot operate without Google Cloud TPUs and NVIDIA GPUs. Any disruption, price increase, or allocation reduction would materially harm our business."
  2. Model commoditization — "Open-weight models (DeepSeek, Llama, Qwen) may reach parity on key benchmarks, eroding API pricing power."
  3. Regulatory uncertainty — CA AG lawsuit (P6 Story 1), NY RAISE Act, EU AI Act, potential federal legislation.
  4. Talent concentration — Key researchers (Kaplan, Olah, Amodei) are "difficult to replace."
  5. Safety incidents — "A significant misuse or alignment failure could trigger regulatory bans, customer exodus, and reputational collapse."
  6. Circular financing — Cornerstone investors are also vendors (detailed above).

The Fable 5.2 / Mythos 5.2 Catalyst

Anthropic's November 9 model launch (Fable 5.2 general access, Mythos 5.2 vetted cyber/bio access) — covered in P6 Story 3 — is timed for the roadshow. The models deliver:

  • 30% faster output via speculative decoding and kernel fusion
  • 20% price cut — $8/$40 per M tokens (input/output) vs. $10/$50 for 5.1
  • Fast mode at 2.5× speed for latency-sensitive agentic workloads
  • Breakthroughs in coding (Terminal-Bench), math proofs (Fermat's Last Theorem formalized in 30 hrs), and robotics

This demonstrates the inference efficiency flywheel: better models → lower cost per token → higher margins → more compute budget → better models. The IPO pitch is that this flywheel is just starting to compound.

What the Market Will Watch on November 19

Signal Bullish If Bearish If
First-day pop >20% (SpaceX did 35%) <5% or break below $70
Volume >500M shares traded Thin volume, price supported only by underwriters
Retail participation 20% allocation filled, hold through day Retail flips, institutions only
Options flow Call skew, high OI at $80–$90 Put skew, hedging dominant
Analyst initiation Buy ratings, $100+ PTs Hold/Sell, "fully valued"
OpenAI reaction Accelerates IPO timeline Delays, cites "market conditions"

The Bigger Picture

Anthropic's IPO is not just a liquidity event — it's a referendum on the AI lab business model. The questions it answers:

  1. Can a pure-play model company sustain a >$1T valuation without owning the compute layer?
  2. Does the consumption model (API revenue) scale to $200B+ run-rate with positive margins?
  3. Will public markets accept PBC governance where mission control trumps shareholder votes?
  4. Is the supplier-investor circular financing sustainable or a conflict-of-interest time bomb?
  5. Does the $901B compute obligation represent moat or millstone?

Pricing on November 18, trading on November 19. The outcome sets the anchor for every private AI valuation — OpenAI, xAI, Cohere, Adept, and the next generation. If Anthropic trades at $1.4T, the AI boom has a public market benchmark. If it struggles, the "AI winter" narrative gains its first data point.


Related AIPress Coverage

Sources

  • TechCrunch (Nov 10, 2026): "Anthropic sets IPO price range of $70 to $78 per share"
  • CNBC (Nov 10, 16, 2026): IPO pricing, date, retail allocation details
  • The Guardian (Nov 10, 2026): "Anthropic targets valuation of up to $1.4tn in stock market flotation"
  • Reuters (Nov 10, 2026): "Anthropic eyes $1.4 trillion IPO valuation in test of Wall Street's AI fervor"
  • The Verge (Nov 10, 2026): "Anthropic targets $1.4 trillion IPO"
  • The Desk Research (Nov 10, 2026): Valuation arithmetic analysis
  • ABC News (Nov 10, 2026): "Anthropic says it could be worth as much as $1.4T"
  • StreetInsider/Reuters (Nov 10, 2026): Cornerstone investor details
  • Anthropic S-1 Amendment No. 2 (filed Nov 10, 2026): Full financial terms, risk factors, compute obligations

Building something with AI?

DevsIsle designs and ships AI systems, agents and integrations for teams that need it done properly.

Talk to our team →