An interactive model of what it costs to serve one token of a frontier LLM — and the gross margin that implies at today's prices. Defaults are grounded in public evidence (DeepSeek's serving disclosure, independent GB300/GB200 benchmarks, X-sphere analyses); every slider documents its sources. Drag things. Nothing here is Anthropic's actual ledger.
The claim under examination: Anthropic's marginal cost of serving a Claude token is so far below its API list price that the gross margin on serving is 90–95%. This is a claim about unit serving economics — one token, on a warm GPU, at scale — not about Anthropic's income statement (see §7 for why those diverge).
The loudest and most quantitative proponent is @teortaxesTex:
"math for DeepSeek Serving Opus is at most $4/1Mt. Anthropic is not retarded. They are fleecing you even on subscriptions, except maybe at 100% utilization" — Jun 28, 2026
"No, they'll just increase the batch size, have the same speed, and drive margins from 90% to 95%. You're welcome" — Jun 27, 2026
Earlier and more conservative from the same account: "if we exclude R&D and look at inference alone, Anthropic and OpenAI are making like 80% margins" (Mar 2025).
@zephyr_z9 (Zephyr) makes the same claim from the semis side: "At least xAI isn't juicing up the gross margins to 90%-95% and scamming consumers" (Jul 8, 2026) — while separately pegging Anthropic's company-level gross margin at ~70% with 15–20% FCF margin (Jun 24, 2026). Note those two numbers coexisting in one analyst's head: 90–95% on the unit, ~70% on the books.
A correction from the sweep: Jukan — identified by the Grok sweep as @jukan05, though GPT Pro's independent browsing instead surfaced posts under @Jukanlosreve (the two engines disagree; both are reported in §6) — was not found asserting 90–95% token margins. His widely-cited threads are about adjacent economics: GPU cluster leasing (xAI's ~220k GPUs at ~$2.60/GPU-hr ≈ $5–6B/yr; "Anthropic converts roughly $5B of spend into what it expects to be $15B of ARR", May 2026) and the memory-bound nature of inference ("Inference is memory", Jun 2026). His famous 90%+ margin posts are about SK hynix DRAM, not Anthropic.
The counter-camp: @fleetingbits — "we know approximately what frontier lab inference margins are; it's like 40-50%; it's been reported a bunch of times. anthropic labels cloud provider commissions as a sales and marketing expense; so the gross margins are mostly inference compute costs." And in the middle, Dylan Patel of SemiAnalysis: Anthropic's margin on an Opus API token is "north of 80%".
On March 1, 2025 (Open Source Week "Day 6"), DeepSeek published actual production serving statistics for V3/R1 on H800 clusters: 73.7k input / 14.8k output tokens per second per 8-GPU H800 node, 56.3% input cache-hit rate, ~$87,072/day GPU cost (at $2/H800-hr) against $562,027/day theoretical revenue at R1 list prices — the famous "545% cost-profit ratio," which is a markup; as a margin it is 84.5% (a distinction TeorTaxes himself insisted on). Caveats DeepSeek listed: much traffic was free web/app users and off-peak V3 pricing, so realized revenue was materially lower than the theoretical figure.
Why it matters for Anthropic: DeepSeek proved ~85% serving margins at prices 10–45× below Anthropic's list, on export-restricted hardware, in early 2025. Every input to that calculation has since improved: better chips (§4), better serving software, higher cache rates. The a-fortiori argument — if DeepSeek could do 85% at $2.19/Mtok output on H800s, what does $25/Mtok output on GB300s and TPUv7 imply? — is the strongest single piece of evidence in the bull case, and it survives scrutiny.
The recollection "Musk leaked that Opus is much smaller than expected" has the sign flipped on total parameters. What Musk actually posted (Apr 9, 2026): "0.5T total. Current Grok is half the size of Sonnet and 1/10th the size of Opus. Very strong model for its size." The community deduction (e.g.): Sonnet ≈ 1T, Opus ≈ 5T total parameters — Opus is big.
The "smaller than expected" intuition belongs to active parameters: TeorTaxes, observing Fable served at ~90 tok/s, concluded it had "shockingly FEW active parameters for what it was", and Zephyr puts OpenAI's frontier at "~100B active range" with Opus/Fable "the highest active parameters" among peers. For serving cost, active parameters are what matter (FLOPs/token ≈ 2 × active); total size mostly sets the HBM footprint. A 5T-total/~300B-active MoE is the shape our defaults assume — and the active number is the single most uncertain, most consequential slider on this page. Even the total is contested: the "incompressible knowledge probes" estimation method (arXiv:2604.24827) carries ±3× error bars per model, and a methodological re-analysis lands nearer 1.1T total for Opus 4.7. GPT Pro's sanity check cuts the other way: a dense 5T Opus would cost more than its own $25/Mtok list price to serve — so if Musk's number is right, heavy sparsity isn't optional, it's implied.
The researcher is @_xjdr; the product is ncode on the Noumena platform (code.noumena.com), served on GB300 NVL72 racks from Prime Intellect. His free-week postmortem (Jun 30, 2026) is the best public look at frontier-style serving on Blackwell Ultra:
"final GLM 5.2 served stats: ~12000 unique api keys, served ~300B tokens total, 232 tok/s/gpu output average, 431 tok/s/gpu output max sustained, 2.1 sec TTFT average (1M ctx), 81k tok average input size, 41% cache hit rate"
Configuration: 60× B300 ("15 trays"), bf16 attention with fp8 experts/KV ("virtually 0 measurable quality difference"), no MTP, no Eagle — a custom Rust stack. A third-party estimate on his thread put the implied cost at ~$0.35/Mtok in, ~$1.50/Mtok out at $6/GPU-hr. Note his 232 tok/s/GPU average is output-only bookkeeping on a workload with 81k-token average inputs at 1M context — the GPUs were mostly doing prefill; it is not comparable to SGLang's 12k tok/s/GPU throughput records (different model, batch regime, and MTP).
Adjacent results worth separating from the ncode story (GPT Pro initially conflated them, §6): GLM-5.2's architecture is public via Baseten's serving writeup — 744B total / 40B active, NVFP4-clean, 280+ tok/s/user on Blackwell (a per-user speed record, not aggregate throughput). And the headline GB300 aggregate number belongs to SGLang + NVIDIA serving DeepSeek-V4: 2,200 → 11,200 tok/s/GPU from April to June 2026 on the same racks — five-fold, from software alone.
Covered in §8 — the short version: heavy Claude Max users demonstrably extract 15–40× the subscription price in API-equivalent tokens, which is only economically sane if the marginal cost of those tokens is a small fraction of list price. The plans are themselves evidence for the bull case.
The calculator above prices a token from first principles:
output $/Mtok = ($/GPU-hr ÷ 3600 ÷ tokens/s/GPU) × 10⁶ ÷ utilization, with tokens/s/GPU = (dense FP8 FLOPS × precision factor × effective-MFU × interactivity) ÷ (2 × active params). Prefill (input) uses the same form at higher MFU; cache reads cost a few percent of prefill. The workload mix (input:output ratio, cache-hit rate) blends the three, and the same mix prices revenue at list minus cache/batch/negotiated discounts.
Calibration, not vibes: the effective-MFU defaults reproduce the three best public measurements within a few percent:
| Anchor (published) | Published | Model reproduces |
|---|---|---|
| DeepSeek disclosure, H800 decode (37B active, FP8) | 1,850 tok/s/GPU | 1,873 (effective MFU 7%) |
| DeepSeek disclosure, H800 prefill | 9,212 tok/s/GPU | 9,097 (MFU 34%) |
| vLLM GB200, R1 decode | 10,100 tok/s/GPU | ~10,140 (MFU 15%) |
| SGLang GB300 record, V4 Pro 1.6T FP4 + MTP | >12,000 tok/s/GPU | ~11,800 (MFU 17% × 1.85 FP4) |
Those single-digit decode MFUs are not a bug — decode is memory/interconnect-bound; this is why "inference is memory" and why bigger NVLink domains and HBM keep beating raw FLOPS.
| Platform | Dense FP8 (PF) | Dense FP4 (PF) | HBM | BW | TDP | Rental (Jul 2026) |
|---|---|---|---|---|---|---|
| H100 SXM (2022) | 1.98 | — | 80 GB | 3.35 TB/s | 700 W | $1.99–3.90/hr, spot ≈ $2.40 |
| H200 (2024) | 1.98 | — | 141 GB | 4.8 TB/s | 700 W | $2.45–4.50/hr |
| B200 / GB200 NVL72 (2024-25) | 4.5 | 9 | 186 GB | 8 TB/s | 1.0–1.2 kW | $3.50–6.00/hr; rack ≈ $3–3.5M |
| GB300 NVL72 "Blackwell Ultra" (2025-26) | 5.0 | 15 | 288 GB | 8 TB/s | 1.4 kW | $4–7/hr early; rack ≈ $3.5–4.5M |
| TPU v7 Ironwood (GA Nov 2025) | 4.61 | — | 192 GB | 7.37 TB/s | ~1 kW | undisclosed; Anthropic deal: up to 1M chips |
| Trainium2 / Trainium3 (GA Dec 2025) | 1.3 / 2.51 | — | 96 / 144 GB | 2.9 / 4.9 TB/s | ~0.5–0.8 kW | AWS-internal; Project Rainier = $11B, 2.2 GW, ~500k chips |
| Vera Rubin VR200 NVL144 (H2 2026) | ~3.5× GB200 | ~50 PF/pkg eff. | HBM4 | ~13 TB/s | ~1.8 kW | projection |
Measured, not marketing: SemiAnalysis InferenceX (the successor to InferenceMAX, the field's independent benchmark) finds the most-optimized GB300 NVL72 delivers ~17× the best H100 config in FP8 and ~32× in FP4 on DeepSeek R1 (Jun 27, 2026) — and, critically, that software alone was a 14× gain on the same silicon (baseline FP8 ~1k → wideEP+disagg ~8k → +MTP ~14k tok/s/GPU). SGLang's production numbers agree: 12k tok/s/GPU on V4 Pro 1.6T, 6.5× over B200 with Dynamo disaggregation. At rack level: an H100 rack two years ago did ~8.8k tok/s; a GB300 NVL72 does ~370k — 42× in two years (8× of it HBM growth).
But cost per token falls slower than throughput rises, because rental prices track capability: $2.40/hr H100 → $6/hr GB300 eats ~2.5× of the 17×. Net hardware-economics gain per generation in $/Mtok: H100→H200 ~1.2×, H200→GB200 ~2–3×, GB200→GB300 ~1.1–1.5× (GB300's edge is HBM capacity for reasoning/long-context, not FLOPS), Rubin ~2–3× again. Cumulative 2024→2027: roughly 6–25× cheaper per token at the hardware level, before model-side efficiency (sparser MoEs, MTP, quantization) which historically contributed as much again. This is why margins at constant list prices mechanically march toward 95%+ — and why in practice prices fall instead (Opus's 2025-11 cut to $5/$25 was 3×; fast-mode at $30/$150 shows how the latency premium is being monetized separately).
Verdict: the 90–95% claim is approximately right for what it actually claims — the marginal margin on list-priced API tokens — and 95% is not even the ceiling for output tokens on the newest hardware. But it is the top of a stack of margins that shrinks at every step toward the income statement.
My point estimates for mid-2026, at list prices, balanced latency, 50% fleet utilization, neocloud-level hardware costs:
Three things the bull case gets right: (1) the DeepSeek a-fortiori argument is sound and the only primary-source-grounded argument anyone has; (2) caching is a margin machine — cache reads bill at 10% of input price but cost ~1–5% of prefill to serve, so agentic traffic stays high-margin even at deep effective discounts; (3) each hardware generation adds margin at constant prices, and Anthropic sits on three platforms (GB-class NVIDIA, up to 1M TPU v7, ~500k Trainium) with pricing leverage none of its open-weight competitors have.
Three things it elides: (1) utilization — fleets are provisioned for peak; at 35% utilization my Opus median drops ~8 points; (2) the latency premium is not free — interactive serving at user-acceptable speed costs ~1.4–3× throughput-optimal serving (fast mode's 6× price exists for a reason); (3) subscriptions and whales — the plans hand tail users 15–40× their subscription in tokens (§8), and at marginal cost the top decile is genuinely underwater, which is margin leakage the API numbers never show.
Forward: on GB300→Rubin economics, at unchanged list prices, Opus-class output margins pass 92–96% by 2027; I expect Anthropic to keep converting the surplus into price cuts and reasoning-token volume instead of letting the sticker margin print, so the quoted marginal margin stays in the high-80s/low-90s while $/token falls another 3–5×. The interesting question isn't whether 90–95% is right — it's how long the industry can hold list prices this far above a falling cost floor with open-weight models (DeepSeek V4, GLM 5.2, Kimi) selling adequate quality at 10–20× less.
GPT-5.6 Pro's verdict: "Standard-list-price marginal serving margin: approximately 92–94% for Opus and 94–96% for Sonnet on a mature 2026 fleet." And: "Is 95% an upper bound? No." — Sonnet at normal $3/$15 pricing exceeds it (96.3% on output), Opus on strategic TPU contracts exceeds it (97.2%), and the electricity-only marginal margin is >99.8%. But 95% is "not conservative for every token": at public cloud rates the same math yields only 55–85%.
Method. It priced an "economic marginal cost" — accelerator opportunity cost at Anthropic-scale contract rates, 75% fleet utilization, +10% CPU/network/reliability overhead — explicitly not electricity-only and not full COGS. Architecture: Opus ≈ 5T total / 300B active (independently landing on the same activation ratio as §5, ~6%, matching DeepSeek's 5.5% and GLM-5.2's 5.4%), FLOPs/token = 2.3 × active. Fleet blend: 40% TPU v7, 25% GB300, 15% GB200, 15% Trainium2, 5% H200.
Central estimates (Opus, $/Mtok serving cost, input → output):
| Basis | Input cost → margin @ $5 | Output cost → margin @ $25 |
|---|---|---|
| Fleet blend (central) | $0.47 → 90.6% | $1.68 → 93.3% |
| GB300 NVL72 only | $0.31 → 93.8% | $1.44 → 94.2% |
| TPU v7 at est. strategic rate (~$1.60/hr per SemiAnalysis) | $0.24 → 95.1% | $0.70 → 97.2% |
| H100/H200 (economic rates) | ~$1.2 → ~76% | ~$4.9 → ~80% |
| Public cloud rates (AWS H100 $5.19/hr … B300 $14.04/hr) | margins collapse to 55–85% — "why both sides of the online argument can produce superficially plausible calculations" | |
Sensitivity it flagged as dominant: active parameters. Opus at 150B active → 95–97% margins; at 600B active → 81–87%. Identical conclusion to §5: the whole debate compresses into one undisclosed number.
New evidence it surfaced (not in the X sweep): PitchBook/Morningstar's June 2026 estimate of Anthropic gross margin ≈ 44%, with compute spend of $0.71 per revenue dollar in Q1 falling to a projected $0.56 in Q2 — the freshest business-margin datapoint anyone has; GLM-5.2's actual architecture (744B total / 40B active) via Baseten's "world's fastest GLM-5.2 API" writeup (280+ tok/s/user on Blackwell NVFP4); and the SGLang+NVIDIA GB300 progression 2,200 → 11,200 tok/s/GPU between April and June 2026 — a 5× gain from software alone on fixed hardware, its exhibit for why "serving software is a first-order economic variable." Its illustrative bridge from 93.5% marginal margin down to the ~44% reported gross margin allocates: −12 pts price realization, −8 peak provisioning, −7.5 cloud markup/older hardware, −9 free traffic + subscription whales, −13 CPU/network/retries/safety/internal inference.
Attribution divergence, preserved: GPT Pro could not locate "Numina Code" and concluded the recollection conflated the SGLang/NVIDIA DeepSeek-V4 result with Baseten's GLM-5.2 work. The Grok X sweep, however, found the actual account — @_xjdr's ncode/Noumena, with primary post URLs (§2c) — while GPT Pro's browsing surfaced hardware-cost posts under the handle @Jukanlosreve, contradicting Grok's @jukan05 identification. Where the two engines' account attributions conflict we show both; the direct post links in §1–2 are the ground truth we could verify.
Where GPT Pro and I (Claude) differ — and why it's smaller than it looks: its central Opus numbers run ~10 points above my rental-based median (93.3% vs ~80% blended). The gap is almost entirely procurement assumptions, not physics: it prices hardware at Anthropic's negotiated strategic rates (TPU ~$1.60/hr) and 75% utilization; my median uses open-market neocloud rental at 50% utilization. Run my model in owned-TCO mode at its utilization and the two converge at ~93% — which is exactly what the "SemiAnalysis / Dylan Patel" and "GPT-5.6 Pro fleet model" presets above let you reproduce. The honest summary: the marginal margin at list price is 78–85% if you think Anthropic pays market rates for compute, 90–95% if you think it pays hyperscaler-partner rates. Everyone agrees on the token math; the argument is about the invoice.
Reported/leaked figures for Anthropic's business-level gross margin paint a different picture from the unit math. The Information (Jan 2026, from people with knowledge of its financials): Anthropic lowered its 2025 gross-margin projection to 40% (down 10 points from the earlier 50% plan) because inference costs on Google/Amazon servers ran 23% higher than anticipated; including free-tier chatbot inference the margin would be ≈38%. Same reporting: 2025 revenue ≈ $4.5B (~12× 2024's $381M), 86% of it API. SemiAnalysis put the 2024 accounting gross margin at −94% (yes, negative). Zephyr's mid-2026 read is ~70% GM with 15–20% FCF margin; a detailed July 2026 thread claims quarterly gross profit swung from −$55M to ~$453M with inference cost/token down ~40× since early 2024 (@IvanaSpear). The freshest independent estimate is PitchBook/Morningstar (Jun 2026): gross margin ≈ 44%, with compute spend of $0.71 per revenue dollar in Q1 2026, projected $0.56 in Q2. The trajectory — deeply negative (2024) → ~40% planned (2025) → ~44% estimated (mid-2026), with bulls like Zephyr already seeing 70% — is the margin ladder being climbed in public, and The Information ran a follow-up on why the labs kept missing their own gross-margin forecasts.
The bridge from ~90% unit margins down to ~40–70% books, roughly in order of size:
| Bridge item | Mechanism | Rough magnitude |
|---|---|---|
| Cloud markup | Anthropic buys most compute from AWS/GCP, who take their own margin; some channel revenue also shares 30–40% with the clouds (Bedrock/Vertex economics). DeepSeek owns its H800s; Anthropic mostly doesn't own its fleet. | −10 to −20 pts |
| Utilization & peak provisioning | Capacity sized for Monday-morning peak; the unit math assumes a busy GPU. 35% vs 70% utilization is a 2× on cost. | −5 to −15 pts |
| Subscription over-consumption | Flat-fee plans where the tail extracts 15–40× the fee (§8); weekly caps (Aug 2025) exist to clamp exactly this. | −5 to −10 pts |
| Free tier & internal inference | claude.ai free traffic, evals, RL/synthetic-data generation all burn serving compute against zero revenue (some booked as R&D, treatment varies). | −5 to −10 pts |
| Discounts & mix | Enterprise/committed-use discounts, 50%-off batch tier, cache-read-heavy agentic mix pulling realized $/Mtok toward $0.99 observed vs $5/$25 sticker. | −3 to −8 pts |
| Accounting choices | What lands in COGS vs S&M vs R&D differs by lab — fleetingbits' point that Anthropic books cloud commissions under sales & marketing cuts the other way, flattering GM. | ± |
Training compute, the famous money pit, sits below gross margin in R&D — it explains why the company loses money overall (training reportedly falling from 400%+ of revenue toward ~36% by end-2026 per the IvanaSpear thread), not why gross margin is below the unit margin.
The best single investigation found: ksred's Claude Code pricing guide — longitudinal tracking of ~10B tokens over 8 months: ~$15,000 at API list prices against ~$800 of Max subscription fees (~19×). The canonical X-thread calculation is @melvynx (Jan 2026): $50 of tracked usage = 6% of a week's Max-20x allowance ⇒ ~$800/week ⇒ ~$3,200/month of API-equivalent capacity for $200. Corroborating longitudinal data points: 14B+ tokens / $7.5k ccusage in 2.5 months, $8,030 in 4 weeks, 8.6B tokens / ~$8.5k over 60 days (~21× leverage), and the pathological single-session case: $4,243 at token rates in one session on a $100 plan.
Two readings, both correct: (a) at list prices these plans are massively underwater — Anthropic sells $3,200 of sticker for $200; (b) at marginal cost (this page's subject) $3,200 of API-equivalent usage costs perhaps $300–900 to serve depending on mix — so the median subscriber is comfortably profitable and only the whale tail is subsidized. That tension is exactly why weekly limits appeared (Aug 2025) and why the 5-hour windows keep being retuned (May 2026 doubling). The plans are a price-discrimination scheme that only works because marginal cost per token is tiny — the subscription card in the calculator lets you find the exact break-even.
Method note: X posts located and quoted via a Grok 4.5 agent sweep (2026-07-09); hardware/TCO data via Parallel deep research + Exa across SemiAnalysis, MLCommons, NVIDIA/Google/AWS primary pages; cross-model verification via an independent GPT-5.6 Pro research run. Attribution corrections (Jukan's handle, the Musk-leak direction, ncode vs "Numina") were made against primary posts, not memory. Model sizes for closed models remain estimates — the calculator exists so you can disagree with a slider instead of a reply-guy.