Model

Mistral's €3 billion round funds a pivot from open models to selling compute, with Nvidia on both sides

Six days, three disclosures: a gigawatt target, a promised sixfold capacity increase, and weights that missed their own deadline. Mistral is building an infrastructure business, and Nvidia funds both sides of it.

Richard TangRichard Tangthree-minute read
Share
Mistral AI co-founder and CEO Arthur Mensch meets Prime Minister Keir Starmer at 10 Downing Street
1 / 7Slide 1 of 7
Arthur Mensch, Mistral's co-founder and CEO, at 10 Downing Street in January 2025. He now argues Europe's €260 billion annual cloud and software bill should be redirected to homegrown infrastructure.

Mistral$14B — Mistral, private, latest valuation $14B is no longer pitching itself as Europe's open-model lab. In six days it disclosed a gigawatt-of-compute ambition 2, a promised sixfold capacity increase with another raise to pay for it 3, and still had not shipped the open weights it promised for the end of October 4. That combination points to a different business: infrastructure, where the model is the way to sell compute, agents and custom training to Airbus, BMW and ASML+3.91% — ASML, up 3.91 percent today, the industrial customers Mistral named at its Paris summit in May 1.

Arthur Mensch, Mistral's co-founder and CEO, made the argument himself in the Frankfurter Allgemeine Zeitung on October 24: European companies pay more than €260 billion a year to American cloud and software providers, and he wants that bill redirected 3.

The money is following the infrastructure story

Mistral's €3 billion Series D in September, at a valuation above €21 billion, was Europe's largest round of the period 1, and Tracxn, a funding-data tracker, shows late-stage rounds absorbing 53% of Europe's $61.9 billion in tech funding from about 9% of deals 6. Capital is concentrating on the company that owns the physical layer.

Selling the weights and selling the compute pull against each other

The tension is that Mistral sells two opposite things. One is Large 4, a trillion-parameter model whose weights customers could download and run without Mistral 2. The other is Mistral Compute, the infrastructure sale that needs them to stay.

On October 26 no source checked had confirmed the weights were live, and Mistral had not named their license 4. If Mistral follows rivals such as MiniMax, whose open weights carry conditions including prior authorization above $20 million in annual revenue 7, every tightening of the license makes the compute sale easier and the "open" claim weaker.

The price reinforces the point. Artificial Analysis, an independent benchmarking firm, puts the cost of running its Intelligence Index on Large 4 Preview at $1.13 per task, more than four times the cost of similar-intelligence open-weights models 8. That premium is only defensible if what the customer buys cannot be swapped out: European data residency, air-gapped deployment, custom training through Forge 1.

Mistral's input price sits far above its open-model rivals

00.511.5Mistral Large 41.36Nemotron 3.5 Super VL0.23MiniMax M3.1 Flash0.15
Data
Value
Mistral Large 41.36
Nemotron 3.5 Super VL0.23
MiniMax M3.1 Flash0.15
Listed public-preview or published API rates, US dollars per million input tokens. Mistral from its preview listing via shattered.io; Nemotron from Artificial Analysis via Trending Topics; MiniMax from its rate card via OrcaRouter.4,5,7

Nvidia owns the plumbing either way

Nvidia+5.22% — Nvidia, up 5.22 percent today, the chipmaker, supplied the 3,800 Grace Blackwell GPUs that trained Large 4 in Mistral's own European data centers 2. It is also co-building an open-weight model with Mistral in the Nemotron Coalition 2, and selling its own Nemotron 3.5 Super VL at roughly $0.23 per million input tokens against Mistral's $1.36 45. It has separately agreed to buy Hugging Face, the open-model platform, for $12.9 billion 6. Nvidia funds the challenger and owns the plumbing either way.

The next test is not quality: Large 4 already holds 38 on Artificial Analysis's Intelligence Index, which the firm calls the most intelligent model from outside the US and China 8. What to watch is the license the weights ship under, and whether Airbus and BMW deploy the stack in production rather than pilot. If the deployments come, the sixfold build-out Mensch promised has a paying customer. If they stall, Europe will have funded a data-center operator with a model nobody had to buy.

Get Model by email, free

The AI labs, their models, and the business behind them. In our daily email, every weekday.

Loading NVDA

More about NVIDIA

NVDAFiscal Q2 2027
Fiscal Q2 2027 · quarter to 26 Jul 2026 · changes vs a year earlier

Revenue rose 106% to $96.2B, 92.5% of it from Data Center, “driven by the ramp of our Blackwell Ultra infrastructure”. Gains on equity stakes of $7.8B lifted net income to $59.7B.

Revenue$96.2B▲ 106%
Gross margin75.0%▲ 2.6 pts
Operating margin66.2%▲ 5.4 pts
Net income$59.7B▲ 126%
Hyperscale$48.7B▲ 102%AI clouds & enterprise$40.3B▲ 138%Data Center$89.0B▲ 117%Edge Computing$7.2B▲ 27%Revenue$96.2B▲ 106%Gross profit$72.1B75.0% marginCost of revenue$24.1BOther income$7.8Bequity-stake gainsOperating income$63.7B66.2% marginOperating expenses$8.4BNet income$59.7B▲ 126%Tax$11.8B16.5% rateR&D$7.1BSG&A$1.4B
Show as a table
LineValue
Revenue$96.2B
Gross margin75.0%
Operating margin66.2%
Net income$59.7B
Hyperscale$48.7B
AI clouds & enterprise$40.3B
Data Center$89.0B
Edge Computing$7.2B
Gross profit$72.1B
Cost of revenue$24.1B
Other income$7.8B
Operating income$63.7B
Operating expenses$8.4B
Tax$11.8B
R&D$7.1B
SG&A$1.4B

Reader comments

0 comments

    Model

    Anthropic hired a UK lobby firm the week it priced its $1.4 trillion IPO

    The same week Anthropic set its $70-to-78 IPO range, the UK lobbyist registry showed it paying for political access in three countries. Goldman's $685 billion revenue case runs on data centers that voters and regulators must permit, and buyers fund that insurance without seeing its cost.

    Model

    Mistral built its enterprise stack in eight days and priced none of it

    Between October 29 and November 5 the Paris lab published its biggest model's weights, expanded its agent platform and bought a fifth startup, all funded by a September mark of €21 billion. Not one piece of the stack carries a published price, while free Chinese open-weight models still outscore it.

    Model

    xAI's Grok 4.8 entered training 57 days late with no price card; Anthropic pays SpaceX $1.25 billion a month

    The 57-day wait bought a serving stack xAI wrote itself, and the delay ends with no scores and no price. For a SpaceX shareholder, the unpublished price card for 4.8 is the next receipt.

    Model

    Meta's Muse Ember costs $0.14 a task, a third of Claude Haiku, as Anthropic prices its IPO

    Meta's smallest model does the work agents repeat all day at $0.14 a task, matching Anthropic's Haiku at a third of the cost. Anthropic prices its IPO this week on about $81 billion of run-rate token revenue, and the floor under that business just moved.

    Model

    Anthropic cuts Fable 5.2 prices 20% on IPO roadshow eve, and Google and Broadcom shareholders carry the bill

    Anthropic discounted its flagship the day its IPO quiet period ended, for buyers who had already stopped climbing. The compute bill behind the discount is owed chiefly to Google and chipmaker Broadcom, which leaves their shareholders carrying the exposure.

    Model

    AfterQuery's agent training data lifted OpenAI's small model from 3.1% to 17% on Terminal-Bench

    The November 4 run, showcased on Thinking Machines' Tinker, sits beside rival Mercor's proof on the same page, giving labs a direct comparison of measured benchmark uplift for the first time. It is the claim beneath AfterQuery's reported $3.2 billion valuation, now competing in public for labs' data budgets.

    Get the newsletter

    The day's AI and market news, in your inbox every weekday.