GPT-56T 861 —
MUSE-SPK 837 +0.2%
GPT-56SC 790 -4.6%
GLM-5 781 -0.4%
CL-OP55X 780 -5.1%
GROK-46H 780 -5.1%
QWEN-38X 748 -9.2%
GPT-6A 743 -9.4%
KIMI-K3X 742 —
CL-FAB5H 698 -6.1%
CL-OP5H 675 -6.2%
GEM-38FH 672 -0.7%
CL-OP5X 670 -5.5%
CL-OP55H 668 —
CL-OP46H 657 -5.9%
CL-OP47H 648 -6.1%
GPT-56S 618 -0.6%
GEM-37FH 610 -7.2%
GEM-36FH 593 —
CL-OP48H 588 —
CL-OP47 581 -0.2%
GEM-35FH 580 —
GPT-55H 541 -7%
INKL 531 —
GEM-31P 512 -0.2%
CL-OP46 498 +0.4%
GEM-3P 498 -0.2%
CL-OP48 492 +0.4%
GPT-52 464 —
GPT-55 423 —
GPT-56T 861 —
MUSE-SPK 837 +0.2%
GPT-56SC 790 -4.6%
GLM-5 781 -0.4%
CL-OP55X 780 -5.1%
GROK-46H 780 -5.1%
QWEN-38X 748 -9.2%
GPT-6A 743 -9.4%
KIMI-K3X 742 —
CL-FAB5H 698 -6.1%
CL-OP5H 675 -6.2%
GEM-38FH 672 -0.7%
CL-OP5X 670 -5.5%
CL-OP55H 668 —
CL-OP46H 657 -5.9%
CL-OP47H 648 -6.1%
GPT-56S 618 -0.6%
GEM-37FH 610 -7.2%
GEM-36FH 593 —
CL-OP48H 588 —
CL-OP47 581 -0.2%
GEM-35FH 580 —
GPT-55H 541 -7%
INKL 531 —
GEM-31P 512 -0.2%
CL-OP46 498 +0.4%
GEM-3P 498 -0.2%
CL-OP48 492 +0.4%
GPT-52 464 —
GPT-55 423 —
← Back to feed

Mistral Raises $830M for 13,800 Nvidia GB300 GPUs, Hits $14B Valuation — and xAI Is Calling

Mistral AI closed $830 million from a consortium of European banks to fund a Paris data center filled with 13,800 Nvidia Grace Blackwell GB300 GPUs. The facility delivers 44 megawatts of compute capacity and is expected to be operational in Q2 2026. The raise values Mistral at $14 billion.

The lenders: Bpifrance, BNP Paribas, Crédit Agricole CIB, HSBC, La Banque Postale, MUFG, and Natixis Corporate and Investment Banking. No US venture capital.

The Sovereignty Pitch

CEO Arthur Mensch has a simple argument: “Europe needs an ambitious AI cloud infrastructure and an independent AI stack.” The Paris data center is the first of what Mistral is positioning as a 200 MW European compute footprint, with a second facility already announced in Sweden.

The customer base for this pitch is real. Governments, regulated enterprises, and research institutions that cannot route sensitive workloads through US hyperscaler infrastructure represent a captive market that OpenAI and Anthropic cannot credibly serve. Mistral’s open-weight model distribution means enterprise customers can fine-tune and self-host — making Mistral the only major frontier lab that offers a genuine on-premises path.

The xAI Complication

While Mistral raises European capital on a sovereignty message, Business Insider reported that xAI has held discussions with Mistral and Cursor about a potential three-way partnership. The exact structure is not public, but the logic is legible: xAI owns Colossus, one of the largest AI training clusters in the world, but lacks competitive models and distribution. Mistral has models and a clear regulatory advantage in Europe. Cursor has developer distribution.

The discussions come after xAI already announced a separate deal with Cursor in which SpaceX secured the right to acquire Cursor for $60 billion, with $10 billion owed in compute access if the acquisition does not proceed. Adding Mistral to that triangle would give xAI European regulatory cover and an open-weight model line — something the Grok models currently lack.

Whether Mistral’s investors, who funded the company specifically on a non-US-dependency pitch, would accept a partnership with Musk-owned infrastructure is an open question.

Market Position

Founded in 2023, Mistral is the only company among the eight or so frontier AI labs that is not American. Its open-weight models — particularly Mistral Large and the MoE variants — have meaningful enterprise adoption through Microsoft Azure, AWS, and Google Cloud partnerships, as well as through self-hosted deployments.

At $14 billion, Mistral is valued at roughly half of what Anthropic was raising at in early 2026 and well below OpenAI’s $852 billion. The bet is that the addressable market for sovereign, non-US AI infrastructure is large enough to support a premium valuation independent of the frontier model race — and that regulators across Europe, the Middle East, and Asia will increasingly mandate it.

Key Numbers

  • Raise: $830 million in debt financing from seven European banks
  • Paris facility: 13,800 Nvidia GB300 GPUs, 44 MW, Q2 2026 operational
  • Europe target: 200 MW total compute capacity
  • Valuation: $14 billion
  • Forge platform: launched March 2026, enables enterprise-grade model training on proprietary data