Key Takeaways
- Mistral AI secured $830M in debt financing to purchase approximately 13,800 Nvidia GB300 GPUs for a Paris data center, representing potential chip revenue of ~$575M
- Starcloud closed a $170M funding round at $1.1B valuation to expand its satellite GPU infrastructure following successful H100 orbital deployment
- Nvidia introduced its Space-1 Vera Rubin computing module on March 16, designed to solve data transmission challenges in space-based computing
- Cloud infrastructure spending worldwide reached $110.9B in Q4 2025, climbing 29% year-over-year, with 27% expansion projected for 2026
- Broadcom (AVGO) declined 1.3% while AMD rose 1% during Monday’s opening session
Nvidia began Monday’s trading session with modest gains, lifted by two significant announcements in the AI infrastructure space.
The more substantial development emerged from Mistral AI, the French artificial intelligence firm, which disclosed an $830 million debt financing round — the company’s first — earmarked for constructing a data center facility in the Paris region. This installation will deploy 13,800 of Nvidia’s GB300 GPUs. Industry analysis from HSBC suggests each GB300 NVL72 rack carries a price tag near $3 million, pointing to approximately $575 million in potential chip revenue for Nvidia from this single project.
Mistral has declined to verify specific pricing details, and Nvidia maintains confidentiality around individual chip pricing. Yet the magnitude of this purchase aligns with patterns analysts have observed throughout recent quarters: demand for AI processing hardware continues at elevated levels.
Worldwide spending on cloud infrastructure totaled $110.9 billion during Q4 2025, marking a 29% jump compared to the prior year period, per data from research group Omdia. The firm anticipates another 27% growth trajectory through 2026 — the type of market momentum that has propelled Nvidia’s performance over the past 24 months.
Orbital Computing Enters the Picture
The day’s more unconventional development involves operations beyond Earth’s atmosphere. Starcloud, a company developing data center infrastructure in space, announced completion of a $170 million investment round valuing the business at $1.1 billion. The firm previously launched one of Nvidia’s H100 GPUs into orbit in late 2024 through its Starcloud-1 satellite — marking the first instance of AI model training conducted in space.
Starcloud now targets a second satellite deployment scheduled for later this year, equipped with a complete GPU cluster and the largest commercial deployable radiator system ever placed in orbit, delivering 100 times the computational capacity of its initial spacecraft.
Proponents of space-based data centers position them as potential answers to mounting challenges surrounding ground-based facilities — including power grid constraints, water consumption concerns, and opposition from local communities, all of which have complicated terrestrial construction efforts. Orbital platforms could harness solar power directly and eliminate conventional cooling requirements entirely.
Still, the engineering complexity and financial barriers remain substantial, and the market exists in nascent form.
Nvidia addressed this emerging sector two weeks prior. The company revealed its Space-1 Vera Rubin computing module on March 16 — specialized hardware engineered to tackle a fundamental challenge facing space-based systems: bandwidth limitations. Given the restricted data transmission capacity between satellites and ground stations, Nvidia’s new module performs processing at the point of data collection rather than transmitting raw information to Earth for analysis.
Industry Landscape and Valuation
Nvidia faces competition in the orbital computing sector. SpaceX, led by Elon Musk, has been associated with initiatives to establish solar-powered orbital data centers, potentially financed through a future public offering. Blue Origin, founded by Jeff Bezos, is pursuing regulatory clearance to deploy nearly 52,000 satellites equipped with AI computing capabilities.
Nvidia currently carries a forward price-to-earnings ratio near 21.4, below recent quarterly levels while still reflecting growth expectations. The company’s market capitalization exceeds $4 trillion.
Nvidia has yet to announce a delivery timeline for the Space-1 Vera Rubin module to commercial customers.
