NVIDIA Invests in Cloverleaf for AI Data Center Sites

According to the PR Newswire release, Cloverleaf will apply the NVIDIA DSX platform earlier in facility design so customers can evaluate site, power, cooling, computing, and facility tradeoffs together. The release describes the objective as building AI factories that produce more useful AI output within available power, water, and grid constraints.

Data Center Dynamics describes DSX as NVIDIA's reference design for data centers that support its latest CPU-GPU racks, Spectrum-X Ethernet networking, and storage. That makes the agreement more than a land-development investment: Cloverleaf's site-selection workflow is being linked to a vendor-specific infrastructure design framework.

For engineering teams, early coordination of electrical capacity, cooling design, rack configuration, and networking can reduce the risk that a site is physically available but cannot support the intended compute deployment. Comparable large-scale projects often encounter these constraints before hardware procurement, particularly where grid interconnection timelines and water availability differ by location.

The companies' announcement identifies powered, shovel-ready sites as critical to bringing new capacity online quickly. The issue is especially relevant for dense AI deployments, where the facility's usable compute capacity depends on the interaction of grid power, cooling systems, and the rack-level design rather than on land availability alone.

Data Center Dynamics reported that Cloverleaf has sold projects capable of supporting more than 7 GW to data center developers and claims a pipeline exceeding 10 GW. It also reported that projects include Wisconsin sites for Oracle and OpenAI. Those figures are Cloverleaf-reported project capacity, not an announcement of installed AI compute capacity or operational NVIDIA systems.

Reuters reported that the partnership follows NVIDIA's recent investment in SB Energy, a SoftBank-owned data center developer. TechCrunch reported that NVIDIA also committed $1.5 billion to the OpenAI-linked Ohio project associated with SB Energy. Data Center Dynamics further reported that NVIDIA took a 20% stake earlier in 2026 in power developer Lancium, which is connected to the Abilene, Texas campus operated by Crusoe for Oracle and OpenAI.

The Cloverleaf agreement does not disclose deployment dates, GPU quantities, customer contracts, or a valuation for the investment. It therefore offers no direct basis for estimating near-term hardware demand or new cloud capacity.

Still, public reporting places the transaction within a broader infrastructure pattern: AI hardware suppliers are increasingly participating in the power, land, and data-center development layers needed to deploy high-density systems. For ML platform and capacity-planning teams, that pattern increases the importance of facility-level variables, including power delivery, cooling architecture, network topology, and interconnection schedules, alongside accelerator availability.