What happened
Rimac Technology, a Tier‑1 automotive battery and power‑train supplier, and Dubai‑based AI‑ provider ECOBLOX announced a Joint Development Agreement to create a Modular Data Center (MDC) for high‑performance AI workloads. The MDC will integrate Rimac’s UNI battery modules, battery‑management system (BMS), and high‑voltage power electronics (400 VDC/800 VDC) into ECOBLOX’s pre‑fabricated, air‑ and liquid‑cooled enclosure, DCIM platform, and GPU‑as‑a‑Service (GPUaaS) software stack. The first unit is scheduled for a live demonstration at Rimac’s campus near Zagreb, Croatia, in November 2026, where it will serve as a showcase for customers and partners.
The joint press release, dated October 7, 2026, outlines a Joint Development Agreement between Rimac Technology and ECOBLOX. Rimac will supply its UNI battery modules—originally designed for high‑performance electric vehicles—and associated power‑electronics capable of operating at 400 VDC and 800 VDC. ECOBLOX will incorporate these components into its modular data‑center architecture, which includes pre‑engineered enclosures, cooling, UPS, and a proprietary Data Center Infrastructure Management (DCIM) platform that aggregates telemetry from the battery management system.
ECOBLOX describes its offering as a turnkey, exascale‑ready AI solution that can be delivered in six to eight months. The company positions itself as an NVIDIA Preferred Partner, handling everything from hardware procurement to managed services. The integration of Rimac’s battery technology is intended to “remove grid constraints” and enable deployments in locations with limited power infrastructure.
The first modular data center will be installed at Rimac’s campus near Zagreb, Croatia, in November 2026. This site will act as a live demonstration platform for prospective customers and partners, allowing them to evaluate the combined system’s performance, reliability, and scalability.
Why it matters
The partnership brings automotive‑grade energy storage and power‑conversion technology to AI , addressing two persistent bottlenecks in AI data‑center deployment: power density and cooling efficiency. By high‑voltage battery packs and a unified BMS directly into the modular enclosure, the solution promises longer uninterrupted operation, reduced reliance on external grid capacity, and faster site‑to‑service timelines (six to eight months). If the pilot proves successful, cloud providers, sovereign entities, and large enterprises could deploy exascale‑ready AI clusters in locations with limited grid infrastructure, potentially reshaping the economics of AI compute in regions where power availability has been a limiting factor. The collaboration also illustrates a broader trend of cross‑industry technology transfer—automotive power‑train expertise being repurposed for data‑center power delivery.
Power density and cooling are two of the most costly and time‑consuming aspects of scaling AI . By leveraging automotive‑grade battery packs, the solution can provide high‑energy storage in a compact footprint, reducing the need for large, dedicated power substations.
The integrated BMS feeding real‑time data into ECOBLOX’s DCIM platform enables automated alerts and analytics, potentially improving operational efficiency and reducing downtime compared with legacy data‑center monitoring systems.
If the modular approach proves reliable, it could accelerate AI infrastructure rollout in emerging markets or remote locations where extending the electrical grid is prohibitive, thereby expanding the geographic reach of AI services.
The partnership signals a growing convergence between the automotive electrification sector and AI , suggesting that other high‑voltage, high‑density technologies may soon be repurposed for data‑center use.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
What should a useful AI forecast state?
What to watch next
Key indicators to monitor include: (1) performance data from the Zagreb pilot, especially uptime, thermal metrics, and power‑efficiency versus traditional data‑center designs; (2) customer uptake beyond the demonstration, particularly any contracts with cloud providers or government agencies; (3) regulatory or safety approvals required for deploying automotive‑grade battery systems in data‑center environments; and (4) pricing and financing models, which have not been disclosed and will affect the solution’s competitiveness against conventional power‑distribution architectures.
Performance metrics from the Zagreb pilot, including power‑efficiency (kW per GPU), thermal headroom, and backup duration, will be the first concrete evidence of the solution’s value proposition.
Announcements of commercial contracts or pre‑orders from cloud providers, sovereign entities, or large enterprises will indicate market acceptance.
Regulatory scrutiny around the deployment of automotive‑grade batteries in data‑center environments could affect rollout speed, especially concerning fire safety and grid interconnection standards.
Pricing, financing, and service‑level agreements have not been disclosed; these details will be critical to assess whether the solution can compete with traditional power‑distribution and UPS systems.