TensorX, an Irish‑based AI infrastructure startup, announced an €8 million seed round in June 2026 that is being led by Darius Cubed Ventures. The financing is earmarked for the acquisition of NVIDIA’s latest Blackwell GPUs—including the cutting‑edge B300 chips—and for the rapid deployment of a privacy‑first AI inference service that runs exclusively on dedicated hardware located in Europe. By keeping the compute and data processing strictly within European borders, TensorX is targeting highly regulated sectors—such as finance, healthcare, and legal services—where data sovereignty is not just a preference but a regulatory requirement under GDPR, the forthcoming EU AI Act, and the constraints imposed by the U.S. CLOUD Act. The company’s vision is to create a pan‑European sovereign AI stack that can scale from its initial sites in Dublin and Helsinki to additional data‑centre hubs across the United Kingdom, Germany, France and the Nordic region, thereby offering enterprises a compliant alternative to the dominant U.S. cloud providers.
TensorX Launches with €8M Seed Funding Led by Darius Cubed Ventures
The €8 million seed round, disclosed in June 2026, provides the capital needed to purchase a first wave of NVIDIA Blackwell GPUs. Co‑founder Shane Morton has personally committed €4 million of the total, with €2 million of GPUs already delivered to the company’s data‑centre locations and a further €2 million on order. TensorX is a participant in NVIDIA’s Inception programme, which gives it early access to the latest hardware and technical support, and it has secured a strategic partnership with Dell for the physical sourcing and integration of the GPUs.
Beyond the initial hardware spend, the round also funds operational expansion. TensorX is in “advanced talks” to secure a multi‑year financing facility that will underwrite the rollout of additional GPU capacity across a planned network of sites in Ireland, the United Kingdom, Germany, France and the Nordics. The company’s roadmap envisions a modular, region‑based architecture that can be scaled quickly in response to demand spikes, while preserving the same zero‑data‑retention guarantees at every location.
The seed round’s lead investor, Darius Cubed Ventures, highlighted the market opportunity: “Demand for sovereign AI infrastructure is outpacing supply across Europe. We’re seeing it directly from enterprises in Germany, France, the Netherlands and the Nordics. Our €8 million investment is the opening move. There is a far bigger build‑out to come, and the infrastructure partnerships we have in Ireland mean we can move at the speed this market demands.” This sentiment reflects broader industry data—Accenture reports that 62 % of European organisations now actively seek sovereign AI solutions, a figure that rises to 76 % within the banking sector, while Gartner forecasts that by 2030, 75 % of AI workloads in Europe will be handled by local providers.
Architecture Focused on Zero Data Retention
TensorX’s technical architecture is built around a strict zero‑data‑retention policy. The platform runs open‑source AI models—such as LLaMA, Falcon and other community‑driven frameworks—exclusively on NVIDIA Blackwell GPUs that are housed in two purpose‑built data‑centres: one in Dublin, Ireland, and the other in Helsinki, Finland. Because the inference compute never leaves these facilities, the service does not store, log, or reuse any input data after a request is completed. This design directly mitigates the legal exposure created by the U.S. CLOUD Act, which can compel U.S.-headquartered cloud providers to hand over customer data irrespective of its physical location, often under gag orders that prevent the data owner from being notified.
By keeping both the hardware and the data processing within the European Economic Area, TensorX offers enterprises a clear compliance pathway under GDPR and the upcoming EU AI Act, which will impose stricter obligations on AI system transparency, risk management and data handling for regulated sectors. The company’s “chips‑up” philosophy—building the stack from the hardware layer upward—means that every layer of the service, from the GPU firmware to the orchestration software, is under the direct control of TensorX and its European partners. This contrasts with the typical public‑cloud model where the underlying infrastructure is owned by U.S. corporations, creating a jurisdictional mismatch for European customers.
Early Revenue and Target Customer Segments
Within weeks of the platform’s commercial launch, TensorX began generating revenue from three clearly defined customer cohorts, confirming that market demand aligns with the company’s sovereign‑AI thesis.
- Large regulated enterprises – Financial institutions, healthcare providers and law firms are signing multi‑year sovereign‑infrastructure contracts that guarantee that all AI inference runs on European soil. These contracts often include service‑level agreements (SLAs) that specify latency, uptime and data‑privacy metrics required for mission‑critical workloads.
- Partnership channels – Platforms such as OpenRouter are integrating TensorX’s compute offering into their developer‑facing APIs, routing AI‑model inference requests from third‑party developers onto the European‑based GPU fleet. This channel expands TensorX’s reach without the need for direct sales effort for every developer.
- SMEs building AI products – Small‑to‑medium enterprises, including APEX:E3, TradeLocker and Cor Prime, are leveraging the platform to embed AI capabilities—such as coding assistants, risk‑analysis tools and document‑review bots—directly into their products.
Customers consistently cite the platform’s zero‑data‑retention guarantee as a decisive factor. Usman Khan, founder of APEX:E3, explained, “TensorX is simply the only platform we trust with our most sensitive data which we manage on behalf of regulated institutional financial services companies.” The company also reports a surge of inbound interest from Germany, France, Denmark and the Netherlands, driven by the anticipation of the EU AI Act’s tighter compliance regime.
In total, early revenue streams are already covering a portion of the operational costs associated with the initial GPU deployment, providing a proof point that the business model can be sustainable while the company continues to scale its hardware footprint.
Key Takeaways
- TensorX secured €8 million in seed funding, led by Darius Cubed Ventures, to acquire NVIDIA Blackwell GPUs for a European sovereign AI inference platform.
- The service runs on dedicated hardware in Dublin and Helsinki with a zero‑data‑retention policy, addressing GDPR, EU AI Act and U.S. CLOUD Act concerns.
- Revenue is already being generated from regulated enterprises, partnership channels, and SMEs building AI products, with early growth driven by inbound demand from Germany, France, Denmark and the Netherlands.
TechInsyte's Take
TensorX’s funding and hardware commitment illustrate a concrete response to rising demand for sovereign AI infrastructure in Europe. While the company has secured early customers, its ability to scale capacity across multiple countries will determine how effectively it can serve the broader market projected to shift 75 % of AI workloads to local providers by 2030. Buyers should monitor the rollout timeline and any additional financing that may accelerate the planned multi‑region expansion.
Source: Businesswire
