TensorX raises €8M for sovereign AI on Nvidia Blackwell
💡Discover how European startups are leveraging Blackwell GPUs to solve data residency challenges for enterprise AI.
⚡ 30-Second TL;DR
What Changed
Raised €8M to purchase Nvidia Blackwell B300 GPUs.
Why It Matters
TensorX addresses the critical need for sovereign AI in Europe, potentially capturing market share from US-based cloud providers for sensitive data workloads.
What To Do Next
If you operate in a regulated European sector, evaluate TensorX's infrastructure for your next AI deployment to ensure data compliance.
🧠 Deep Insight
AI-generated analysis for this event.
🔑 Enhanced Key Takeaways
- •TensorX is headquartered in Dublin, Ireland, leveraging the country's status as a major European data hub to facilitate compliance with the EU AI Act.
- •The funding round was led by Atlantic Bridge, a venture capital firm specializing in deep-tech investments within the European market.
- •The platform utilizes a proprietary 'Privacy-First' orchestration layer that ensures data never leaves the sovereign cloud environment during inference.
- •TensorX plans to integrate with existing European private cloud providers to offer hybrid deployment options alongside their dedicated Blackwell clusters.
- •The B300 GPU deployment is specifically optimized for low-latency inference of large-scale models (LLMs) exceeding 70B parameters, addressing the performance gap in current European sovereign clouds.
📊 Competitor Analysis▸ Show
| Feature | TensorX | Scaleway (Sovereign Cloud) | OVHcloud AI |
|---|---|---|---|
| Primary Hardware | Nvidia Blackwell B300 | Nvidia H100/A100 | Nvidia H100 |
| Target Market | Regulated (Legal/Med/Bank) | General Enterprise | General Enterprise |
| Data Residency | Strict EU-Only | EU-Only | EU-Only |
| Inference Focus | High-Performance/Low-Latency | General Purpose | General Purpose |
🛠️ Technical Deep Dive
- Deployment utilizes Nvidia Blackwell B300 GPUs, which feature the second-generation Transformer Engine for accelerated FP4 precision inference.
- Infrastructure architecture relies on NVLink Switch System to provide high-bandwidth, low-latency interconnects between GPU nodes, minimizing bottlenecks for distributed inference.
- The software stack incorporates a custom-built Kubernetes-based orchestration layer designed to enforce strict data residency policies at the container level.
- Implements hardware-level encryption and Trusted Execution Environments (TEEs) to ensure data remains encrypted even during active processing in memory.
🔮 Future ImplicationsAI analysis grounded in cited sources
⏳ Timeline
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
Same topic
Explore #sovereign-ai
Same product
More on tensorx-inference-platform
Same source
Latest from The Next Web (TNW)
China claims supercomputing crown without US silicon
Cerebras revenue doubles but stock drops on margins
OpenAI unveils Jalapeño, its first custom inference chip
Apple suppliers pivot to AI hardware and humanoid robots
AI-curated news aggregator. All content rights belong to original publishers.
Original source: The Next Web (TNW) ↗