Anthropic has inked a massive 10 billion dollar cloud computing agreement with UK-based AI cloud startup Volta to secure processing power over the next six years. The partnership highlights the escalating race among top AI labs to lock in long-term compute capacity for frontier model training and inference.
⚡ The Norway Infrastructure Deal
Under the six-year contract, Volta will deliver dedicated cloud infrastructure powered by cutting-edge hardware. To build out the required capacity, Volta is partnering with crypto-mining firm Bitdeer to establish a high-density data center in Norway.
Key infrastructure specifications include:
- Total Power Capacity: 133 Megawatts (MW) dedicated to AI compute workloads.
- Hardware Architecture: Powered by Nvidia Vera Rubin systems, Nvidia's latest flagship AI chip architecture.
- Geographic Focus: Nordic data center placement in Norway leveraging renewable energy sources.
- Ecosystem Integration: Volta operates within the official Nvidia Cloud Partner program consortium.
🏗️ Compute Architecture & Partnership Flow
+-----------------------+ $10B 6-Year Agreement +-----------------------+
| | ------------------------------> | |
| Anthropic (Claude) | | Volta (Cloud Provider)|
| | <------------------------------ | |
+-----------------------+ 133 MW Compute Access +-----------+-----------+
|
| Infrastructure
v
+-----------------------+
| Bitdeer Norway Facility
| (Nvidia Vera Rubin) |
+-----------------------+
💡 Why Neocloud Startups Are Winning Big
The deal underscores a growing trend in the AI sector: frontier model developers are increasingly turning to specialized "neocloud" providers alongside traditional hyperscalers like AWS and Google Cloud. Specialized AI cloud providers offer flexible deployment structures and rapid hardware provisioning specifically tailored for massive LLM workloads.
For Anthropic, diversifying its cloud compute portfolio across multiple partners ensures redundancy and guarantees compute availability as model parameter sizes continue to grow exponentially.
🔮 What's Next for Frontier AI Compute
As demand for high-performance AI acceleration outstrips global supply, multi-billion-dollar long-term commitments are becoming standard practice for leading labs. Securing 133 MW of Nvidia Vera Rubin compute positions Anthropic to scale future iterations of the Claude model family without facing hardware bottlenecks. Expect further multi-gigawatt infrastructure deals across Europe and North America as frontier model training demands expand.
🔗 Reference
- Original Article: Read the full story on TechCrunch
