RHAELM
Category: AI Infrastructure
UK-based sovereign AI cloud and infrastructure developer that combines power, data centers and compute to deliver gigawatt-scale secure AI capacity across Europe and APAC. The company is led by Bradd Lewis. Based in London, United Kingdom. Team size: 101-500. Latest round: Undisclosed. Key investors include Apollo Global Management (intended strategic investment and financing partner for the Chiba Project).
- Headquarters
- London, United Kingdom
- Team size
- 101-500
Value proposition
Removes the main constraint on AI deployment, which is access to power and land. It builds AI infrastructure next to power generation, including behind-the-meter sites at power stations, and lets customers choose their sovereign controls and semiconductor architecture.
Products and solutions
Ring-fenced AI infrastructure (compute, networking, storage), end-to-end power-plus-compute platforms, Bare Metal, Kubernetes and AI-as-a-Service offerings, optional sovereign controls including verifiable compute, kill-switch firewalls and black-swan resistance, large-scale AI data center development such as the Chiba Project in Japan (up to 400 MW)
Unique value
Sovereign AI cloud 'from the soil to the token': it integrates power generation, site, data center and compute, with selectable sovereignty controls and architecture-agnostic GPU/CPU clusters.
Target customer
Governments, defence organisations, regulated enterprises (healthcare, finance) and AI developers that need sovereign, secure, large-scale AI compute
Industries served
Government, Defence, Healthcare, Financial Services, Enterprise AI
Technology advantage
Places AI data centers behind the meter at power stations, such as JERA's Chiba Thermal Power Station, so capacity can come online years ahead of grid-connected builds. Compute is standardized on Dell AI Factory rack-scale systems with NVIDIA GPUs, and Palantir serves as the infrastructure operating system. Apollo intends to act as strategic investment and financing partner.
How they differentiate
Combines sovereignty-grade controls with integrated power generation, rather than offering GPU leasing alone. Partners directly with utilities (JERA) for behind-the-meter power. Is architecture-agnostic and lets government and defence workloads configure their own sovereign controls.
Main competitors
Nscale, Nebius, Fluidstack
Key partnerships
JERA (MoU and Heads of Terms for the Chiba Project, Oct 2026), Dell Technologies (Dell AI Factory, MoU partner), Apollo Global Management (intended strategic investment and financing partner for Chiba), Accenture (EMEA delivery partner, Jan 2026), Palantir (infrastructure operating system, EMEA), NVIDIA (compute)
Major milestones
Jan 2026: as Sovereign AI (S-AI), selected Accenture and Palantir to build sovereign AI infrastructure across EMEA using Dell AI Factory and NVIDIA. Sep 2026: rebranded from Sovereign AI to RHAELM. Oct 1, 2026: signed an MoU with JERA and Dell Technologies, plus Heads of Terms with JERA, for the US$15B, 400 MW Chiba AI infrastructure project, with Apollo as intended financing partner.
Growth metrics
Chiba Project: up to 400 MW, with total capital deployment expected above US$15B (¥2.3T) across all phases and operations targeted around 2028. Partners plan to explore multi-gigawatt capacity across JERA sites in the 2030s.
Market positioning
Emerging sovereign AI infrastructure developer and neocloud targeting gigawatt-scale, power-led AI capacity in Europe and APAC, with Japan as its flagship market.
Geographic focus
UK/EMEA and APAC (Japan, Korea)
Patents and IP
RHAELM® trademark; no public patents found
About Bradd Lewis
Founder and CEO of Sovereign AI (S-AI), the company's name before the Sep 2026 rebrand to RHAELM. Earlier background not verified.
Latest news about RHAELM
More AI Infrastructure companies
Official website: https://rhaelm.com