Amphiform raises $5.5M to build cheaper, denser fuel cells for data centres
What's the deal? Amphiform, an Oxford-based deep tech startup, has closed a $5.5M pre-seed round led by General Catalyst and co-led by Main ObjectDealroom has a profile for this one. Try Dealroom →. Embassy Ventures, K5/Tokyo BlackDealroom has a profile for this one. Try Dealroom →, Hugging Face co-founder Thomas WolfDealroom has a profile for this one. Try Dealroom →, and Charlie Songhurst also participated.
The company is building fuel cells using materials engineered atom by atom. It claims the approach can deliver 30x higher power density than current fuel cells and cost 85% less, by turning the entire volume of a cell into active surface rather than relying on a thin catalyst interface.
Amphiform's fuel cells run on liquid fuels like methanol and ethanol, both of which can be produced from biomass or captured CO₂. The startup plans to deploy first in AI data centres, where power demand is outstripping grid capacity.
Why now? The buildout of AI infrastructure has created an urgent energy bottleneck. Data centres need more power than grids can supply, and alternatives remain scarce. Defence and space applications face similar density constraints.
Amphiform also positions itself as a Western alternative to energy supply chains that depend heavily on China and volatile oil markets — a selling point with growing geopolitical weight.
What could go wrong? This is a pre-seed company with lab-stage technology making bold performance claims. Atomically engineered materials are notoriously difficult to manufacture at scale, and the gap between laboratory results and commercial production has swallowed many deep tech startups.
Fuel cell companies have a long history of overpromising on timelines. Amphiform will need to demonstrate that its nano-catalyst approach works outside controlled conditions before it can compete with lithium-ion batteries or grid-scale alternatives already in deployment.
The signal: This round reflects a broader pattern of AI-adjacent infrastructure attracting serious VC capital at ever-earlier stages. General Catalyst, one of the world's most active venture funds, co-leading a pre-seed for a lab-stage materials science company underscores how acutely investors feel the energy bottleneck threatening AI scaling. The participation of Hugging Face co-founder Thomas Wolf — an angel investor rooted in the AI ecosystem — further suggests that those closest to AI's compute demands see hardware and energy breakthroughs, not just software, as the binding constraint.
Read more: globenewswire.com
