Liquid Energy

Website

Company Bio

Visit website ↗Explore our full portfolio & program companies
Founders

LvlUp insightsGenerated from LvlUp’s portfolio data and the company’s own updates
Similar companies

More in the LvlUp portfolio

See all →

Building something ambitious?

LvlUp Ventures backs founders from first check to late stage.

Apply for funding
Arsh Sekhon, Sid Ravikumar
Seed
Canada
AI, B2B, DeepTech, Energy
Liquid Energy manufactures and deploys NanoDCs, self-contained, off-grid solar-powered compute modules. Each unit generates its own electricity, stores it in onboard LFP batteries, and runs AI inference silicon 24/7. We don't sell power. We sell compute. Our customers are the companies actually consuming AI tokens at scale: startups fine-tuning and running their own models, tech companies embedding inference into their products, and enterprises deploying AI agents, reasoning systems, and automated workflows on internal data. This is the fastest-growing and most underserved segment of the compute market. Hyperscalers serve themselves. We serve everyone else. We have agreements with large inference platforms that supply billions of dollars of inference and compute, giving us immediate access to demand without building our own marketplace. Tokens are delivered over the internet, meaning our units can be deployed anywhere there is sunlight, independent of grid infrastructure. Revenue is recurring and begins on day one of deployment. Each NanoDC generates approximately $1,500 per year in compute revenue at commodity token pricing, with zero marginal energy cost. Sunlight is free. Battery cycling costs are negligible at current LFP cell prices. Hardware payback is under 24 months. Solar panels degrade at roughly 0.5% per year, giving each unit a 10 to 15 year revenue tail after payback. The inference market will mature like the oil market. Today, every provider has a different cost of production. Some run on grid power at retail rates, others negotiate PPAs, others burn natural gas on site. As the market matures and token prices compress, high-cost producers will be culled. Only the lowest-cost suppliers will survive. Our energy cost is zero. We are building the Saudi Arabia of inference, the lowest-cost producer on earth, powered by the one input that is free and infinite. We scale like a manufactured product, not a data center. Every unit we build and ship is a self-sufficient, revenue-generating asset. No grid interconnection, no construction, no permits, no utility approval. Our cost structure improves with every doubling of units produced. Wright's Law, not Moore's Law. The components we buy, solar panels, batteries, inference silicon, are all on independent cost-decline curves, compounding our margins over time.

We turn sunlight into AI. The NanoDC is a self-contained, off-grid solar module that generates power and runs inference in the same unit - serving AI tokens at the lowest cost on earth. No grid. No construction. No assembly.