Liquid Energy, a five-person Y Combinator startup from Waterloo, Canada, says it can manufacture 2 megawatts a month of NanoDC units and plans to scale that to 20 megawatts a month over the next year, according to its Y Combinator company page. NanoDC units are factory-built boxes that carry their own solar panels, batteries and AI inference computers, the company says. It launched them on 5 November in a Launch YC post.
The company says a unit can be deployed anywhere in four weeks and turned on in 30 minutes, with no grid connection. By our arithmetic, at the 20 megawatts a month it is aiming for, units adding up to one gigawatt, the scale of a large AI campus, would take 50 months; at today’s stated 2 megawatts a month, 500 months.
Business Pill · Who Pays for the Grid
A short explainer on who pays for grid capacity. It teaches the general idea only and does not discuss this story.
The key insight: As we read the launch, Liquid Energy is pitching speed to power rather than more compute: units it says carry their own solar and batteries, a stated 2 megawatts a month of manufacturing today, and a target of 20 megawatts a month within a year, at which a gigawatt of units would take about 50 months by our arithmetic.
The launch post calls NanoDC “a solar-powered autonomous datacenter that can be deployed anywhere on the planet in 4 weeks”. It says each unit is “built rapidly at a factory and shipped to the site ready to be turned on in 30 minutes”, and that “You open the door and the solar panels fold-out like an accordion”.
The company says the computers are cooled with “patent-pending high-density air cooling” and connect to the internet via Starlink, and that the designneeds “No construction. No grid interconnect. No Water supply.” Its website puts the time “From signed contract to first token delivered in under 8 weeks”.
The same text ran on the company’s X account on 5 November, where it had about 114,000 views when we read it the next day, and in the post the company says “The first units will be live this month. Thousands will be shipping starting 2027.”
Months to build one gigawatt of NanoDC units at the monthly manufacturing capacity Liquid Energy states on its Y Combinator page: 2 megawatts a month today and 20 megawatts a month targeted over the next year. The months are our arithmetic.
Who It Sells To
Liquid Energy says it sells inference, not boxes. Its website says “We sell inference to companies of every size”. Asked on X whether the units are portable, the company replied: “We don’t sell or install individual units outside of demo units. The idea is to do MW-GW sized large solar compute farms.”
The Y Combinator page says the company has “20 units” and has “spent the last year in building the supply chain and manufacturing”. It lists a team of five, founded in 2024, with Sid Ravikumar as chief executive, Arsh Sekhon as president and Sajan Khosa as a founder; the website also names Romit Bali as founding engineer.
The Grid Wait It Is Pitching Against
The launch post states “11 Trillion will be invested in AI compute over the next 4 years”, without naming the source of that estimate, and says most of that spending will be for inference and that most of it “will be severely delayed: stuck in grid interconnect queues, site permits, or supply chain issues”.
For scale, hooking up 100 megawatts to the grid in Northern Virginia takes about seven years, according to the International Energy Agency, as TechCrunch reported and we noted in our jet-turbine coverage. By our arithmetic, the four weeks Liquid Energy states is about 1% of those seven years (84 months). The two figures measure different things: one company’s stated deployment time for a unit, and an agency’s estimate for a large grid connection.
As we read the company’s posts, the product is inference sold from fleets of units, not boxes sold to buyers: it says it does not sell or install individual units outside demo units and aims for MW-GW sized solar compute farms.
As we read the numbers it does publish, scale is the open question: 20 units today and 2 megawatts a month of manufacturing are small next to gigawatt campuses, by our arithmetic.
As we read what is missing, the company gives no power or compute figure per unit and no account of night-time or cloudy-weather output beyond batteries, so the four-week claim cannot yet be checked against delivered capacity.
Liquid Energy, Launch YC post, 5 November 2026
“No construction. No grid interconnect. No Water supply.”
Three Implications
FOR AI INFERENCE BUYERS Liquid Energy says it sells inference to companies of every size, from signed contract to first token in under 8 weeks.
FOR GRID-CONSTRAINED DEVELOPERS Our earlier coverage notes that a 100 MW grid hookup in Northern Virginia takes about seven years, according to the IEA, as TechCrunch reported.
FOR INVESTORS The company does not disclose its funding or prices; it lists Y Combinator’s Fall 2026 batch.
The framework says: “The structural bottleneck isn’t computing, it’s electrons.”
As we read it, Liquid Energy’s pitch is built on the same premise: it says its units bring their own power so new capacity does not wait for a grid connection. The framework predates this launch and does not discuss Liquid Energy.
What Is Not Established
Liquid Energy does not publish the computing or power capacity of a single unit, the number of chips inside, the size of its solar array or battery, its prices or its funding. It does not say how units keep serving tokens at night or in cloudy weather beyond storing the day’s solar power in batteries, which its website describes as “The day is banked”. The four-week and 30-minute figures are the company’s claims. We did not contact Liquid Energy.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
Liquid Energy launched NanoDC, a factory-built solar-and-battery unit for AI inference that it says needs no grid connection, and says it can make 2 megawatts a month of units today and plans 20 megawatts a month within a year. At that target, by our arithmetic, a gigawatt of units would take about 50 months.
A note on sourcing. Product claims and quotations are from Liquid Energy’s Launch YC post of 5 November 2026, its Y Combinator company page, its website and its posts on X. The Northern Virginia grid figure is the International Energy Agency’s, as TechCrunch reported and as cited in our earlier coverage. The months-to-a-gigawatt figures and the comparison are our arithmetic. We did not contact Liquid Energy.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
