Multiverse Computing targeting up to $570M in latest round
Tech.eu John Reynolds ● Covered by 3 sources
Spanish startup Multiverse Computing wants up to $570M more, valuing it at $1.7bn. That's five times its Series B value, betting on shrinking AI to run on phones and drones instead of data centres.
Based on reporting by Tech.eu, John Reynolds — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Multiverse Computing, a scaleup out of San Sebastián, is chasing as much as $570 million in fresh funding at a $1.7 billion valuation — a number that would look ambitious even in Silicon Valley, let alone in Spain's usually modest startup scene. The round is still open, led by Forgepoint Capital International, BNPP SIVF, and Bullhound Capital, with Santander, Tikehau Capital, Orange Ventures and Scania Invest also writing checks. Once it closes, the company will have raised roughly $800 million total, and that valuation marks a fivefold jump from the $215 million Series B it closed not long ago.
The pitch is straightforward: AI is too big, too power-hungry, and too dependent on data centres, and Multiverse thinks it has a fix. Its core product, CompactifAI, borrows techniques from quantum physics to compress large language models by 80 to 95 percent, the company claims, without meaningfully hurting accuracy. Shrink the model enough, and suddenly it can live on a drone, a satellite, a security camera, or a phone — no round trip to a server farm required.
That edge-AI bet is the real story here, not just the funding number. Multiverse is wagering that as AI spreads into physical devices — vehicles, telecom gear, industrial sensors — the winners will be the companies that make models small enough to run locally, cheaply, and without burning through a small country's worth of electricity. It's already got paying customers testing that thesis: Bosch, Iberdrola, Telefónica, PwC, Allianz, Bank of Canada, and Indra span manufacturing, energy, defense, finance and healthcare, according to the company.
Forgepoint's Damien Henault, whose firm co-leads the round, described Multiverse as having moved beyond being a compression vendor into something closer to a full AI operating system — sitting across both infrastructure and application layers. That's a big claim for a company most people outside quantum-computing circles have never heard of. But the investor list, and the size of the round itself, suggests some serious money believes the edge-AI thesis is about to get a lot more relevant as enterprises tire of ballooning cloud compute bills.
My take — AI-written commentary, not fact-checked reporting
A five-x valuation jump on the promise that quantum-flavored compression solves AI's cost problem is exactly the kind of story that makes me want more independent benchmarks before I get excited. Model compression is real and useful, but 80-95% size reduction with 'immaterial' accuracy loss is a claim that needs third-party scrutiny, not just investor enthusiasm. Still, if edge AI actually takes off the way this round is betting, European deep tech finally gets a genuine home-grown contender instead of just reselling US infrastructure — that part I'll cheer for.
Read more about this at: Tech.eu