Multiverse Computing targets €500m Series C at €2bn valuation

Multiverse Computing is targeting a €500m Series C at a €2bn post-money valuation, a financing that would make the Donostia, Spain-based AI model-compression company a unicorn if completed as planned. The Multiverse Computing Series C is still open, according to the company, so the final size and investor roster have not been locked.

The round is being co-led by Forgepoint Capital in the US, Bullhound Capital in the UK and Solar Impulse Venture Fund, the climate-focused fund backed by BNP Paribas. Multiverse said the financing also has commitments from the European Innovation Council Fund, Spanish state investor SETT, Qatar Development Bank and NAVentures, the venture arm of the National Bank of Canada.

Other named backers include Santander Alternative Investments, HP Inc and Orange Ventures. The company said the round is in its final stages and remains open to more strategic investors. If completed at the announced target, it would take Multiverse’s total funding to more than €700m and would rank among the largest private tech financings raised by a Spanish company.

What does Multiverse Computing do?

Multiverse sells technology for shrinking large AI models so they need less computing power to run. Its main product, CompactifAI, uses a method the company describes as inspired by quantum physics to compress large language models and other AI systems.

The company claims CompactifAI can cut the size of some large-language models by as much as 95% while causing little loss in accuracy. That claim matters because AI infrastructure costs remain a constraint for enterprises that want models running outside large cloud environments, although Multiverse did not disclose independent benchmark data in the announcement.

CEO and cofounder Enrique Lizaso said in a statement that the AI industry had treated expensive infrastructure as a requirement for powerful models, and claimed Multiverse had shown AI could run “on a smartphone, inside a sovereign data center, on a factory floor with no cloud connection.” The company says its models are already used in devices and systems including drones, cameras, satellites, vehicles and telecom infrastructure.

Multiverse is also building a broader platform intended to match models to workloads and manage deployment more efficiently. The company says that platform combines compression, GPU management, AI services and control systems.

Who is using Multiverse Computing?

Multiverse says it has more than 100 customers in Europe and North America across manufacturing, finance, energy, aerospace, cybersecurity, defence and life sciences. Named customers include the Bank of Canada, Bosch, Telefónica and Allianz.

Lizaso previously told Sifted that Multiverse was on track to reach €200m in annual recurring revenue in 2026. The company now says annualised revenue has grown more than tenfold from the same period a year earlier, but it did not provide the revenue base for that comparison.

The new capital is earmarked for expanding the company’s catalogue of compressed models, increasing research and development, investing in AI infrastructure and growing in Asia, the Middle East and North America. That puts Multiverse in the part of the AI market focused less on building ever-larger frontier models and more on making existing models cheaper and easier to deploy at the edge or in controlled enterprise environments.

This story draws on original reporting from Sifted.