Sources: Spain-based Multiverse, which shrinks LLMs to reduce energy and compute costs, is in talks to raise ~€500M at a €1.5B+ valuation and has hit €100M ARR
Multiverse Computing SL, a Spanish artificial intelligence software company, is in discussions to raise about €500 million …
Context & Ripple Effects
Multiverse Computing’s reported financing talks follow a sharp progression from its €25M Series A at a €100M valuation to a €189M Series B round less than two years later. The company’s pitch has remained focused on compressing large language models so customers can use less compute and energy.
The reported €100M ARR gives the prospective round a commercial marker alongside its technical claim: model compression is being sold as operating-cost infrastructure, not solely as an AI research capability.
First-order effects
- A roughly €500M raise, if completed, would give Multiverse substantial capital to expand delivery and support around its LLM-compression software while valuing the business above €1.5B.
- Existing and prospective customers gain a better-capitalized supplier for reducing the compute footprint of deployed language models; Multiverse gains stronger credibility in enterprise procurement.
Second-order effects
- Compression and model-optimization vendors face a clearer pressure to show measurable savings and recurring revenue, rather than only technical benchmarks.
- Cloud and AI application buyers may gain additional leverage to evaluate software-based efficiency measures alongside spending more on underlying compute.
Third-order effects
- If enterprises continue to reward tools that lower model-running costs, value in AI infrastructure could shift toward the software layer that makes deployed models cheaper to operate, not only toward providers of more compute.
- Large financing rounds for efficiency vendors would test whether cost reduction can sustain infrastructure-scale valuations through recurring software revenue; that remains dependent on durable customer savings.
The trend: AI infrastructure investment is broadening from funding compute supply to funding software that reduces the compute required for useful model deployment.