This Spanish startup’s breakthrough could finally make AI affordable for everyone. Spanish quantum computing startup Multiverse Computing just pulled off something remarkable: a massive €189 million ($215 million) Series B funding round that could finally solve AI’s biggest problem. Their breakthrough compression technology promises to make artificial intelligence accessible to businesses that have been priced out of the AI revolution.

The Technology That Could Change Everything

Here’s what makes this interesting: most AI compression techniques are like trying to fit into skinny jeans after Thanksgiving dinner. Sure, you might squeeze in, but you’re not going to perform your best. Multiverse’s CompactifAI technology is different. It can shrink Large Language Models by up to 95% without the usual performance penalty that makes compressed AI models feel like diet soda versions of the real thing.

CompactifAI models deliver impressive results:

  • 4x to 12x faster processing than uncompressed versions
  • 50% to 80% cheaper to run
  • So efficient that they can work on everyday devices like phones, laptops, and even tiny Raspberry Pi computers

Why This Matters for Your Wallet

Let’s talk real numbers. Multiverse’s compressed Llama 4 Scout Slim costs 10 cents per million tokens on Amazon Web Services versus 14 cents for the regular version. That 4-cent difference might not sound like much until you scale it up. For companies processing millions of AI requests daily, this could mean the difference between profitable AI deployment and budget-busting infrastructure costs.

Perfect Timing for a Massive Market

The timing here is almost too good to be true. The global AI inference market is worth $106.15 billion in 2025 and is expected to hit $254.98 billion by 2030. That’s a lot of zeros, and Multiverse is positioning itself right in the middle of this gold rush.

What’s been holding back many businesses isn’t a lack of interest in AI, but the brutal economics. Running sophisticated AI models requires expensive, specialized hardware that puts advanced AI capabilities out of reach for smaller companies. If Multiverse can genuinely deliver enterprise-grade AI performance at a fraction of the cost, they’re not just solving a technical problem but potentially democratizing access to cutting-edge AI.

Smart Money is Betting Big

The investor lineup reads like a who’s who of tech investing. Bullhound Capital led the round, and these are the folks who spotted winners like Spotify and Revolut early. When investors of this caliber write checks this big, it’s worth paying attention.

The full investor roster includes some heavy hitters:

  • HP Tech Ventures (because HP knows something about making technology accessible)
  • Forgepoint Capital International (cybersecurity experts who understand infrastructure)
  • CDP Venture Capital (Italy’s national fund)
  • Quantonation (quantum tech specialists)
  • Toshiba (bringing serious corporate validation)

What the Smart Money is Saying

Per Roman from Bullhound Capital, put it best: “Multiverse’s CompactifAI introduces material changes to AI processing that address the global need for greater efficiency in AI.” Translation: this isn’t just another incremental improvement, but potentially a fundamental shift in how AI gets deployed.

HP’s Tuan Tran highlighted something crucial: “By making AI applications more accessible at the edge, Multiverse’s innovative approach has the potential to bring AI benefits to companies of any size.” This matters because until now, sophisticated AI has been largely the domain of tech giants with massive budgets.

The Brains Behind the Breakthrough

This isn’t some garage startup stumbling onto success. Román Orús, the Chief Scientific Officer, is a professor at the Donostia International Physics Center and a pioneer in tensor networks. This is the quantum-inspired computational framework that makes CompactifAI possible. When someone at his level co-founds a company, it signals serious scientific credibility.

CEO Enrique Lizaso Olmos brings an interesting mix: mathematical expertise combined with real-world business experience as former deputy CEO of Unnim Bank. It’s rare to find leaders who can navigate both complex mathematics and corporate boardrooms.

The Science Made Simple

Tensor networks sound complicated, but here’s the key insight: they let you peek inside a neural network and identify which connections actually matter. As Orús explains it:

“For the first time in history, we are able to profile the inner workings of a neural network to eliminate billions of spurious correlations to truly optimize all sorts of AI models.”

Think of it like editing a massive document. Instead of randomly deleting paragraphs and hoping for the best, you can identify which parts actually contribute to the meaning and which are just filler.

What You Can Actually Use Today

Multiverse isn’t just talking about future possibilities. They already offer compressed versions of popular open-source models:

  • Llama models (various sizes from 8B to 70B parameters)
  • Mistral Small 3.1
  • DeepSeek R1 (coming soon)

You can access these through Amazon Web Services or license them directly. One notable limitation: they only work with open-source models. If you’re hoping for a compressed version of GPT-4, you’re out of luck for now.

Real Companies, Real Results

With 160 patents and 100 customers globally, this isn’t vaporware. Major companies are already using their technology:

  • Iberdrola (major Spanish utility company)
  • Bosch (German engineering giant)
  • Bank of Canada (yes, an entire country’s central bank)

When organizations this diverse and sophisticated are willing to bet on your technology, it suggests the real-world performance lives up to the hype.

Why This Could Be Huge

Damien Henault from Forgepoint Capital nailed it:

“The company is well-positioned to be a foundational layer of the AI infrastructure stack.”

If he’s right, Multiverse isn’t just building a product but potentially becoming essential infrastructure for the entire AI ecosystem.

Here’s what makes this particularly compelling: while everyone’s been focused on making AI models bigger and more capable, Multiverse is solving the opposite problem. How do you make AI practical and affordable for everyday use? In a world where AI capabilities are advancing faster than most organizations can afford to implement them, that’s exactly the right question to be asking.

The Bottom Line

With nearly $250 million in total funding, Multiverse Computing has the resources to scale rapidly. More importantly, they’re tackling one of the biggest barriers to AI adoption: cost. If their technology delivers on its promises, we could see AI capabilities become accessible to businesses that have been sitting on the sidelines.

As CEO Lizaso Olmos put it:

“What started as a breakthrough in model compression quickly proved transformative, unlocking new efficiencies in AI deployment.”

Sometimes the most important innovations aren’t about making technology more powerful but making it more practical. Multiverse might just be proving that point.The real test will be whether they can scale their compression technology across more models and use cases. But with this level of funding and the quality of customers already signed up, they’re certainly well-positioned to find out.