AI’s Next Frontier: Multiverse Computing Bets on Smaller, Local Models Amid Financial Instability
As global financial instability shakes the tech sector, venture capital firm Lux Capital has issued a stark warning to companies relying on artificial intelligence (AI): get your compute capacity commitments in writing. With private company defaults hitting a staggering 9.2%, the highest in years, the AI supply chain is under unprecedented strain. Against this backdrop, a Spanish startup, Multiverse Computing, is emerging as a potential game-changer with its innovative approach to AI: smaller, locally run models that bypass external compute infrastructure altogether.
Founded in 2019 and headquartered in San Sebastián, Multiverse Computing has quietly carved out a niche in the AI industry, leveraging its quantum-inspired compression technology to shrink large language models (LLMs) into lightweight versions capable of running directly on users’ devices. Its flagship product, CompactifAI, is a testament to this vision, offering a glimpse into a future where AI operates on the edge — offline, private, and resilient.
The Rise of Edge AI
The AI industry has long been dominated by cloud-based infrastructure, with giants like OpenAI, Google, and Meta relying on vast data centers to power their models. However, this reliance comes with risks: financial instability in the private credit market, rising compute costs, and growing concerns over data privacy. Multiverse Computing’s approach addresses these challenges head-on by enabling AI models to operate locally, eliminating the need for cloud providers or external data centers.
CompactifAI, named after Multiverse’s proprietary compression technology, is a chat-based AI app that showcases the potential of this approach. Embedded within it is Gilda, a compact model so small that it can run entirely on a user’s device, offline and without sending data to the cloud. For users, this means unparalleled privacy and reliability, especially in settings where connectivity is unreliable or sensitive data must remain on-premises.
However, the app is not without limitations. Older smartphones with limited RAM and storage may struggle to support Gilda, forcing the app to revert to cloud-based models via an API. This transition, handled automatically by Multiverse’s Ash Nazg system (a nod to Tolkien’s “One Ring”), compromises the app’s privacy benefits. Still, the app’s ability to switch seamlessly between local and cloud processing highlights the flexibility of Multiverse’s technology.
Targeting the Enterprise Market
While CompactifAI’s consumer-facing app offers a compelling demo of edge AI, Multiverse’s real focus is on the enterprise market. On March 6, 2026, the company launched a self-serve API portal, giving developers and businesses direct access to its compressed models. Unlike traditional AI deployments, which often require integration with platforms like AWS Marketplace, Multiverse’s API portal offers greater transparency and control over model usage.
Enrique Lizaso, CEO of Multiverse Computing, emphasized the portal’s value in a statement: “The CompactifAI API portal gives developers direct access to compressed models with the transparency and control needed to run them in production.” Real-time usage monitoring is a key feature, enabling businesses to optimize their AI workflows and reduce compute costs — a major advantage over traditional LLMs.
The enterprise appeal of smaller AI models is further bolstered by their growing capabilities. For instance, French AI startup Mistral recently launched Mistral Small 4, a compact model optimized for general chat, coding, and complex reasoning tasks. Similarly, Multiverse’s HyperNova 60B 2602, derived from OpenAI’s publicly available gpt-oss-120b model, boasts faster response times and lower operational costs, making it particularly suited for agentic coding workflows.
Competing with Tech Titans
Multiverse’s ambition to shrink AI models puts it in direct competition with tech giants like Apple, which has adopted a hybrid approach for its Apple Intelligence system. By combining on-device and cloud-based models, Apple sidesteps the limitations of local processing while maintaining a degree of privacy. Multiverse, however, is betting on a future where local AI models are powerful enough to stand alone, offering businesses and consumers a more resilient and private alternative.
The company’s technology has already found traction in critical industries, with over 100 global customers including the Bank of Canada, Bosch, and Iberdrola. These partnerships underscore the potential of edge AI in sectors where connectivity cannot be taken for granted, such as drones, satellites, and remote industrial settings.
A Bold Vision Amid Financial Headwinds
Multiverse’s rise comes at a pivotal moment for the AI industry. As financial instability disrupts the private credit market, companies are increasingly seeking cost-effective alternatives to traditional AI infrastructure. Multiverse’s quantum-inspired compression technology offers a compelling solution, promising to “radically slim AI costs,” as Lizaso puts it.
The startup’s financial trajectory reflects this optimism. After raising $215 million in a Series B funding round in June 2025, Multiverse is rumored to be pursuing a fresh €500 million investment at a valuation exceeding €1.5 billion. Such funding would position the company as a serious contender in the AI landscape, challenging established players with its innovative approach.
The Road Ahead
While CompactifAI’s consumer app has yet to achieve mass adoption — Sensor Tower reports fewer than 5,000 downloads in the past month — its real value lies in its potential to transform how businesses deploy AI. By enabling local processing, Multiverse is not just reducing costs but also unlocking new use cases for AI in environments where connectivity and privacy are paramount.
As the AI industry grapples with financial uncertainty and evolving consumer demands, Multiverse Computing’s vision of smaller, locally run models offers a glimpse into the future — a future where AI is not just smarter but also more accessible, private, and resilient. Whether this vision will reshape the industry remains to be seen, but Multiverse’s ambition is clear: to make AI work for everyone, everywhere.
For now, as Lux Capital’s warning reverberates across the tech sector, Multiverse’s approach serves as a timely reminder that innovation often thrives in times of disruption.
