SAN SEBASTIÁN, Spain, September 24 2025 -- Multiverse Computing, the leader in quantum-inspired AI compression, today announced a collaboration with SOHMA AI to deploy the first behavioral AI platform for youth mental health support operating fully on user devices. Leveraging compressed models from Multiverse, SOHMA AI’s deployment demonstrates a critical breakthrough enabled by Edge AI: sophisticated behavioral analysis that runs entirely on consumer devices in real time, an early intervention opportunity previously thought to be impossible due to cloud-based computational limitations.

New AI Approach To Answer Youth In Crisis

Mental health challenges affect up to 1 in 3 elite youth athletes, with 58% of adolescent mental health conditions going undiagnosed in the United Kingdom alone. Current solutions remain largely reactive, clinical, and stigmatized, resulting in continued low adoption, late intervention, and widespread underreporting of psychological distress.

SOHMA AI selected Multiverse’s compression technology to make possible what no digital system has yet achieved: reading unstructured non-verbal human signals and translating them into structured emotional insights – mood, focus, stress, mindset – in real time. This creates adaptive, emotionally responsive systems that work across mobile, wearables, AR/VR, and robotics.

SOHMA’s first deployment is with elite athletes, integrating preventative mental health tools seamlessly into their digital lives and responding to their emotional and cognitive states in real time. The framework is designed to scale into digital health, education, gaming, and workforce wellbeing – anywhere human emotion and resilience matter.

The collaboration addresses what SOHMA AI’s CEO Lisa Franke calls a critical gap in current approaches. Franke adds, "Digital systems today can’t understand how we feel. SOHMA changes that by creating the emotional intelligence layer for AI itself. To change the future of mental health, we need to meet young people where they are – on the pitch, in their games, and inside digital spaces they trust. Together with Multiverse, we’re making AI human-aware."

The technical challenge was substantial. Traditional behavioral AI requires significant computational power, limiting applications to research settings or cloud-dependent solutions unsuitable for real-time intervention.

Quantum-Inspired Compression Solves Fundamental Compute Barrier

Multiverse's CompactifAI uses quantum-inspired techniques to compress large language models by up to 95% with only 2-3% precision loss. This enables 4x–12x faster processing and 50%–80% lower inference costs, making it possible to run advanced AI models on devices ranging from smartphones to Raspberry Pi, and unleashing what Multiverse Computing CEO Enrique Lizaso describes as a new category of edge AI applications:

"Edge AI holds the potential to reshape industries of all shapes and sizes, but first the models need to actually fit on device," said Lizaso. “CompactifAI’s quantum-inspired compression algorithms open the door to high-performance AI at a fraction of the compute and cost.”

Lizaso adds: “Real-time behavioral AI was impossible on mobile devices until now – the computational requirements were simply too massive for consumer hardware. Our compression technology doesn't just make AI faster, it fundamentally changes what's possible. We're talking about shrinking models by 95% while maintaining accuracy, which means SOHMA can deliver the same behavioral analysis that required cloud supercomputers directly through a teenager's smartphone. From there, you can imagine infinite possibilities in the fields of digital therapeutics, education, gaming, and workforce wellness."

SOHMA AI will first pilot the technology in elite football academies across the U.K. and U.S. The long-term roadmap expands across digital health, education, gaming, and robotics, making emotional intelligence a native capability of AI systems.

Multiverse Computing is currently offering compressed versions of top Llama, DeepSeek, and Mistral models, with additional models planned for release soon. In June 2025, Multiverse announced that it had secured an additional $215 million in Series B funding from leading investors Bullhound Capital, HP Tech Ventures, Forgepoint Capital, and Toshiba. The investment will accelerate widespread adoption of CompactifAI models to transform the $106 billion AI inference market.