Multiverse Computing

What if you could compress a model by 95%… without losing what actually matters? 💡 In this article, our Machine Learning Engineer Javier Boix explores how we tackled one of the biggest challenges in AI deployment: making high-performance models viable in real-world, resource-constrained environments. At Multiverse Computing, efficiency isn’t a trade-off, it’s the foundation. Using our quantum-inspired compression techniques, we rethink how models are structured to dramatically reduce their footprint while preserving performance. ✅ Over 95% model size reduction (from 258 MB to 8 MB) ✅ No meaningful loss in output quality ✅ Up to 2x higher throughput & ~50% lower energy consumption ✅ Deployment-ready models for edge environments 🔗 Find the full article here: ow.ly/wwz750YUtFn

⚛️ Multiverse Computing is proud to sponsor Q-Expo 2026 at Palacio Euskalduna (Bilbao), May 18–19. Come meet us at Booth 10 and join our Co-Founder & CSO @OrusRoman for the panel: 🎤 “The Global Quantum Computing Race: Are Today’s Leaders Tomorrow’s Winners?” See you in Bilbao 🤝 🔎 qexpo.org #QExpo2026 #QuantumComputing #QuantumAI #Bilbao #DeepTech #MultiverseComputing

La IA ya es una decisión estratégica para las empresas. Hoy @AlumniDBA y @PwC_Spain , hemos contado con @Enrique Lizaso, CEO de @MultiverseCompu , en la conferencia: “IA eficiente o irrelevancia: decisiones que solo competen al CEO”. 👉acortar.link/BC207E @deustoDBS

🚀 CompactifAI API now powers the leading coding agents at up to 75% lower cost Plug it into your preferred coding agent and run leading models like GLM 5.1 as the backend, without changing your IDE or workflow. Main highlights: ✅ Drop-in integration with Cursor, OpenCode and other leading coding agents ✅ Up to 75% lower cost per token vs the comparable frontier setup ✅ Externally validated by Artificial Analysis's Agentic Index: GLM 5.1 on CompactifAI API scores 6% higher than Sonnet 4.6 Same agent. Same workflow. Lower bill. ➡️ Request access: ow.ly/9iQY50YZqQ6 📄 More info: ow.ly/wOir50YZqQ4 #CodingAgents #GLM5.1 #InferenceProvider #CompactifAI #MultiverseComputing

@MultiverseCompu and @ARQUIMEACenter are partnering to advance efficient AI for unmanned systems in Defense & Space 🛰️ Combining CompactifAI with ARQUIMEA’s expertise in autonomous systems, we are enabling AI models that are: ⚡ Smaller ⚡ Faster ⚡ More energy-efficient Built for next-generation edge AI and autonomous platforms 🛡️ 🔗 multiversecomputing.com/resources/arqu… #AI #DefenseTech #EdgeAI #SpaceTech #CompactifAI

⚛️ What happens when a production LLM can directly interface with quantum hardware? Our research team at Multiverse Computing has demonstrated a practical framework connecting modern LLMs with quantum computing infrastructure moving hybrid AI + quantum systems beyond theory. This research explores how AI models can orchestrate quantum resources as part of real-world reasoning and execution workflows. 📰 Read the paper: multiversecomputing.com/papers/talking… #QuantumComputing #AI #LLM #QuantumAI #MultiverseComputing

🚀 @MultiverseCompu has been named by @TechCrunch as one of the “21 European startups to watch.” We are building smaller, more efficient AI models that reduce inference costs and energy consumption while delivering enterprise-grade performance. Proud to represent Europe’s growing AI and deeptech ecosystem 🇪🇺 Read more 👇 techcrunch.com/2026/05/02/bey… #AI #GenAI #DeepTech #Europe #TechCrunch #MultiverseComputing #CompactifAI

🚀 New: HyperNova 60B 2605 is here. An updated version of our open-source 60B signature model, with stronger coding capabilities and higher accuracy across several benchmarks vs. HyperNova 60B 2602. The headline number: LiveCodeBench 68.68, up from 51.53 in the previous release, and ahead of gpt-oss-120B at 62.75. 📄 Read more here: ow.ly/pmlJ50YVpC2 🔗 Available today on Hugging Face: ow.ly/Cgce50YVpAf #AI #LLM #CompactifAI #HyperNova #MultiverseComputing

How do you double GPU utilization without buying a single piece of new hardware? 📈 In our latest blog, Mikel Irastorza, MLOps Manager at @MultiverseCompu , shares the playbook behind how we doubled GPU utilization across a 1,000+ GPU fleet, using the exact same hardware footprint. At Multiverse Computing, our research and engineering teams need flexible access to high-end GPUs (H200, B200, B300) to push the limits of model compression. Building that capability in-house meant designing a multi-cloud orchestration layer that treats compute as a commodity, picks the best available option for every job, and turns efficiency into a strategic advantage. We built this GPU Compute layer in collaboration with SkyPilot, a trusted partner of Multiverse Computing. ✅ 2x GPU utilization on the same hardware: twice the work, same compute budget, no extra CapEx. ✅ Reduced dependency on any single cloud vendor: freedom to switch based on price, capacity or hardware availability. ✅ Lower operational overhead for our MLOps team: no more provider-specific tooling or manual job juggling. ✅ One unified workflow: engineers focus on the code and the job they need to run, not on the underlying cloud or hardware. 🔗 Find the full article here: ow.ly/LO0j50YUtax #MLOps #GPU #LLM #CloudComputing #MultiverseComputing

The AI-SHIELD project, coordinated by @MultiverseCompu, has been selected for funding under the European Defence Fund (EDF) 2025 🇪🇺 We’ll coordinate a 7-partner European consortium to deliver a secure, efficient, and deployable LLM for mission-critical defence use powered by quantum-inspired compression (up to 80% reduction in size & energy). Advancing sovereign, privacy-preserving AI in Europe. #AI #Defence #EdgeAI #QuantumInspired #EDF2025

🐑 Meet the LittleLamb 0.3B family Tiny. Fast. Edge-ready. 3 CompactifAI-compressed models (-50%) → Gemma-270M footprint, Qwen3 reasoning, 32K context, EN/ES: ⭐ Flagship: general-purpose reasoning 🧠 Tool-Calling: agentic + structured outputs 📱 Mobile: built for on-device Smaller models. Smarter deployment. 🔗 Learn more: multiversecomputing.com/resources/intr… 🎯Access here: huggingface.co/MultiverseComp… #AI #LLM #EdgeAI #CompactifAI #OnDeviceAI

40 expertos mundiales elegidos por @ONU_es se han reunido en el Congreso en su 1º encuentro presencial. Llevan tiempo estudiando cómo legislar la inteligencia artificial. Hemos charlado con el único representante español @OrusRoman (@MultiverseCompu) en este panel de expertos.

📺 @OrusRoman Co-Founder & Chief Scientific Officer of @MultiverseCompu shares the risks of not regulating AI and how policymakers need to closely watch the latest developments of the technology at @rtvenoticias

🇯🇵 Multiverse Computing partners with Marubeni to expand CompactifAI in Japan. As AI adoption surges, so do costs, compute limits, and energy demands ⚡ CompactifAI tackles this head-on: ✅ ~90% smaller models ✅ Minimal accuracy loss ✅ Lower costs & energy use A key step in bringing efficient, scalable AI to one of the world’s most advanced markets 🌍 🔗 multiversecomputing.com/compactifai #AI #Japan #EfficientAI #CompactifAI

AI will define our future, but are we ready for its impact? 🤔 Our CSO @OrusRoman , Spain’s representative on the UN AI Scientific Panel 🌍, warns: AI is comparable to the “birth of the atomic bomb” in both power and risk. Yesterday, he joined the UN expert meeting at the Spanish Parliament with PM @sanchezcastejon A new Industrial Revolution is already underway ⚙️🚀 📰 efe.com/ciencia-y-tecn… #AI #ArtificialIntelligence #DeepTech #UN

🏆 Zorionak @RealSociedad ! Proud to support such a forward-thinking club on their King's Cup triumph. Together, we are bringing AI & Quantum into elite football with our Player Injury Risk Monitor helping predict injuries and optimize performance ⚽️ The future of sport is here. 👉 multiversecomputing.com/player-injury-… #RealSociedad #CopaDelRey #SportsTech #AI #FootballInnovation

#AIAMSUMMIT26 | "Cariño, he encogido a la IA”: de la nube al dispositivo industrial Jabier Ruiz (@inetum_es ) · Iraia Ibarzabal (@MultiverseCompu) El diálogo abordó uno de los grandes frenos actuales en la adopción de la IA industrial: pasar del piloto a la implantación real. Modelos sobredimensionados, altos costes, consumo energético y falta de control están ralentizando su despliegue. La conversación puso de relieve la necesidad de una IA más eficiente, estable y predecible, adaptada a casos de uso concretos. No se trata de más IA, sino de una IA que funcione, que sea segura, escalable y capaz de generar valor desde el inicio en entornos industriales reales. #InteligenciaArtificial #Confianza #Innovación #TransformaciónDigital

AI is accelerating breakthroughs in immunotherapy and vaccine design 🧪 At @MultiverseCompu , we built epiGPTope a generative model fine-tuned from ProtGPT2 to design novel, biologically meaningful epitopes. By navigating the vast protein sequence space, it helps tackle one of the field’s biggest challenges. In collaboration with @CICbiomaGUNE , we are showing how AI + domain expertise can unlock faster, more cost-effective epitope discovery 🔬🤖 Read more 👉 phys.org/news/2026-04-a… #AI #Biotech #Immunotherapy #Vaccines #ProteinDesign