Multiverse Computing

CompactifAI by Multiverse Computing has been awarded the Solar Impulse Foundation Efficient Solution Label, a recognition granted to solutions that deliver both environmental and economic benefits 🌍🏆 As AI adoption continues to accelerate, so does the demand for computing power, energy, and infrastructure. At Multiverse Computing, we are tackling this challenge head-on. With our compression technology, our optimized AI models enable organizations to reduce inference costs, lower energy consumption, and deploy powerful AI models more efficiently 🤖📉 This recognition from the Solar Impulse Foundation reinforces our mission to make AI not only smarter, but also more sustainable and accessible. By reducing the resources required to run advanced AI, we help businesses unlock greater value while minimizing their environmental footprint 🌱🌍 A big thank you to the Solar Impulse Foundation and its network of independent experts for recognizing the impact of efficient AI technologies 🚀 👉🏻 Learn more: https://lnkd.in/eswhnvUc #MultiverseComputing #CompactifAI #ArtificialIntelligence #GenerativeAI #SustainableAI #AIInnovation #ModelCompression #SolarImpulse #ClimateTech #EnterpriseAI #GreenTech

Today we are excited to introduce Pulsar 16B, our newest open reasoning model validated by NVIDIA 🚀✨ Built on NVIDIA Nemotron architecture and powered by Multiverse Computing's proprietary technology, Pulsar 16B delivers the reasoning capabilities of leading 30B-class models in a footprint that's nearly half the size. ⚡🧠 🔹 16.15B parameters delivering 30B-class reasoning performance 🔹 Only 3.1B active parameters 🔹 Up to 43% higher throughput and significantly lower latency 🔹 Available in BF16, FP8, and NVFP4 precisions 🔹 Open-source under the Apache 2.0 license We have created a model that makes frontier-grade AI more accessible for enterprises operating in regulated, on-premises, and latency-sensitive environments 🏢🔒 🌟 This is another step toward our vision of making powerful AI faster, more efficient, and deployable wherever organizations need it most. A huge thank you to NVIDIA for their collaboration and support through NVIDIA Model Optimizer, Megatron Bridge, and NVIDIA accelerated computing infrastructure. 👉 Get the model on Hugging Face: https://lnkd.in/e48Wdhpd 📖 Check out Pulsar 16B's benchmarks: https://lnkd.in/eYDX4Uye 💬 We'd love to hear what you will build with Pulsar 16B. #MultiverseComputing #AI #GenerativeAI #LLM #OpenSourceAI #NVIDIA #ReasoningModels #CompactifAI #OpenSource #MachineLearning

AI at the Edge Takes Center Stage at Hewlett Packard Enterprise Discover 2026 🌐🤖 Last week at HPE Discover 2026, our Strategic Account Executive Mark Rabkin joined Michael Merluzzi (Intel) on stage to discuss how AI agents are moving to the edge, enabling faster decision-making, lower latency, and more efficient AI deployments across industries 🏭📡 Beyond the presentation, Mark also spent time at the Intel booth, where Multiverse Computing showcased a live demo demonstrating how organizations can deploy AI more efficiently while reducing infrastructure costs.💡 A big thank you to Hewlett Packard Enterprise and the Intel team for making this collaboration possible. 🙌 Special thanks to Scott Dai, Winnie Koh Vinton, and everyone involved in organizing the session and creating opportunities to engage with customers and partners on the future of AI 🤝✨ Events like HPE Discover highlight the importance of collaboration across the AI ecosystem as we work together to make AI more accessible, scalable, and efficient from the cloud to the edge. #HPEDiscover #HPEDiscover2026 #AI #EdgeAI #AIAgents #Intel #HPE #MultiverseComputing #CompactifAI #EnterpriseAI #AIInfrastructure #GenerativeAI #EdgeComputing #Innovation

Is bigger really better in AI? Or are we just building an expensive, unsustainable multiverse of models? As enterprises scale GenAI, many are defaulting to oversized LLMs that exceed what their use cases actually need, driving up cost, energy and water use, and emissions. Capgemini’s latest POV highlights a shift: 👉 the future isn’t about more powerful models everywhere 👉 it’s about the right model for the right task Nearly half of executives already see AI increasing their emissions, putting sustainability goals at risk. Rightsizing changes the game. 💡 My take: We’re moving toward a fit-for-purpose AI. The winners will design ecosystems, not just deploy models. Use smaller, optimized models where possible, reduce compute, and keep performance where it matters. Innovations like Multiverse Computing’s CompactifAI show how model compression can dramatically reduce model size while preserving accuracy, making sustainable AI at scale far more achievable. 📄 https://lnkd.in/ggf8uAfz Maik Schwalm | Vincent Charpiot | Etienne Grass | Dr. Philippe Cordier | Victor Gaspar | Asier Arburua | Priya Bohra

Wrapping up an inspiring week at VivaTech 2026!🇫🇷🇪🇺 It was fantastic to connect with innovators, business leaders, and AI experts exploring how to turn AI ambition into real business value 🤝 One of the highlights was seeing Co-CEO France Michel Kurek join an insightful panel discussion alongside Helene Costa de Beauregard (French Ministry of Ecological Transition), Aude Giard (Veolia), and Philippe Gouelleu (Sopra Steria), sharing perspectives on how sustainable AI is boosting european companies 🎤💡🌍 Elyes BOUASKAR and Victor Barbier Saint Hilaire were busy showcasing Multiverse Computing and delivering live demos to some of France’s leading companies 🏢 The conversations confirmed a clear trend: enterprises are looking for AI that is not only powerful, but also efficient, scalable, and cost-effective⚡📈 A huge thank you to everyone who stopped by to meet the team, exchange ideas, and discover how our AI models are helping organizations make AI more efficient and accessible. Paris delivered another incredible edition of VivaTech, and we are excited to continue the conversations that started this week ✨🚀 #VivaTech #VivaTech2025 #AI #EnterpriseAI #GenerativeAI #CompactifAI #MultiverseComputing #Innovation #DeepTech

⭐ 𝐖𝐞𝐥𝐜𝐨𝐦𝐢𝐧𝐠 𝐚 𝐧𝐞𝐰 𝐏𝐥𝐚𝐭𝐢𝐧𝐮𝐦 𝐌𝐞𝐦𝐛𝐞𝐫: 𝐌𝐮𝐥𝐭𝐢𝐯𝐞𝐫𝐬𝐞 𝐂𝐨𝐦𝐩𝐮𝐭𝐢𝐧𝐠! The Spanish Chamber of Commerce in Canada is pleased to welcome Multiverse Computing as a new Platinum Member. Multiverse Computing is a Spanish deep-tech company founded in 2019 in San Sebastián and one of Europe's leading quantum software companies. Through its products CompactifAI and Singularity, the company develops quantum and quantum-inspired computing solutions and AI model compression technology applied to finance, energy, manufacturing, and cybersecurity with an established presence in Canada. The Spanish Chamber of Commerce in Canada proud to have Multiverse Computing as part of our community and look forward to strengthening the business relationship between Spain and Canada, and fostering new opportunities for collaboration and growth. #SpainCanadaCC #Member Invest in Spain, Oficina Económica y Comercial de la Embajada de España en Toronto, Economic and Commercial Office of Spain in Ottawa, European Union Chamber of Commerce in Canada (EUCCAN), Federación de Cámaras Españolas de Comercio en América (FECECA)

What if the most valuable moments in any video could find themselves? 🎬 In our latest blog, Jorge Garcelán Gómez, Pre-Sales Engineer at Multiverse Computing, introduces Highlights Studio: an AI pipeline that automatically finds the moments that matter, from sports broadcasts and parliamentary sessions to legal proceedings and live conferences. Main highlights: ✅ Rubric-driven detection: define "important" in plain text, no model retraining required. ✅ Live-ready pipeline: scores and clips footage within broadcast-tight windows. ✅ Compressed Whisper STT: 2.5x faster transcription, same broadcast-grade accuracy, under 1GB. ✅ Model-agnostic: swap or upgrade the VLM without rebuilding the pipeline. 👇 Find the full article below. ow.ly/Xhuz50ZehQn #VideoAI #MediaTech #AIEngineering #ContentAutomation #MultiverseComputing

Yesterday, our Chief Growth Officer Iraia Ibarzabal spoke at La Rioja's Digitalization Strategy Forum 2026–2030 on what applied, sovereign AI looks like in practice, for governments and businesses alike. The infrastructure is there. The political will is there. What brings it together is efficient, compact AI built to run on existing infrastructure, in Spanish, with data sovereignty by design. That's what Multiverse Computing delivers. Thanks to the organizers from Futuro Digital La Rioja and the Gobierno de La Rioja for the invitation, and for putting the right questions on the table. #SovereignAI #EfficientAI #AIStrategy #MultiverseComputing

🚀 A great first day at VivaTech 2026! We had the pleasure of meeting with Christel Heydemann, CEO of Orange, and LE MOUNIER Frédéric, Chief Product Officer of Orange showcasing Comet 1B running on a Raspberry Pi 🤖🍓 demonstrating how powerful AI can run efficiently on lightweight hardware. 🎤 Our Co-CEO France, Michel Kurek, also participated in several interviews, sharing how model compression is making AI more accessible, efficient, and sustainable. 🤝 We connected with customers, partners, and innovators throughout the day at our PwC and France Deeptech stands, discussing the future of efficient AI. An exciting start, with plenty more to come this week! #VivaTech #VivaTech2026 #AI #EdgeAI #LLM #CompactifAI #MultiverseComputing #DeepTech #Innovation