TLDR: Morph Labs has unveiled ‘Liquid Metal,’ a new technology designed to address the infrastructural bottlenecks in scaling intelligent AI agents. This innovation, building on their Infinibranch solution, promises ultra-fast startup times, memory snapshotting, and the ability to branch virtual machines, enabling a new era of AI reasoning and verified superintelligence.
San Francisco, CA – Morph Labs, a pioneering force in artificial intelligence infrastructure, today announced the launch of ‘Liquid Metal,’ a groundbreaking technology poised to transform the landscape of agentic AI. Unveiled by Founder and CEO Jesse Han at the AI Engineer World’s Fair, Liquid Metal is the latest iteration of Morph’s Infinibranch solution, specifically engineered to overcome the limitations of current cloud computing environments in supporting advanced AI agents.
Han emphasized the critical need for a new infrastructural paradigm as AI agents transition from experimental tools to practical co-workers. He articulated a vision where machines can react to their thoughts with zero latency, stating, “Thinking at the speed of light must be just as lonely as moving at the speed of light.” This sentiment underscores the core challenge Liquid Metal aims to solve: the fundamental bottleneck in scaling intelligent agents.
Liquid Metal directly addresses this constraint by reimagining virtualization, storage, and networking. The new offering boasts impressive performance metrics, including startup times under 100 milliseconds, the ability to snapshot memory and process states in under 200 milliseconds, and the capacity to branch running virtual machines into dozens of instances within milliseconds. This unprecedented speed and flexibility also extend to indefinite persistence and autoscaling capabilities, ranging from zero to infinite instances, with GPU support anticipated in Q4 2025.
The innovation at the heart of Liquid Metal is what Han terms “Git for compute.” This concept allows for “all mistakes become reversible,” empowering AI agents to explore diverse computational paths without irreversible consequences. This ‘reasoning-time branching’ facilitates efficient simulation and parallel testing of multiple solutions, enabling AI to scale against verification processes and identify optimal answers.
Also Read:
- Google Unveils Gemini 2.5 Deep Think: A Leap in Advanced AI Reasoning
- Valory’s Decentralized AI Agents Revolutionize DeFi with Enhanced Transparency and Control
This foundational shift in infrastructure is not merely about enhancing speed; it’s about ushering in a new era of AI reasoning. Han envisions a future where multi-agent systems can branch and reverse their operations, utilizing software during reasoning time and scaling against rigorous verification. To spearhead this ambitious initiative towards ‘verified superintelligence,’ Morph Labs has announced the appointment of Christian Szegedy, former xAI co-founder and Grok-3 code reasoning lead, as their Chief Scientist. Szegedy’s profound expertise in formal verification and large language models will be instrumental in the development of Magi-1, an Infinibranch-native verified reasoning agent slated for release in Q1 2026. This strategic collaboration underscores Morph Labs’ commitment to building the essential infrastructure for AI to seamlessly interact with complex software environments at an unprecedented scale and with robust algorithmic guarantees.
