MaHui | Multiple MaHui Member Companies Named to MIT Technology Review's Annual "50 Smartest Companies"

• This article is reprinted from MIT Technology Review

At the recent EmTech China 2025 Global Emerging Technology Summit, the new annual "50 Smartest Companies" list from MIT Technology Review was officially unveiled. By the publication's definition, smart companies share two traits: they develop and deploy new technologies intelligently, and they understand markets and commercial opportunities with equal savvy. They use technological innovation as their opening move, pair it with sustainable business models, and extend their impact globally.

This year, MIT Technology Review adopted "Breaking Boundaries" as its guiding theme and "China's Future" as its focal point. Evaluating candidates across multiple dimensions — technological innovation, global expansion capability, and commercialization potential — it examined innovative forces that are "born in China, led by China, and serving China" (In China, by China, for China). Through continuous breakthroughs in technology, paradigms, and domains, the list identifies enterprises with the potential to drive the next wave of technological development in China and worldwide.

By sector, the top four represented among this year's "50 Smartest Companies" are: AI and computing, robotics and intelligent manufacturing, semiconductors, and life sciences — a snapshot of where emerging technology is heading. The list is not region-specific; it encompasses Chinese companies, international firms, and China-foreign joint ventures, spanning startups, growth-stage companies, and industry giants alike.

MaHui portfolio companies Unitree, Moonshot AI, WITINMEM, Keenon Robotics, METiS Pharmaceuticals, and DeepWise made the list:

The Chinese Force Driving Robotics Toward Mass Deployment

Unitree reshaped robot cost structures through deep integration and manufacturing optimization of core power units and other key modules. This breakthrough lowered the barrier to entry for quadruped and humanoid robots from top-tier research institutions to industrial developers, providing the critical catalyst for moving this technology from the lab into real-world validation.

The Open-Source Pioneer in Large Model Agent Capabilities

Moonshot AI's Kimi K2 model demonstrated efficient computation and agent capabilities at the trillion-parameter scale, achieving core competencies comparable to the world's leading closed-source models. Its release drew intense global attention, shattering the absolute dominance of closed-source models and offering the world a high-performance open-source alternative from China's AI sector.

Breaking Computational Boundaries with Processing-in-Memory Architecture

WITINMEM broke through the traditional chip architecture where memory and computing are separated, achieving true processing-in-memory. This yielded significant improvements in energy efficiency, computational parallelism, and power consumption. By reconciling the trade-off between computing power and power draw in conventional mobile AI devices, the company enables higher computational performance in visual processing and AI terminals, providing the industry with new momentum through the deep integration of innovative computing architecture, underlying devices, and software ecosystems.

Turning Robots into Productive Forces in Service Industries

Keenon Robotics focuses on large-scale deployment of service robots in real commercial environments. Through engineering capabilities in multi-robot collaboration and stable operation, it transforms robots into reliable productive forces. Its solutions effectively solve the core challenge of stable robot service in complex human traffic environments, and its leading technical performance has secured broad global deployment.

Solving Nano-Delivery Challenges Through an AI Platform

METiS Pharmaceuticals developed a multi-organ-targeting LNP delivery system powered by its self-developed AI-driven nano-delivery system design platform. By building a large-scale lipid library and patent moat, the company has advanced RNA and cell gene-editing drugs in delivery efficiency, safety, and clinical application, laying a solid foundation for breakthroughs and commercialization in related R&D.

Using AI to Upgrade the Scientific Research Paradigm

DeepWise is dedicated to using AI to upgrade the scientific research paradigm. Through AI for Science, it achieves higher-precision molecular dynamics simulations, opening new scales in scientific computing. Relying on scientific foundation models and micro-scale industrial infrastructure, it builds next-generation intelligent research systems and ecosystems for materials science, life sciences, chemistry, and energy materials — enhancing the depth, breadth, and efficiency of scientific exploration.

Original article and full list: https://www.mittrchina.com/news/detail/15231