Fear not China’s giant robots

China has another robot moment, but Tencent and Alibaba still struggle to profit from AI. Meanwhile, some of their alumni are fueling SEA’s AI boom.

Read from your browser The Prompt Welcome to The Prompt, your Monday dive into the world of AI that puts Asia front and center. From Big Tech's power players to the region's scrappy disruptors, we cover it all. Subscribe to get full access to our AI reporting. Ensure our articles appear in your Google searches: Hello Betamax, I'm a fan of the Gundam series and built my fair share of plastic models back in the day. So when I saw Unitree's latest giant robot come to life, the franchise was the first thing that popped into my mind. But even in my dreams, I probably wouldn't have US$650,000 to buy the half-ton monster. This newest robotics debut is just one of the many inventions China wants to wow us with. At the country's spring gala earlier this year, its robots stunned the world with martial arts skills Still, once the spectacle was over, the bigger question was: then what? For all their viral robot videos, China's tech giants still haven't fully cracked the business of AI. Despite the hype around Alibaba's Tongyi Qianwen and Tencent's Hunyuan models, turning AI into meaningful revenue remains an uphill battle. Chinese firms are still widely seen as lagging behind US players in both frontier model capabilities and commercialization. That gap, of course, didn't emerge in a vacuum. Washington's export controls have restricted China's access to Nvidia's most advanced chips, limiting the computing power needed to train and run cutting-edge AI systems. The geopolitics around AI have become so intense that Nvidia CEO Jensen Huang even joined Donald Trump's recent trip to China - albeit so last-minute that the internet quickly turned it into a meme. Still, even if Alibaba and Tencent haven't fully solved the AI monetization puzzle themselves, their ecosystems are already shaping the next generation of AI startups across Southeast Asia. As my colleague Glenn and Duc explore in our top AI read this past week, many of the region's emerging AI founders come directly from the talent pipelines of ByteDance, Tencent, Alibaba, and their sprawling alumni networks. Miguel Cordon, journalist TOP AI READS FROM OUR DESK 1️⃣ Tracking the alumni networks powering SEA's AI startups Be nice to your colleagues. One day, they might leave to start an AI company and bring you along. Jokes aside, this is already happening across the industry. Our latest tracker found that former employees from companies like Microsoft, Grab, Alibaba, ByteDance, and Tencent are increasingly becoming the founders behind Southeast Asia's next wave of AI startups. And the money is following them. AI startups in Southeast Asia founded by former ByteDance and Tencent executives have raised more than US$2 billion each, according to Tech in Asia data, making them the region's most heavily funded founder networks so far. 2️⃣ How AI-powered tools are clearing up construction's data debris As my colleague Scott points out, the construction industry is filled with paperwork, compliance checks, site updates, and endless busywork - exactly the kind of repetitive processes AI startups love targeting. We're seeing a similar pattern in healthcare as well. In a previous story, Glenn explored the rise of AI-powered medical scribes in Southeast Asia, where startups are trying to reduce the administrative burden on doctors. It wouldn't be surprising to see AI crawl its way into more niches as the tech matures. WHAT WE THINK 1️⃣ Qwen may give Lazada new life The Qwen app now has access to more than four billion products across Taobao and Tmall. It can also manage the entire shopping process, from product discovery and comparison to checkout and after-sales service. The rollout marks one of Alibaba's clearest attempts yet to show returns on its massive AI investments. It could also help the company claw back market share from Temu parent firm PDD Holdings. Because Alibaba also owns Lazada, the integration could eventually extend to Southeast Asia. That may give Lazada a much-needed boost as it continues to lose ground to Shopee in the region's ecommerce race. What's interesting in this story is that Alibaba seems to be pushing ecommerce beyond the usual "search, scroll, click" model. Instead, Qwen is becoming the interface. Users may no longer need to manually browse through pages of listings if the AI can recommend products, compare options, complete purchases, and even handle customer support for them. If so, Shopee and TikTok Shop may have to respond quickly because letting an AI agent handle your entire shopping journey is a very different game from simply adding a chatbot to an ecommerce app. 2️⃣ Taking control of the conversation This past week, two AI labs revealed their respective visions for what voice AI should actually feel like. Meta's Muse Spark-powered Voice Conversations are designed to feel more human: users can interrupt mid-sentence, switch topics, or even swap languages while speaking. The feature is rolling out across the Meta AI app first before expanding to WhatsApp, Instagram, and Meta's smart glasses. That gives Meta a major distribution advantage in Southeast Asia, where WhatsApp already dominates daily communications in many markets. Integrating AI voice directly into existing consumer behavior may prove far more powerful than asking users to download yet another standalone AI app. At the same time, Thinking Machines Lab - the startup founded by former OpenAI executive Mira Murati - is making a very different argument about the future of AI voice. The company believes today's leading voice assistants, including those from OpenAI, still rely too heavily on rigid, turn-based conversations: humans speak, the AI waits, then responds. Thinking Machines Lab's approach is more ambitious. Its system processes audio, video, and text simultaneously, allowing the model to listen, reason, and speak at the same time. In theory, that means it can better handle interruptions, pauses, and overlapping speech instead of treating them as mistakes. If Meta wins on consumer distribution, Thinking Machines Lab may be positioning itself to win somewhere else entirely: enterprise environments, where conversations are messy, fast-moving, and rarely wait for your turn to speak. STANDOUT AI DEALS 1️⃣ DeepMind spinout Isomorphic Labs raises $2.1b Isomorphic Labs' latest funding round - backed by sovereign wealth funds including Temasek and MGX - signals growing government confidence in AI-driven drug discovery. The company already has partnerships with pharmaceutical giants like Novartis and Johnson & Johnson. Still, the path to commercialization remains slow. Isomorphic recently pushed its target for beginning human trials from the end of last year to the end of 2026, a reminder that biotech timelines remain stubbornly difficult even in the AI era. Even so, the deal stands as the second-largest fundraising round for a biotech company. If Isomorphic reaches its clinical milestone before the end of the year, that massive bet may start to look justified. 2️⃣ Vertex co-leads $2m for Malaysian chip startup FusionAP FusionAP was founded by former executives from Intel and TSMC and focuses on advanced semiconductor packaging, the process of integrating multiple chips into a single unit. As the startup uses fresh capital to expand its R&D and pilot production capacity, FusionAP is positioning itself to become a key player in integrated chip stacks, a technology increasingly critical to the global AI infrastructure boom. These advanced packaging techniques help deliver higher performance and better energy efficiency as AI workloads become more demanding. 3️⃣ Ex-Meta researcher's AI startup raises $650m Recursive Superintelligence has only just emerged from stealth mode, yet it's already reportedly valued at over US$4.6 billion. Is that valuation justified? The startup is led by engineers from AI heavyweights including OpenAI, Google, and Meta, exactly the kind of pedigree investors are currently rewarding. But unlike many frontier labs focused on scaling ever-larger foundation models, Recursive Superintelligence is betting on something else: automating the AI development pipeline itself. In other words, using AI to build better AI with fewer human engineers involved. If that approach works, fears of an AI "jobpocalypse" may hit the AI industry itself before everyone else. Catch up on past issues of the newsletter here. In the meantime, if you have any feedback or ideas, feel free to get in touch with Huong, our managing editor, at huong@techinasia.com. ADVERTISE | SUBSCRIBE | HIRE | FIND JOBS P.S. Don't miss out on the biggest tech news and analysis. Add newsletter@techinasia.com to your address book, contacts, or safe sender list. Or simply move us into your inbox. Tech in Asia thanks our partners for supporting our newsletters. Sponsorship has no influence on editorial content. Too many emails? Switch to a different frequency or get new content through our preference center, or unsubscribe. You can also remove yourself from all Tech in Asia emails over here. Copyright © 2026 Tech in Asia, All rights reserved. Tech in Asia (Podium Level 3), 1000 Toa Payoh North, News Centre, Singapore 318994 A member of The Business Times.