Alibaba Gives Its LLM Structure a Major Overhaul, Rolls Out the New ‘Token Foundry’ Division
Breaking on June 8, Alibaba just reshuffled its AI game plan behind the scenes. They’re merging their Tongyi LLM division with the Future Living Lab to launch a dedicated unit called Token Foundry, which will report straight to Group CEO Eddie Wu (Wu Yongming). This isn’t just a rebrand—several core AI operations are being restructured to align with the new setup. Meanwhile, Zhou Jingren is stepping in as Alibaba’s Chief Scientist to spearhead the newly formed AI Future Research Institute, focusing entirely on pushing boundaries in next-gen AI. On the operational side, Zheng Bo is bringing teams like Happy Horse and Happy Oyster under the Token Foundry umbrella.
WeChat Finally Opens the Gate: Developers Can Now Plug Into the WeChat AI Ecosystem
As of June 8, WeChat officially rolled out the toolkit developers need to tap directly into its growing AI ecosystem. According to WeChat’s official Open Class account, you’ll just need to head over to your Mini Program admin dashboard, find the “AI Capabilities” section, and give the green light to connect. What’s really smart here is that they’ve handed you two paths depending on how your team rolls: an Auto Mode and a Dev Mode. If you go with Auto Mode, you just grant permission for WeChat to peek at your mini-program’s source code during submission—zero extra coding required. Their system will scan your pages and let the AI handle interactions on the fly. If you’d rather get your hands dirty, Dev Mode gives you full creative control to build custom AI features tailored to your specific app logic. Once your project passes WeChat’s review and QA, it’ll be ready to run seamlessly on the platform’s AI stack.
Just a quick heads-up from WeChat: this feature is still rolling out in closed beta, so average users won’t be able to test it out quite yet.
Nvidia Teams Up with SK Hynix for Long-Term Memory Breakthroughs to Power Next-Gen AI Factories
June 8 marked a major handshake between two hardware giants. Nvidia and SK Hynix officially announced a multi-year tech partnership focused on engineering the next wave of memory chips specifically built for scaling AI data centers worldwide. Under the deal, SK Hynix will co-develop custom memory architectures for Nvidia’s upcoming Vera Rubin supercomputer, Vera CPUs, RTX Spark PCs, and the Jetson Thor robotics platform. This move squarely places them at the forefront of Nvidia’s push into AI infrastructure, consumer-grade AI devices, and physical AI systems. Over on the manufacturing floor, things are getting even more futuristic. SK Hynix will start leveraging Nvidia’s CUDA-X libraries and PhysicsNeMo framework to supercharge chip simulation and lithography workflows. Plus, by integrating Omniverse and cuOpt, they’ll be building digital twins of their fabs to drive fully autonomous factory operations.
Tiangong 3.0: The Full-Size General-Purpose Humanoid Robot Set for Mass Production in Late 2026
Good news for robotics enthusiasts: on June 8, the Beijing Advanced Innovation Center for Humanoid Robots and DiGua Robotics jointly revealed that their collaborative masterpiece, Tiangong 3.0, is finally cleared for volume production. Expect the first batch to hit the ground running in the second half of 2026. Under the hood, this thing runs on DiGua’s Xuri S600 compute chip, packed with serious embodied AI power designed to tackle heavy lifting in factories, retail services, and complex 3D environments.
Here’s the real kicker—thanks to a leaner tech architecture and economies of scale, Tiangong 3.0 is projected to slash total manufacturing costs by over 50%. Once it goes mainstream, you’ll see these robots handling assembly lines, warehouse logistics, smart customer service, and even high-risk maintenance jobs in harsh industrial settings.
JD.com and Tencent Lock In a Strategic Partnership to Supercharge AI Agents
According to reports from the STAR Market Daily on June 8, JD.com and Tencent are joining forces to build out next-generation AI Agent ecosystems. By blending JD’s rock-solid supply chain and fulfillment engine with Tencent’s massive user touchpoints and social ecosystem, the duo plans to roll out a unified cross-platform service model. Basically, they’re taking AI Agents out of isolated apps and wiring them directly into a living, breathing network that can anticipate user needs across shopping, messaging, and daily services.
Nebius Drops £1.7 Billion to Build Out Massive AI Cloud Infrastructure Across the UK
On June 8, AI cloud provider Nebius unveiled a major expansion plan, pledging roughly £1.7 billion to break ground on three new Nvidia-powered deployment hubs in the UK. Alongside these new sites, they’re also beefing up their London-based commercial and R&D centers. Every single one of these new facilities will run on Nvidia’s latest full-stack AI factory tech, and once everything ramps up fully by 2027, the combined power capacity is set to hit a staggering 65 megawatts. For context, Nebius already kicked off its UK journey back in November 2025 by deploying its very first Nvidia Blackwell Ultra cluster there.
AMD Pledges Up to £2 Billion Over Five Years to Fuel Britain’s AI Revolution
Also dropping news on June 8, AMD announced a massive five-year investment pipeline targeting up to £2 billion right here in the UK. The goal? Fast-tracking AI innovation, deep-tech research, and next-level computing infrastructure across the country. To make sure those investments hit the ground running, AMD is also teaming up with Imperial College London for a strategic research partnership. Expect this collaboration to fund cutting-edge projects spanning advanced engineering design, multi-physics simulations, and groundbreaking materials discovery—all aimed at keeping the UK competitive at the bleeding edge of AI hardware and software.
