Alphabet’s $80B AI Infrastructure Play, Backed by a $10B Berkshire Hathaway Commitment
June 1st brought a massive announcement from Google’s parent company, Alphabet: they’re officially planning to raise a staggering $80 billion through equity financing. The entire goal here is to supercharge their AI infrastructure build-out. And they’re not going it alone—Berkshire Hathaway has already stepped up with a firm $10 billion private placement commitment to back the initiative.
Unitree’s Single Humanoid Robot Model Crosses the 11,000-Unit Production Mark
June 2nd delivered a serious milestone for Unitree Tech. They just revealed that one specific model of their bipedal humanoid has officially rolled off the assembly line, hitting roughly 11,000 units total. To keep things crystal clear, Unitree emphasized this tally only covers their two-legged walkers. No wheeled platforms or other humanoid variants are mixed into the count. The data runs right through May 2026, showing some serious scaling momentum.
Alibaba Drops Qwen3.7-Plus: A New Multimodal AI Agent Model
Fresh off their official channels, Alibaba just rolled out the Qwen3.7-Plus on June 2nd, and it’s a serious leap for multimodal AI agents. Building on the solid text foundation of Qwen3.7, this upgrade dramatically enhances vision-to-language processing while keeping all the heavy-lifting agent capabilities intact—think coding, tool integration, and seamless productivity workflows.
UBTECH Launches Pre-Sales for the World’s First Full-Size Hyper-Bionic Humanoid
UBTECH is making history. Their “U-World” hyper-bionic humanoid is now officially up for pre-sale on JD.com, with a full public launch slated for June 30th. If you’re wondering what sets it apart, it’s billed as the very first full-size, hyper-bionic humanoid robot hitting the consumer market.
Dobot Unveils Kongyi DobotWAM: A Breakthrough in Embodied AI
June 2nd also saw Dobot drop their in-house “Kongyi” DobotWAM, an embodied AI model that’s turning heads in the robotics space. It’s been rigorously tested on the LIBERO benchmark suite, crushing tasks across spatial reasoning, object generalization, command comprehension, and long-horizon execution. The result? A jaw-dropping 99.25% average success rate across all four standard LIBERO test sets.
Tencent Cloud Slashes DeepSeek-V4 Model Pricing, Cuts Up to 97.5%
Big savings are coming for developers. On June 2nd, Tencent Cloud announced a massive price drop for the DeepSeek-V4 series on their agent development platform. Effective June 3rd at midnight Beijing time, rates are getting slashed by up to 97.5%. Don’t worry about performance taking a hit—this is purely a pricing play, so all model capabilities and service tiers stay exactly as they are.
Arm CEO: $15B AI Chip Revenue Target Could Be Beat Ahead of Schedule
Rene Haas, CEO of Arm, dropped some exciting news on June 2nd. Thanks to AI demand completely outpacing early forecasts, Arm might just hit its $15 billion autonomous chip sales target well before the end of the decade. Meta Platforms is already locked in as the first major client for Arm’s upcoming AGI CPU chip. Specs-wise, you’re looking at an absolute beast with up to 136 cores running at a 300-watt power draw.
Aptiv and Nvidia Deepen Ties to Fast-Track Mass-Produced Edge AI
On June 1st, Aptiv and Nvidia announced they’re leveling up their partnership to bring mass-production-ready Edge AI to life. They’re teaming up to fully upgrade the NVIDIA Jetson ecosystem—next-gen platforms like the Jetson Thor included—transforming it into a commercially viable, factory-ready edge AI solution that’s primed for immediate real-world deployment.
Anthropic Secretly Files Draft IPO Paperwork with the SEC
June 2nd brought some quiet but massive corporate news: Anthropic has confidentially filed its initial IPO paperwork with the SEC. Right now, the exact number of shares and the final pricing are still under wraps, but the move clearly signals they’re gearing up to go public.
