Nvidia goes all in on AI robotics at Computex 2025 in Taipei
Nvidia came out swinging at Computex Taipei 2025 on Monday, unveiling a futuristic lineup of humanoid robotics tools and customizable AI server infrastructure aimed at accelerating the global AI boom.
At the heart of the announcements was Nvidia Isaac GR00T-Dreams, a powerful new platform that helps developers generate vast training datasets to teach humanoid robots how to function and adapt in dynamic environments.
Jensen just announced NVIDIA’s Isaac GR00T N1.5 and GR00T-Dreams blueprint at COMPUTEX 2025:
⦿ Isaac GR00T N1.5 is the first update to NVIDIA’s open, generalized, fully customizable foundation model for humanoid reasoning and skills.
⦿ “Human demonstrations aren’t scalable —… pic.twitter.com/E77UYIy2c7The move is a major step in what CEO Jensen Huang has called “the next trillion-dollar industry”—physical AI.
Nvidia is now staking its claim beyond just chips, focusing on the software and infrastructure to power a new generation of intelligent machines, both in industrial settings and, eventually, in homes.
Also grabbing headlines was the debut of NVLink Fusion, a next-gen server architecture allowing customers to mix and match Nvidia’s Grace CPU with third-party AI chips, or integrate Nvidia’s GPUs with custom CPUs.
Electricity. The Internet. Now it’s time for another major technology, #AI, to sweep the globe.
NVIDIA founder and CEO Jensen Huang took the stage at a packed Taipei Music Center Monday to kick off #COMPUTEX2025, captivating the audience of more than 4,000 with a vision for a… pic.twitter.com/7ORgr6UmuUThe goal: give hyperscalers and enterprises more flexibility in designing AI-first data centers.
“Using NVLink Fusion, hyperscalers can work with the NVIDIA partner ecosystem to integrate NVIDIA rack-scale solutions for seamless deployment in data center infrastructure,” the company said in a statement.
Adding to the firepower, Nvidia announced RTX Pro Blackwell servers powered by its Blackwell Server Edition GPUs.
These new systems are designed to handle “virtually every enterprise workload,” from engineering simulations to large-scale agentic AI deployments.
#NVIDIA just unveiled a beast at #Computex: the RTX PRO Server.
This next-gen enterprise AI and industrial computing platform is built for real-time digital twin simulation and ultra-high-performance workloads.#RTXProServer #AIInfrastructure #DigitalTwins #Computex2025 pic.twitter.com/vy2QRIWNexTo extend access to AI development tools globally, Nvidia launched DGX Cloud Lepton, a cloud-based AI processing platform built in collaboration with CoreWeave, Foxconn, SoftBank, and others.
The system will tap into a global GPU cloud network, enabling customers to build and roll out AI software at scale.
The announcements land at a pivotal time for Nvidia, fresh off a market boost following the US government's reversal on proposed export restrictions.
The company is also set to deliver hundreds of thousands of AI processors to Saudi-backed startup Humain over the next five years, as revealed during President Trump’s visit to the Kingdom.
#Nvidia CEO Jensen Huang (#黄仁勋) announced that the company would be opening up an office in Taiwan, dubbed "Nvidia Constellation."#COMPUTEX2025 pic.twitter.com/qgWzCsdQea
Despite a rocky few months on the stock market—linked to export controls, data center lease rumors, and semiconductor tariff fears—Nvidia’s long-term trajectory appears more AI-fueled than ever.
Shares are still up 43% over the past year, even as short-term volatility continues.
From cloud to humanoids, Nvidia is clearly signaling that the future isn’t just artificial intelligence—it’s artificially embodied.