CES 2026 is officially underway in Las Vegas, and if the first two days of press conferences tell us anything, AI has completely consumed the tech industry's hardware strategy. Nvidia kicked things off with its new Rubin computing architecture and autonomous vehicle models, while AMD and Amazon made aggressive pushes to bring generative AI to consumer PCs and smart home devices. This isn't just another product cycle. It's a fundamental shift in how every major tech company builds and sells hardware.
Las Vegas is packed with the usual CES chaos right now, but there's a clarity to the messaging that wasn't there before. Every major hardware company is essentially asking the same question: how do we make AI useful on the device people are actually holding, not just in the cloud?
Nvidia CEO Jensen Huang opened Monday with the kind of presentation that's become his signature—part victory lap, part technical deep dive. The company spent two years turning its GPUs into the default computing backbone for AI training and inference. Now it's moving to the next phase. Rubin, the new architecture rolling out in the second half of 2026, brings significant speed and storage upgrades over the current Blackwell generation. But more importantly, it signals Nvidia's commitment to staying ahead of an increasingly competitive chip market where AMD, Google, and others are designing custom silicon for AI workloads.
What caught more attention than raw specs was Nvidia's push into embodied AI. The company showed off its Alpamayo family of open-source models built specifically for autonomous vehicles, framing itself as infrastructure that roboticists and automakers can build on. That mirrors a broader strategy: positioning Nvidia as Android for robots, providing the software layer that lets companies focus on hardware and applications rather than training foundational AI models from scratch.
AMD followed with its own offensive. The company's Ryzen AI 400 Series processors arrived with backing from heavy hitters like OpenAI President Greg Brockman and AI researcher Fei-Fei Lei. The message was direct: AI shouldn't be confined to cloud data centers. AMD is betting consumers want generative AI capabilities running locally on their laptops, preserving privacy and reducing latency. It's a calculated move against Nvidia's dominance, essentially saying "you don't need us to dominate consumer PC chips like we do in data centers."
The robotics announcements proved equally significant. , owned by Hyundai, revealed something almost no one expected: it's partnering with DeepMind to train and operate its humanoid Atlas robots, not competing against it. The announcement came during Hyundai's press conference, and the implication is striking. The companies showed a new generation of Atlas alongside the existing model, both powered by Google's AI research. This suggests that building world-class humanoid robots isn't just about hardware engineering anymore. It's about having access to cutting-edge AI infrastructure.