The recent CES 2026 featured a major presentation outlining a significant shift in the AI computing landscape. The core revelation was a new architectural approach, positioned not merely as a faster chip, but as a systemic re-engineering of the entire AI infrastructure stack. This move addresses the industry’s fundamental challenge: computational demand is growing exponentially, far outpacing what traditional semiconductor scaling alone can provide.
The proposed solution hinges on extreme co-design, optimizing the entire compute path from processors and memory to networking and storage, alongside crucial software and model-layer innovations. A key promise is backward compatibility, allowing performance upgrades via software without requiring clients to rebuild their physical infrastructure from scratch. This architectural leap reportedly delivers a tenfold performance increase while slashing token generation costs by 90%. System-level innovations also focus on operational efficiency, boasting features like hot-swappable components to maximize data center uptime and power-smoothing techniques to eliminate energy waste.
Regarding market strategy, the vision extends far beyond selling hardware. The business model is deeply intertwined with cloud service providers who act as channels to major AI model companies. An unexpected growth driver cited is the explosive rise of open-source models, which now account for a significant portion of all AI-generated tokens and drive demand for public cloud compute. Another anticipated growth vector is the return to a key international market, with specific high-performance chips already in production for that region following regulatory approvals.
The ambition isn’t limited to data centers. Clear timelines were given for other transformative technologies. In autonomous vehicles, the focus is on providing the full technology platform—encompassing training, simulation, and onboard computers—to the automotive industry at large, predicting hundreds of millions of such vehicles in the coming decade. For robotics, a landmark prediction was made: robots with human-level capability are expected to emerge by 2026. The argument presented is that such robots will act as “AI immigrants” to address critical global labor shortages, taking on undesirable or unfillable roles and, in turn, creating more human employment in a growing economy.
The underlying competitive advantage claimed is a “full-stack” ecosystem. This encompasses everything from proprietary CPUs and GPUs for various applications, to advanced AI models for physics and reasoning, to development platforms for virtual training. This vertical integration, combined with deep partnerships across the AI industry and strategic investments in core technologies and the broader ecosystem, is presented as a moat that is difficult for competitors to cross rapidly. The overarching narrative frames this not as just another tech cycle, but as the beginning of a new industrial revolution where AI evolves from a tool into a form of labor itself.