NVIDIA founder and CEO Jensen Huang took center stage at CES 2026 in Las Vegas to reveal the company’s latest developments in artificial intelligence. The highlight of his keynote was the introduction of the Rubin platform, NVIDIA's first extreme-codesigned AI platform, aimed at reducing computing costs and enhancing performance across various sectors. This announcement marks a pivotal moment in the tech industry, as AI continues to integrate deeper into everyday applications, aligning with NVIDIA's vision of ubiquitous AI solutions.
Huang emphasized the monumental economic impact of AI, estimating that over $10 trillion of computing from the last decade is now being modernized through AI. This transformation is not just about accelerating performance; it addresses long-standing challenges such as high inference costs and scaling limitations, which are crucial as businesses increasingly rely on AI for intelligence-driven applications.
As the tech landscape evolves, NVIDIA's push into open models for multiple domains—from healthcare to autonomous driving—positions it advantageously in a rapidly changing market. The company's commitment to open frameworks suggests a strategic focus on collaboration and innovation, which could drive further adoption across industries.
The Rubin platform, named after pioneering astronomer Vera Rubin, marks a significant leap forward from NVIDIA's previous Blackwell architecture. Designed with an extreme codesign approach, it integrates multiple specialized components—including the new Rubin GPUs, Vera CPUs, and advanced networking technologies—aimed at streamlining AI processing and reducing bottlenecks.
With 50 petaflops of inference capability, the Rubin GPUs offer transformative performance, promising a 10x reduction in AI inference costs and reducing GPU training times by 4x. Such capabilities are critical as enterprises grapple with the economics of deploying large-scale models. Prior to the Rubin platform, the Blackwell architecture had already established NVIDIA as a leader, driving 66% year-over-year revenue growth in data center business. The leap to Rubin could signify an accelerated timeline for competitors eager to catch up.
Huang's presentation also detailed features like the NVIDIA Inference Context Memory Storage Platform, enhancing long-context inference and improving cost-efficiency metrics significantly. This level of integration signals NVIDIA's intent to dominate by simplifying complex AI deployments, an area many organizations struggle with under traditional computing models.
A critical aspect of Huang's presentation was NVIDIA's focus on open models, designed to support the development of AI across various industries. Huang noted that 80% of AI startups are utilizing these open frameworks, underpinning a growing ecosystem of innovation that includes models tailored for healthcare, climate science, and robotics.
NVIDIA's initiative to make advanced models like Alpamayo for autonomous vehicle development available openly reinforces its strategy to broaden the participation of businesses and researchers in AI evolution. As Huang articulated, this open model framework allows any enterprise or individual to develop, evaluate, and deploy intelligent applications, creating a democratized platform for AI innovation.
Historically, the reliance on proprietary models has limited access and stifled innovation. By fostering an environment where shared resources are the norm, NVIDIA may catalyze breakthroughs similar to those seen during the advent of open-source software.
A tangible example of NVIDIA's commitment to practical applications of AI was showcased during Huang's discussion of the newly developed Alpamayo model. Designed for autonomous vehicles, it features sophisticated reasoning capabilities that enable vehicles to understand and navigate complex environments. The demo included a Mercedes-Benz CLA, set to hit the roads soon, showcasing how AI-defined driving can become a reality.
This initiative aligns with NVIDIA's broader vision of fully autonomous vehicles influencing transportation norms and safety standards. With a recent EuroNCAP five-star safety rating for the CLA, NVIDIA's entry into the automotive sector promises to drive consumer confidence while also challenging competitors in the autonomous market.
Furthermore, allied with partners like Siemens, NVIDIA is leveraging its technologies to integrate AI across manufacturing and production processes. Such strategic collaborations could redefine operational efficiencies, turning manufacturing plants into adaptive, intelligent ecosystems.
Huang's CES keynote addressed not just the technology but also the future of AI in the workplace and everyday life. With the introduction of the DGX Spark, capable of high-performance AI computations from personal devices, NVIDIA signals a commitment to making AI tools accessible for individual developers and businesses alike.
The company’s future roadmap features continued development and refinement of the Rubin platform, focusing on sustaining its competitive edge as industries adopt these cutting-edge technologies. By emphasizing "technology leadership," Huang articulated a clear path forward for NVIDIA—one that centers on fostering innovation in partnerships and empowering businesses with responsible, scalable AI capabilities.
As the industry watches, the next chapters of NVIDIA's story will undoubtedly shape the future of AI, further solidifying its position as a leader in the field. The blend of advanced hardware infrastructure with open, adaptive software models sets a profound benchmark that could redefine what is possible in AI technology.
Source: Read the full story here
