NVIDIA CEO Jensen Huang took the stage at CES 2026 to outline the company’s expanding vision, blending advancements in automotive AI, supercomputing, and gaming technology. While much of the presentation served as a refresher on ongoing initiatives, several key announcements highlighted NVIDIA’s push into new frontiers. A major focus was the automotive sector with the introduction of Alpamayo, a new family of open-source reasoning models for autonomous vehicles. The flagship model, Alpamayo 1, is a 10-billion parameter system that employs a chain-of-thought approach designed to mimic human-like reasoning. It tackles unexpected driving scenarios by breaking them down into smaller problems, determining the safest path forward, and explaining its logic at each step. A companion model, AlpaSim, enables closed-loop training for rare driving situations. Huang revealed that the 2025 Mercedes-Benz CLA will be the first vehicle equipped with NVIDIA’s full autonomous driving stack, including Alpamayo, reinforcing his statement that the vision is for every car and truck to eventually be autonomous. The presentation also provided an update on the Vera Rubin platform, NVIDIA’s next-generation GPU architecture first announced in 2024. The company has now begun production on a supercomputer utilizing this technology. A single Vera Rubin CPU packs 88 custom Olympus cores and 1.5TB of system memory, totaling 227 billion transistors. A paired Rubin GPU contains 336 billion transistors. Each supercomputer node combines one of each of these powerful components. Following the main event, a separate briefing unveiled significant updates for gamers: DLSS 4.5 and G-Sync Pulsar. The latest iteration of NVIDIA’s Deep Learning Super Sampling was trained on a second-generation transformer model, promising reduced visual artifacts like ghosting and shimmering for a more stable image during fast motion. DLSS 4.5 also introduces two new features arriving in the spring: 6x multi-frame generation and dynamic frame generation. The former allows a GeForce RTX 50-series GPU to generate five frames for every traditionally rendered frame, aiming to fully utilize high-refresh-rate displays, such as pushing a 4K 240Hz monitor to its limit with a card like the RTX 5090. Dynamic frame generation scales this process intelligently, producing more frames in demanding scenes and fewer in simpler ones to optimize performance. G-Sync Pulsar represents the next evolution in NVIDIA’s display technology for motion clarity. By pulsing a monitor’s backlight in a new way, NVIDIA claims it can deliver a perceived motion clarity equivalent to a 1000Hz refresh rate. These new displays will also include sensors to automatically adjust screen brightness and color temperature based on ambient lighting conditions. The first G-Sync Pulsar monitors will be available for pre-order starting January 7, 2026.
