Nvidia has unveiled a series of cutting-edge technologies designed to maintain its pivotal position in the booming AI computing market.
At Asia’s premier electronics expo, Computex in Taiwan, CEO Jensen Huang showcased new product offerings and strengthened strategic partnerships essential to global tech supply chains.
During his address, Huang announced the timeline for Nvidia’s next-generation GB300 systems, aimed at AI workloads and due in the third quarter. These systems will succeed the premier Grace Blackwell AI models currently being deployed by major cloud service providers.
Innovatively, Nvidia is now providing a revamped version of comprehensive data center systems, featuring the NVLink Fusion series. These products offer flexibility, allowing users to integrate their own CPUs with Nvidia’s AI chips or opt for Nvidia CPUs alongside other providers’ AI accelerators—a strategic shift from Nvidia’s prior approach of exclusively using its components.
This move enhances Nvidia’s offerings by incorporating critical connectivity features essential for rapid processor-accelerator communication, giving data center clients more options while maintaining Nvidia’s core technological presence.
In response to tech giants like Microsoft and Amazon, who are developing custom processors and accelerators, Nvidia’s partnerships are diversifying. MediaTek, Marvell Technology, and Alchip Technologies will develop tailored AI chips compatible with Nvidia systems.
Simultaneously, Qualcomm and Fujitsu are crafting processors designed to work with Nvidia accelerators, further embedding Nvidia technology within the competitive landscape of data centers.
