Nvidia launches Vera Rubin AI platform with six custom chips per unit, boosting AI computing power and efficiency, adopted by major tech firms, with products expected late 2026.

With systems scalable into pods of more than 1,000 units, Nvidia's Vera Rubin AI platform boasts six custom chips per unit that can deliver up to five times the AI computing power and ten times greater efficiency in processing AI tokens. Large AI models can be trained more quickly thanks to the platform's Rubin GPU, Vera CPU, and new networking and storage technologies, which also improve performance during lengthy conversations. Major tech companies like Microsoft, Amazon, Google, and CoreWeave are adopting it, and products are anticipated in late 2026. Additionally, Nvidia purchased Groq's AI chip knowledge and made its Alpamayo self-driving car software open-source. Older H200 chips are still in high demand in China despite increased competition, and export licenses are still pending.