New chips, new strategy
Nvidia has recorded a gigantic sales increase since the premiere of ChatgPT in 2022. Her graphic systems (GPU) have become a standard in training AI models, and the largest cloud companies – Microsoft, Google and Amazon – spend billions of dollars on data centers built on the basis of these chips.
During the GTC conference, Jensen Huang announced The premiere of Blackwell Ultra – a new version of the Blackwell family, which offers greater performance in generating content and calculations of AI. In the future, NVIDIA intends to introduce new generations of chips every year, and not every two years, as before.
Vera Rubin – a breakthrough in architecture
The second key premiere is Vera Rubin – The new GPU generation, which will hit the market in 2026. It includes a CPU processor called Vera and a new GPU architecture called Rubin. This is the first CPU construction developed from scratch by Nvidia, based on the original Olympus core – writes CNBC.
The new systems are to double their performance compared to the previous generation, reaching 50 petaflts during AI calculations. Rubin will also handle up to 288 GB of fast memory, which is a key parameter for the creators of artificial intelligence models.
Nvidia also announced that after the generation of Rubin it will introduce the chips named Feynman, planned for 2028. The company continues the tradition of giving its products the names of outstanding scientists, thus emphasizing their innovative character.
Nvidia is dominated by the AI market
During the conference, NVIDIA also presented new hardware solutions and in the scope of software, including laptops and desktop computers to operate large AI models and Dynamo software, optimizing the use of chip. The company also showed network systems enabling the combination of thousands of GPU processors into a uniform computing unit.
Thanks to these innovations, NVIDIA not only strengthens its leader’s position on the AI market, but also shapes the future of this technology. In the face of growing competition and developing artificial intelligence models, the company consistently sets new standards of performance and efficiency in AI calculations.
