NVIDIA Unveils Blackwell Platform to Slash AI Costs by 25x
NVIDIA launches its new Blackwell GPU platform, enabling real-time generative AI on trillion-parameter models at a fraction of the previous cost. Major tech companies including Google, Amazon, and Microsoft are already preparing to adopt the technology.
NVIDIA officially launches the Blackwell platform, marking a significant leap in accelerated computing that powers a new era of generative AI. The new Blackwell GPU architecture enables organizations to build and run real-time AI on trillion-parameter large language models at up to 25 times less cost and energy consumption than previous generations.
The platform incorporates six transformative technologies, including new Tensor Cores, advanced NVLink, and the TensorRT-LLM compiler, which work together to drastically reduce inference operating expenses. Beyond generative AI, these accelerators unlock critical breakthroughs in data processing, engineering simulation, electronic design automation, computer-aided drug design, and quantum computing.
Major players across the tech industry show widespread and immediate adoption of the Blackwell architecture. Every major cloud provider and leading AI company, including Amazon Web Services, Google, Meta, Microsoft, OpenAI, Oracle, Tesla, and xAI, integrates the new chips into their infrastructure to power future AI discoveries.