On March 19, Nvidia unveiled its latest AI chip, Blackwell, along with several partnerships and software tools at its annual developer conference. The aim is to maintain its lead in powering the artificial intelligence boom. Here are the key highlights from the event:
Blackwell Chips:
- The B200 “Blackwell” AI chip boasts a remarkable 30 times speed improvement over its predecessor, the Hopper series, particularly in tasks like aiding chatbots such as OpenAI’s ChatGPT and Google’s Gemini.
- Notably, the chip can handle large language models while consuming 25 times less energy and cost compared to the highly popular Hopper series, which played a pivotal role in Nvidia’s growth into a $2 trillion company.
- The new processor is a combination of two silicon squares, each the size of the company’s previous offering.
- Anticipated to hit the market later in 2024, as confirmed by CFO Colette Kress.
- Nvidia is collaborating with TSMC to circumvent delays in chip packaging, as CEO Jensen Huang mentioned in an interview with CNBC, stating the chip’s price range to be between $30,000 and $40,000.
Software Tools:
- Nvidia introduced new software tools designed to simplify the integration of AI systems into business operations.
- Notable among these is Nvidia Inference Microservices (NIM), a cloud-based tool enabling businesses to create and deploy custom applications on their own platforms while retaining full ownership and control of their intellectual property.
Partnerships:
- Nvidia expanded its collaboration with Oracle to assist governments and enterprises in running “AI factories” on local cloud services.
- Additionally, engineering and chip software makers like Ansys, Cadence, and Synopsys will utilize processors based on the Blackwell platform to enhance their software for designing and simulation.
Drive Auto Chips:
- Nvidia further strengthened its partnerships with Chinese automakers, announcing that electric vehicle maker BYD will utilize its next-generation in-vehicle chips, Drive Thor.
Overall, Nvidia’s annual developer conference showcased significant advancements in AI chip technology, software tools, and partnerships, positioning the company for continued leadership in the AI space.