By Jacob Gronholt-Pedersen and Supantha Mukherjee COPENHAGEN/STOCKHOLM (Reuters) - Nvidia CEO Jensen Huang said on Wednesday a design flaw with its latest Blackwell AI chips which impacted production has been fixed with the help of longtime Taiwanese manufacturing partner TSMC. Nvidia unveiled Blackwell chips in March and had earlier said they would ship in the second quarter but were delayed, potentially affecting customers such as Meta Platforms, Alphabet's, Google and Microsoft. "We had a design flaw in Blackwell," Huang said. "It was functional, but the design flaw caused the yield to be low. It was 100% Nvidia's fault." According to media reports, the delay in production had caused tensions between Nvidia and TSMC but Huang dismissed that as "fake news". "In order to make a Blackwell computer work, seven different types of chips were designed from scratch and had to be ramped into production at the same time," he said. "What TSMC did, was to help us recover from that yield difficulty and resume the manufacturing of Blackwell at an incredible place." Nvidia's Blackwell chips take two squares of silicon the size of the company's previous offering and binds them together into a single component that is 30 times speedier at tasks like serving up answers from chatbots. At a recent Goldman Sachs conference the CEO said the chips will now ship in the fourth quarter. Huang was in Denmark on Wednesday to launch a new supercomputer named Gefion, which boasts 1,528 graphic processing units (GPUs) and was built in partnership with Novo Nordisk Foundation, Denmark's Export and Investment Fund and Nvidia What obout this

This news article provides additional insights into the production and launch of Nvidia's Blackwell AI chips.

1. Design Flaw and Production Delay: Nvidia CEO Jensen Huang confirmed that there was a design flaw with the Blackwell AI chips that impacted production. This design flaw caused a low yield rate, leading to a delay in the shipment of the chips. The company had initially planned to ship the Blackwell chips in the second quarter, but the delay has been pushed to the fourth quarter.
2. TSMC's Role in Resolving the Issue: The production issue was resolved with the help of TSMC, Nvidia's longtime Taiwanese manufacturing partner. TSMC assisted in recovering from the design flaw and resumed the manufacturing of Blackwell chips, allowing for a more rapid production ramp-up.
3. Blackwell Chips' Performance: The Blackwell chips are designed to be significantly faster than Nvidia's previous offerings, with a single component being 30 times speedier for tasks like serving up answers from chatbots. This performance enhancement is a key selling point for the new chips.
4. Launch of the Gefion Supercomputer: In a separate development, Nvidia launched a new supercomputer named Gefion in Denmark. This supercomputer is equipped with 1,528 graphic processing units (GPUs) and was built in partnership with organizations such as the Novo Nordisk Foundation, Denmark's Export and Investment Fund, and Nvidia. The Gefion supercomputer is likely to be used for various AI and computational applications, showcasing Nvidia's capabilities in the field of high-performance computing.

In conclusion, the design flaw and production delay in the Blackwell AI chips have been resolved, and the chips are now expected to ship in the fourth quarter. TSMC's assistance has been crucial in addressing the production issues, and the Blackwell chips are poised to deliver significant performance enhancements. Additionally, the launch of the Gefion supercomputer highlights Nvidia's continued innovation and leadership in the field of AI and high-performance computing.