After yesterday's keynote speech at Computex Taipei, Jen-Hsun Huang returned today for a Q&A session with around 150 journalists and analysts from home and abroad.

The company will seek to diversify its supply chain, which it has already done, he said. The H100 is manufactured by Taiwan Semiconductor Manufacturing Co., LTD., and some of its products are also manufactured by Samsung. The company said it is open to cooperating with Intel on AI chips in the future.

Huang revealed that the company has recently received test chips based on manufacturing of Intel's next-generation process nodes, and the test results are good.

"You know, we're also working with Samsung, and we'd love to work with Intel. Pat Gelsinger previously said they were evaluating the process, and we recently received the results of their next generation test chip process, and the results look good, "Huang said.

Huang hinted nearly a year ago that Nvidia was in talks with Intel's Foundry Services (IFS) to bring in Intel nodes to make some of its chips. Thanks to Intel's IDM 2.0 model, other manufacturers can also take advantage of its latest process nodes to produce chips.

IT's important to note that the test chip mentioned above is a reference chip designed by Intel based on its own development platform to demonstrate to customers that the data on the test node is healthy, so it does not mean that Nvidia has placed an order. Huang also didn't specify which process node Intel's test chip represented or the architecture of the test chip.

However, the industry believes the chip could indeed be based on Nvidia's design. In January, Intel suddenly announced that it had signed an order for chips based on its Intel 3-node. The company initially said its first IFS chips would be based on its 20A, 18A and "Intel 16" nodes.

"I am very pleased that we are able to add a leading provider of cloud, edge and data center solutions as a leading customer for Intel 3," said Pat Gelsinger, CEO of Intel. Intel did not disclose the name of the customer, but given that the Intel 3-node is scheduled to begin production in the second half of 2023, a good partnership between the two companies cannot be ruled out, and given that all of Nvidia's cutting-edge Gpus are manufactured in China, this will help ease supply chain concerns about the situation in the US and China.