Financial Associated Press, October 9th (Editor Shi Zhengcheng) After the opening on Tuesday local time, the global computing power chip leader Nvidia's stock price surged more than 3%, with the latest quote at $132.24 at the time of publication, just a step away from the historical closing high of $135.58 in June.

Strictly speaking, Nvidia itself did not have particularly explosive good news these days. However, what is very subtle is that in the past few days, when various technology giants were showcasing business progress, they all could not avoid Nvidia's cutting-edge chips.

Prominent technology manufacturing company Hon Hai/Foxconn held its annual Technology Day on Tuesday (October 8), during which it revealed that the company is building a large artificial intelligence server manufacturing facility in Mexico to meet the 'crazy' demand worldwide for Nvidia's latest GB200 system.

Hon Hai's Chairman Liu Yangwei stated at the conference that, as Huang Renxun said, the demand for Blackwell chips is extremely high, and the production capacity of the company's Mexican factory will be enormous. The company is also using Nvidia's Omniverse software to build a highly automated Mexican factory, using AI to create a factory producing AI chips.

Meanwhile, technology giant microsoft's Azure cloud computing service also boasted on Tuesday that the company has obtained servers equipped with the GB200 chip.

Azure emphasized that it is the first company among global cloud service providers to adopt the Blackwell system, while the new servers from the technology giant also feature Infiniband interconnect technology and innovative closed-loop liquid cooling technology.

Interestingly, while microsoft was boasting about its servers on Tuesday, The Information cited sources reporting that due to dissatisfaction with microsoft's inability to provide sufficient computing power, the over $150 billion valued OpenAI is planning to break free from its reliance on microsoft, independently lease datacenters, and develop AI chips.

Reports have stated that OpenAI is in discussions with Oracle to lease the entire Abilene datacenter. The datacenter was expected to achieve a power supply scale of 1 gigawatt by 2026 - to support a cluster of hundreds of thousands of nvidia chips.

In addition to finished chips, nvidia's presence is also penetrating the semiconductor manufacturing sector. The company's official website announced on Tuesday that the global chip manufacturing leader Taiwan Semiconductor is using the nvidia cuLitho computing lithography platform to accelerate production and drive the physical limits of next-generation advanced semiconductor chips.

For semiconductor fabs, computational lithography is a critical step in chip production - this involves transferring circuits onto silicon wafers. NVIDIA explains that a typical chip mask set may require the use of 30 million or more CPU compute hours, by switching to NVIDIA's GPU hardware, 350 H100s can replace 0.04 million GPU systems, resulting in faster production speeds and lower costs, space, and power consumption. With the support of AIGC, AI workflows can provide an additional doubling of speed on top of cuLitho acceleration.

By the way, there is also an apology for those who couldn't keep up with NVIDIA's pace. South Korea's memory giant Samsung issued a rare public apology on Tuesday after issuing a profit warning. The company disclosed that there was a delay in delivering HBM3E chips to a major customer. Currently in the high-bandwidth memory supplier field for AI services, SK Hynix and Micron are NVIDIA's main suppliers.