At the end of the US stock market on Thursday, the general AI leader$NVIDIA (NVDA.US)$There was a wave of fast dives, and the final decline was close to 4%. At a time when tech giants are working together to devise new ways to challenge Nvidia's dominance, negative news broke out at the policy level in the Middle East.

On the other side, the aftermath of Nvidia's announcement of the 10-Q watch yesterday is still rippling; and as Hwang In-hoon enters the “100 Billionaire Club,” his work style has also begun to spread as a new “philosophy of success.”

Challengers League assembled!

Although various tech giants have launched their own AI chips in the past few years, they have never been as “arrogant” as Thursday.

In an announcement issued on Thursday,$Microsoft (MSFT.US)$,$Alphabet-A (GOOGL.US)$,$Meta Platforms (META.US)$,$Intel (INTC.US)$,$Advanced Micro Devices (AMD.US)$,$Cisco (CSCO.US)$,$Broadcom (AVGO.US)$,$Hewlett Packard Enterprise (HPE.US)$It was announced that a consensus has been reached to collectively promote a large-scale AI system communication standard for data centers called Ultra Accelerator Link (UALink).

(Source: BusinessWire)

The purpose of this specification is to set specifications for the connection of different AI accelerators. Of course, it is to facilitate the mixing of different AI chips within the same data center. The announcement stated that the relevant specifications will be provided to companies participating in the alliance in the third quarter of this year.

Obviously, everyone's first reaction to this announcement was — Nvidia wasn't among them!

Considering that Nvidia accounts for nearly 80% of the total AI chip market, there is no doubt that it will insist on using NVLink, which claims to be the “world's fastest high-performance computing protocol,” so it makes sense for these tech giants to “play without Nvidia.”

Of course, it's not easy to say what climate this alliance will become. After all, the “big brothers” Microsoft, Google, and Meta are all big customers waiting in line for Nvidia to provide graphics cards. At the same time, since all “investors” are currently Nvidia's major customers, it is unlikely that an agreement will be reached to exclude Nvidia's video cards.

Nvidia released the 10-Q schedule late Wednesday, which is a more complete quarterly report. The company disclosed that during the quarterly reporting period, “two indirect customers each accounted for 10% or more of the company's total revenue.” UBS analyst Timothy Arcuri said bluntly that Microsoft is one of these two major customers. Timothy said that according to previous estimates, the Microsoft family probably contributed close to 19% of Nvidia's revenue in the last fiscal year.

The Middle East business is rumored to be unfavorable

There were also media reports on Thursday that the rate at which the US Department of Commerce issued export licenses for computing power chips from Middle Eastern countries to tech giants such as Nvidia and AMD has “slowed down” in recent days. A person familiar with the matter said that some applications to sell chips to customers in the UAE, Saudi Arabia, and Qatar have been delayed or simply not responded after being submitted in the past few weeks.

According to reports, earlier this month, the US Department of Commerce Assistant Secretary for Export Management Thea Kendler visited Saudi Arabia, the United Arab Emirates, Qatar, Kuwait and other countries to discuss semiconductor export licenses. In October of last year, the US Department of Commerce included most of the Middle East in the scope of chip export restrictions. Companies such as Nvidia, AMD, and Intel need special permits to export cutting-edge chips to the region.

Nvidia directors plan to sell 435,000 shares

Also disclosed along with Form 10-Q is a Form 144, which is a document that the directors and supervisors of a listed company, or shareholders holding a large number of shares, need to be disclosed in advance before reducing their holdings.

Mark Stevens, director of Nvidia and a well-known Silicon Valley investor, revealed his intention to reduce holdings this time. He has been a director of Nvidia since 2008. The announcement revealed that he plans to sell 435,000 Nvidia shares, with a corresponding market value of close to 500 million US dollars.

(Source: SEC)

What makes investors even more envious is that he has held these stocks since he bought them in 2006. Based on current circulation, the cost for Stevens to hold these Nvidia shares should be around $3 per share, so this is another “300x wealth myth.”

Wong's Email Guidelines

With the latest rise in Nvidia's stock price, Hwang In-hoon's worth also broke through the $100 billion mark for the first time this week, making him the 15th richest person in the world.

Like many new super-rich people, beautiful conversations surrounding Hwang In-hoon's way of working have also appeared in various media.

According to the “Business Insider” website, citing former Nvidia employees, Hwang In-hoon has a practice of writing emails within the company called “TL; DR” (too long, don't read), and he pays great attention to the creed of simplicity and efficiency.

According to a former Nvidia executive, sending an overly long email to Hwang In-hoon would cause trouble. At Nvidia, emails are double spaced. Each line is one sentence, and an email may add up to no more than five or six lines. He further added, “The logic is to make it clear what must be said. If Hwang In-hoon or anyone else needs more information, he will come to talk to you rather than send another email.”

Of course, this kind of corporate culture depends entirely on the founder's temper. For example, Bezos, the richest man in the world today, is famous for liking beautiful and detailed internal letters. He has publicly stated that PPT is not allowed to be used during meetings within Amazon, and that employees need to prepare a “six-page narrative structure” memorandum before the meeting and read it silently together at the beginning of the meeting.

edit/new