With the help of AI, Huang Renxun not only brought NVIDIA to the top of the world - becoming the world's number one in market value, but also made himself the eleventh richest person in the world, click here for more details, taking his wealth to a new level.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/cba0a9e4-3e8c-4a63-84bd-a0d8938c4bbf.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/cba0a9e4-3e8c-4a63-84bd-a0d8938c4bbf.png"/>
As the recognized biggest seller of AI shovels, NVIDIA's AI chips have already captured about 80% of the global data center market share.
However, beneath this shiny exterior and well-deserved victories, Huang Renxun remains focused on addressing the next threat - the potential slowdown in demand for NVIDIA GPUs.
According to tech media The Information, Huang Renxun convened a series of executive meetings around Christmas last year to discuss an increasingly serious issue: whether NVIDIA's major customers would be unable to install its AI chips due to insufficient data center space, thus affecting chip sales.
The report stated that to offset the impact of slowing chip demand on performance, NVIDIA has started selling more software to AI developers and even established its own server leasing business, DGX Cloud, a year ago. This move will directly compete with major customers such as Microsoft and AWS.
<h2>Huang Renxun's Concerns</h2>
Huang Renxun told colleagues that he is worried that cloud server providers like Amazon AWS and Microsoft have not quickly built new data centers and power supplies to accommodate the chips they have ordered. Over the past few quarters, these major cloud service providers have collectively purchased about half of NVIDIA's AI server chips.
Following the meeting, NVIDIA's management accelerated inquiries to cloud providers to confirm their ability to accommodate and use the ordered chips, according to an NVIDIA employee and several customers and data center operators.
Raul Martynek, CEO of data center provider DataBank, said:
<blockquote>
Unless customers can prove that their data centers can accommodate these GPUs, NVIDIA will not ship.
</blockquote>
At the 2023 GTC conference, NVIDIA unveiled its cloud product DGX Cloud for the first time, directly competing with cloud giants like Microsoft and AWS.
Interestingly, DGX Cloud runs on NVIDIA chip servers rented by cloud service providers, and then NVIDIA rents the servers to its customers at a higher price, promising them better computing performance.
NVIDIA has even considered leasing its own data centers for DGX Cloud, completely excluding third-party cloud providers.
According to sources, NVIDIA recently hired Meta executive Alexis Black Bjorlin to manage its cloud business
At present, it is unclear whether NVIDIA will continue to advance the plan to establish its own data center for DGX Cloud.
While taking these measures, NVIDIA chip salespeople are also trying to understand how customers are using the chips.
The Information reported that NVIDIA is collecting customer information from cloud providers and the contents of leasing contracts signed between them in advance to plan sales strategies and understand potential customers for their cloud server leasing business.
Huang Renxun also realized another factor that may affect chip sales:
Companies such as Microsoft, Meta, and xAI under Musk are using chips to train experimental new AI models. Since these models do not immediately generate revenue, investors may pressure these companies to reduce chip purchases.
At the end of April, after Zuckerberg admitted the uncertainty of AI revenue, Meta's stock price plummeted, and investors expressed clear dissatisfaction with the situation.
Finally, according to industry insiders, although AI computing currently dominates the business of major cloud service providers, there has been a recent resurgence in demand for traditional computing businesses. Coupled with the high cost of data center expansion, they are unable to fully accommodate the installation and use of NVIDIA chips.
<h2>Controlling GPU allocation, vigorously promoting in-house installation solutions, angering major customer Microsoft</h2>
To address these issues, Huang Renxun has been carefully managing the allocation of NVIDIA chips to prevent any company from hoarding too many chips.
He also tries to influence how customers assemble GPUs in data centers, urging them to adopt NVIDIA-approved server rack designs with higher computing performance.
However, some believe that if they follow NVIDIA's advice, they will find it more difficult to switch to other companies' chips in the future.
Sources say that NVIDIA and major customer Microsoft often clash over the installation method of NVIDIA's next-generation Blackwell chips.
NVIDIA also hopes to generate as much revenue as possible from cables, racks, and other hardware connecting chip servers, but this may harm manufacturers like Dell who have long produced NVIDIA chip servers.
<h2>&#34;All the Leverage&#34;</h2>
Although software and cloud service revenue still pales in comparison to NVIDIA's core server chip business, some customers and former employees believe that this business may ultimately shield NVIDIA from the decline in chips and make it harder for customers renting their servers to seek alternative chips.
Former NVIDIA executive and current partner at venture capital firm Playground Global, Sasha Ostojic, said that cloud and software products as a business that can generate billions of dollars in revenue annually have never been fully appreciated by analysts and the tech community.
He added that NVIDIA has &#34;all the leverage&#34; to develop services complementary to its chips.
In addition to the huge revenue potential, DGX Cloud has also become a way for NVIDIA to help some customers transition to the next generation of chips
Taking ServiceNow as an example, this software manufacturer has always purchased NVIDIA servers for its data centers in the past, and now also directly leases these servers from NVIDIA.
<h2>Avoiding Cisco's old path</h2>
Facing the inevitable decline in future chip demand, NVIDIA may face difficulties in the future if it does not have a stable new source of income.
Cisco is a cautionary tale.
During the peak of the Internet bubble in 2000, Cisco suddenly became the world's most valuable company by selling routers. At that time, telecommunications companies built new data centers, but due to the failure of Internet revenue to materialize as expected, these centers were left idle.
At the same time, Cisco's hardware was widely copied by competitors, leading to a continuous decline in sales, a sharp drop in stock price, and it has not recovered to this day.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/70fed6aa-9feb-48a5-a5b4-0fa855ce23c4.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/70fed6aa-9feb-48a5-a5b4-0fa855ce23c4.png"/>
Huang Renxun privately told colleagues that NVIDIA must ensure that it does not rise rapidly like Cisco or Sun Microsystems, only to quickly decline.
Sun Microsystems became a giant in servers and computer hardware in the 1990s, but after the bubble burst, it failed to seize the emerging software market, lost to Microsoft in the competition, and then declined rapidly

SOXL

NVDA

AMZN

NVDL

Huang Renxun is worried that the demand for NVIDIA GPUs may eventually slow down. To avoid following in the footsteps of Cisco, Huang Renxun is actively promoting NVIDIA's entry into the software and cloud services market, competing with major customers Amazon AWS and Microsoft

1. Core descriptive facts:
   - Huang Renxun led NVIDIA to become the world's most valuable company and himself to become the 11th richest person globally.
   - NVIDIA's AI chips hold around 80% market share in global data centers.
   - Concerns arise about a potential slowdown in demand for NVIDIA GPUs due to data center space constraints.
2. Points and supporting logic:
   - Despite NVIDIA's current success, Huang Renxun is focused on addressing the future threat of slowing demand for NVIDIA GPUs.
   - NVIDIA is exploring new strategies such as selling more software to AI developers and launching the DGX Cloud server rental service to counter the impact of slowing chip demand.
   - Huang Renxun worries that major cloud providers like AWS and Microsoft may not expand data centers quickly enough to accommodate ordered chips, prompting NVIDIA to verify data center capacity before shipping.
   - NVIDIA's move to compete directly with cloud giants like Microsoft and AWS through DGX Cloud indicates a shift towards cloud services.
   - NVIDIA's consideration of renting its own data centers for DGX Cloud and hiring a Meta executive for cloud business management demonstrate strategic diversification efforts.
   - NVIDIA's sales team is actively studying how customers utilize their chips, indicating a focus on understanding client needs and optimizing chip usage.
Overall, the article highlights NVIDIA's dominance in AI chips, challenges posed by potential GPU demand slowdown, strategic shifts towards cloud services, and efforts to ensure chip utilization and address data center capacity issues.

Having reached the top of the world, why is Huang Renxun still worried?