The business structure is also extremely simple.
Nvidia designs semiconductors and supports them with software.
They do not manufacture semiconductors themselves, and their B2C products used by general consumers are limited to gamers.
We only serve four types of customers: data centers, gaming, autonomous driving, and graphics professionals.
Unlike Microsoft, which has massive data centers and research facilities, Nvidia doesn't have many tangible assets.
Just how valuable is their $3 trillion enterprise value? The combined market capitalization of all companies listed on Korea's KOSPI and KOSDAQ falls short of NVIDIA's alone.
To put it more simply, as of the end of February 2024, Samsung Electronics' market capitalization was $378 billion.
This means that eight companies, such as Samsung Electronics, which manufactures everything from home appliances to mobile phones and memory semiconductors, would have to come together to have the same value as Nvidia.

--- 「Part 1.
From "The Emergence of Nvidia, the Destroyer of the Semiconductor Ecosystem"

In a February 2024 article, The New York Times commented that “NVIDIA’s unique AI-related intellectual property, rooted in decades of pioneering investments, sets it apart in the semiconductor league.”
It is really difficult to break the economic moat of Nvidia GPUs thanks to their overwhelming performance and ecosystem dominance that makes them 'unavoidable to use even if they are expensive'.
It is not easy to change the semiconductors used unless there is a major change that will change the industrial landscape.
As it becomes harder to find alternatives to Nvidia GPUs, Nvidia will only strengthen its moat.
This is why we should continue to pay attention to NVIDIA now and in the future.
--- 「Part 1.
From "The Emergence of Nvidia, the Destroyer of the Semiconductor Ecosystem"

Nvidia saw potential for GPUs in these supercomputers, and in the early 2000s, it introduced what it called the GPGPU (General Purpose GPU). They believed that GPUs could play a complementary role in supercomputers that primarily relied on CPUs.
Starting with the 'Nvidia Tesla' launched in 2007 targeting the supercomputer market, Nvidia entered the supercomputer market in earnest.
And this challenge led to Nvidia, which dominates the AI ​​era today.
It could be said that the AI ​​era we are experiencing now was started by Nvidia in 2007.
The know-how and network accumulated through entering the supercomputer market in this way led to an absolute market share in the AI ​​semiconductor market over a decade later.

--- 「Part 2.
From “From a Startup to an AI Semiconductor Dinosaur”

What's even more surprising is the profitability.
The operating profit margin reached 65%, and the gross margin ratio was 78.4%, overwhelming other tech companies.
Sales increased by 262% and operating profit increased by 690% compared to the same period last year.
In one year, the company's sales increased 3.6 times and operating profit increased 7 times.
This means that a whopping 65% of sales remain as company profits, which is an unimaginably high operating profit rate for a manufacturing company.
In this way, Nvidia has announced earnings surprises that far exceed market expectations every quarter, and its stock price has also surged each time.
Nvidia, which was not even in the top 20 in terms of market capitalization among US companies until 2020, temporarily surpassed Microsoft in June 2024 to become the most valuable US company.
The status of a company called Nvidia in Silicon Valley has also completely changed.
Before jumping into data centers, Nvidia was perceived as a B2C-focused gaming company.
However, it is currently a representative AI company and is recognized as a company leading the future of computing.
--- 「Part 3.
What made them an 'irreplaceable' company?

Some say that Nvidia best combines the culture of Silicon Valley and East Asian companies.
Horizontal corporate culture, horizontal information exchange, honesty, and autonomy are typical Silicon Valley-style corporate cultures.
On the other hand, it is typical of East Asian companies to not fire people easily and to keep them employed for a long time.

The only time Nvidia underwent a Silicon Valley-style mass layoff was during the 2008 financial crisis (when it laid off 350 people, or 6.5% of its global workforce).
In Silicon Valley, where job changes and layoffs are frequent, Nvidia is considered one of the companies with the most stable employment.
Even employees prefer to stay at NVIDIA rather than leave.
This is true not only in Silicon Valley but also at Nvidia's Korean branch, where it is said that there are many employees who have worked there for over 10 years.
--- 「Part 4.
From "NVIDIA's corporate culture that never loses its original intention"

Can CSPs' proprietary AI semiconductors truly take over the market dominated by NVIDIA? The prevailing view in the industry is that it won't be easy.
Even though Nvidia's GPUs are expensive, their performance is far superior to that of other competitors, which is something that even competitors acknowledge.

Beyond performance issues, customers' preference for NVIDIA GPUs is also a barrier to entry for competitors. While CSPs use AI semiconductors for their own purposes, they also provide them to customers.
If customers want to use NVIDIA's GPUs rather than AI semiconductors designed by CSPs, they have to provide it.
For this reason, considering that CSPs must maintain a certain percentage of Nvidia GPU usage, the current outlook is that Nvidia's market share will remain at least 50%.

--- 「Part 5.
From "The ongoing war between chipmakers, what is the future of Nvidia?"