$NVIDIA (NVDA.US)$It is known for its excellent performance in the field of AI chip manufacturing. However, the company's core design is to build a business barrier, which is composed of a close combination of software and hardware, and effectively prevents customer and competitive intrusion.

Over the past two decades, NVIDIA has carefully built a "walled garden" in the tech industry, similar to the ecosystem created by Apple (AAPL.US). Apple's ecosystem is mainly aimed at consumers, while NVIDIA focuses on serving developers who use its chips to build artificial intelligence systems and other software.

This closed system design explains why Nvidia can maintain its dominant position in the AI market among many competitors, including other chip manufacturers and technology giants such as Microsoft, Amazon and Intel, and it is unlikely to lose its market share in the next few years.$Alphabet-A (GOOGL.US)$And.$Amazon (AMZN.US)$Bill, vice president of Intel's cloud customer AI, said that the industry's cooperation is mainly focused on developing open source alternatives to CUDA. Intel engineers are participating in two such projects, one of which even involves collaboration with ARM, Google, Samsung, and other companies. The company behind ChatGPT, OpenAI, is also developing its own open source project.

In the long run, the competition for NVIDIA's market dominance will focus more on the company's coding ability, not just circuit design. Competitors are vying to develop software that can bypass NVIDIA's software barriers.

CUDA: The cornerstone of the "walled garden"

Understanding NVIDIA's "walled garden" is key to its software platform called CUDA. Since its launch in 2007, CUDA has solved a problem that other companies have not been able to solve: how to run non-graphic software, such as encryption algorithms and cryptocurrency mining, on dedicated chips designed by NVIDIA for labor-intensive applications such as 3D graphics and video games.

CUDA not only supports diversified computing tasks on these graphics processing units (GPUs), but also enables AI software to run on NVIDIA chips. In recent years, the explosive growth of AI software has pushed NVIDIA to become one of the most valuable companies in the world.

More importantly, CUDA is still evolving. Year after year, NVIDIA meets the needs of software developers by releasing specialized code libraries that enable a wide range of tasks to be executed at much faster speeds on NVIDIA GPUs than on traditional general-purpose processors such as Intel (INTC.US) and AMD (AMD.US) products.

The importance of full-stack computing and software platforms

The importance of NVIDIA's software platform also explains why the company has invested more resources in recruiting software engineers than hardware engineers for many years. CEO Jensen Huang recently emphasized that NVIDIA focuses on full-stack computing that combines hardware and software, manufacturing from chips to AI software.

Whenever a competitor announces the launch of an AI chip that competes with NVIDIA, it is actually competing with a system that has been in use for more than 15 years and has written a lot of code. It is difficult to port this software to a competitor's system, which is the real advantage of NVIDIA's coding capability.

At the shareholders' meeting in June, NVIDIA announced that CUDA now includes more than 300 code libraries and 600 AI models, and supports more than 40,000 companies, 5 million developers and 3,700 GPU-accelerated applications.

Market forecasts and competitive situation

The huge scale of the AI computing market has prompted many companies to join forces to compete with NVIDIA. Atif Malik, a semiconductor and networking equipment analyst at Citigroup Research, predicts that the AI-related chip market will reach $400 billion annually by 2027. At the same time, NVIDIA's revenue for the fiscal year ended in January was about $61 billion.

Not only start-ups, but also technology giants are actively deploying. Google and Amazon are developing their own AI training and deployment chips, while Starhub also announced in 2023 that it would join this group. These moves are not only a challenge to Nvidia's market position, but also a driving force for industry innovation.$Qualcomm (QCOM.US)$Nvidia's two major customers, Microsoft and Amazon, have also started to purchase AMD's AI chips. This behavior reflects the market's demand for diversified suppliers and the desire for high-end product competition.

Investors are flocking to start-ups that are dedicated to developing CUDA alternatives. Part of the reason for these investments is that engineers at many tech giants may join forces to push companies to use any chips they want, rather than pay what some in the industry call the "CUDA tax."

Market forecasts and competitive situation

In the field of AI chips, Nvidia maintains a significant lead but faces increasing competition. Groq, a start-up with a valuation of $2.8 billion, has successfully raised $640 million and is committed to developing chips that are comparable to Nvidia's. This marks the rise of open-source software, bringing new vitality and possibilities to the industry.

This transformation is not easy. Pahlavan admits that there have been many challenges and mistakes in the process. Gadi Hutt, director of Amazon Web Services, also acknowledged that both sides made mistakes in early cooperation, but now they are on the right track. Amazon's AI chip customer base is expanding, including companies like Anthropic, and others. Although Amazon provides customers with the option to use Nvidia chips, the cost is high and customer conversion takes time.$Microsoft (MSFT.US)$N/A

In this competition, AMD's Instinct AI chip series has become one of Nvidia's strongest competitors. AMD's Executive VP Andrew Dieckman said that although AMD still lags behind Nvidia in market share, the company is expanding its software resources by heavily investing in software engineers to close the gap. AMD announced last month that it would acquire Silo AI for $665 million to further enhance its AI R&D capabilities.

N/A$Meta Platforms (META.US)$N/A

However, Nvidia's market barriers are not unbreakable. Babak Pahlavan, CEO of start-up NinjaTech AI, revealed that if the cost allowed, he would prefer to use Nvidia's hardware and software. But faced with the shortage and high cost of Nvidia's H100 chip, NinjaTech AI had to turn to Amazon, which offered its own AI training chip, Trainium. After months of effort and collaboration, NinjaTech AI successfully trained their AI models on the Trainium chip and launched their AI "agent" in May, which now has over 1 million monthly active users, all supported by models trained and operated on Amazon chips.

N/A$Airbnb (ABNB.US)$,$Pinterest (PINS.US)$And.$Snap Inc (SNAP.US)$N/A

NinjaTech AI's experience illustrates one of the major reasons why start-ups like it face development of AI outside Nvidia's 'walled garden': cost. Pahlavan said that to support more than a million users per month, NinjaTech's cloud service costs about $250,000 per month on Amazon. By comparison, running the same AI on Nvidia chips would cost between $750,000 and $1.2 million.

Nvidia's response and future

Facing these competitive pressures, Nvidia is well aware of the high purchase and operating costs of its chips. CEO Jensen Huang has promised that Nvidia's next-generation AI chips will be dedicated to reducing the cost of training AI on the company's hardware.

Malik of Citi Research expects Nvidia's market share in the AI-related chip market to remain around 90% over the next two to three years. This indicates that despite facing competition, Nvidia's leading position remains secure.

In the foreseeable future, Nvidia's fate will depend on the inertia that has historically trapped many companies and customers in its "walled garden" effect.

Editor/ping