Nvidia, with its powerful AI GPU+CUDA ecosystem, has established a dominant position in the AI chip field and its "NVIDIA AI Empire" is becoming stronger and stronger.
$NVIDIA (NVDA.US)$CEO Huang Renxun stated at a technology conference organized by Goldman Sachs on Wednesday that Nvidia's next-generation AI GPU, the Blackwell architecture AI GPU, is so popular that it has caused dissatisfaction among some large customers who were unable to obtain the product in a timely manner. This "Versailles-style" statement by Huang shows Nvidia's near-monopoly position in the AI infrastructure construction field and the market's continued fanatical demand for Nvidia's AI GPU. In addition, statistics show that since 2023, Nvidia has been aggressively investing in AI startups worldwide to consolidate its absolute dominance in the generative AI infrastructure field.
Without a doubt, Nvidia is one of the biggest winners in the global artificial intelligence investment frenzy to date. Due to the hot demand for its high-performance AI GPU from global enterprises and government agencies, Nvidia's stock price has skyrocketed over 700% since early 2023, soaring over 900% since its historic low in October 2022. As Nvidia's market cap and revenue scale surged, the company's management continuously solidified the 'Nvidia software-hardware ecosystem' penetration into the artificial intelligence industry, accelerating investments in AI startups. Over half of the company's investments in startups since 2005 have occurred in the past two years.
Statistics show that the once holder of the title of 'world's highest market cap listed company,' the AI chip leader, has invested over $1.5 billion in AI startups in early 2024, a significant increase from $0.3 billion a year ago. According to Crunchbase, in 2024 alone, this AI chip giant has participated in over ten financings of $0.1 billion or more for AI startups. Since 2023, Nvidia has invested in over 50 AI startups, including important companies in the AI field such as Perplexity AI and Hugging Face.
Furthermore, Nvidia is considering investing in the upcoming ChatGPT developer OpenAI's latest financing round. All these latest dynamics from Nvidia indicate that Nvidia, the undisputed leader in the field of AI chips, is becoming increasingly powerful with the 'NVIDIA AI Empire' established by Nvidia with its powerful performance AI GPU + CUDA ecosystem.
AI startups are crucial to the global AI industry development, especially in the booming enterprise AI application market. Unlike global cloud computing giants like Amazon AWS, Microsoft Azure, and Google Cloud Platform focusing on building AI application development ecosystems or AI underlying infrastructure, these AI startups focus on various niche AI application scenarios, essential for improving enterprise operational efficiency or enhancing global consumer work or learning efficiency.
For example, Perplexity AI from the USA focuses on the cutting-edge field of 'AI search'; the French AI startup Bioptimus focuses on comprehensive integration of leading AI technology with medical science and biotechnology; and AI startup Cognition has introduced what is considered the world's first 'completely autonomous virtual AI software engineer.' This virtual engineer possesses strong programming and software development capabilities, assisting programmers in multiple cutting-edge technologies or independently completing large software development projects.
Here are the startup companies in the AI field that NVIDIA has invested in
Perplexity AI
Huang Renxun does not hide his love for Perplexity AI. Perplexity AI, known as the "Google Killer", unexpectedly became Huang Renxun's favorite AI tool. Huang Renxun was asked in an interview this year, "Do you use ChatGPT or Google AI chatbot frequently? Or do you use other products?" Huang Renxun replied, "I usually use Perplexity and use it almost every day. For example, when I recently wanted to understand how AI assists in drug development, I used Perplexity for related searches."
He even showed his support for Perplexity through practical actions. NVIDIA participated in a financing round of Perplexity AI in April, raising approximately $62.7 million, with a valuation of approximately $1 billion for the AI startup. Top investors led by Daniel Gross, and Jeff Bezos, the founder of Amazon, are among them. This is not the first time NVIDIA has supported the company - the chip giant also invested in Perplexity AI in a financing round in January, when the AI startup raised a staggering $73.6 million.
Hugging Face
Hugging Face is an AI startup that provides open-source AI models or application development platforms. It has had a close relationship with NVIDIA for a long time. In August 2023, the chip giant participated in a financing round of Hugging Face, raising up to $0.235 billion, and the post-financing valuation of Hugging Face is approximately $4.5 billion. Other corporate investors who participated in Hugging Face's financing at that time include Google, Amazon, Intel, AMD, and Salesforce.
Hugging Face has long incorporated NVIDIA's hardware system, CUDA software tools, and library resources into its shared resources. In May, the startup launched a new project, donating NVIDIA GPUs worth up to $10 million for free use by AI developers.
Adept AI
Unlike the generative artificial intelligence chatbots known to people like OpenAI and Anthropic, Adept AI's main product is not centered around text or image generation. Instead, this AI startup focuses on creating a software engineering assistant that can perform tasks on computers, such as generating reports or browsing the web, and can also utilize software tools. NVIDIA is also involved and participated in a funding round in March 2023, raising up to $350 million.
Databricks
Last autumn, Databricks gained a whopping $43 billion valuation, making it one of the most valuable AI startups in the world. As expected, this data analytics software provider has extensively adopted Nvidia's AI GPU and has received support from the chip giant, as well as other venture capital firms such as Anderson Horowitz and Capital One Ventures. All of these investors participated in a $0.5 billion financing round in September 2023. "Databricks is leveraging Nvidia's software and hardware technology to do incredible work in accelerating data processing and large-scale AI model generation," Huang Renxun said in a statement at the time.
Cohere
Canadian renowned AI startup Cohere is a strong competitor to OpenAI and Anthropic, specializing in providing enterprise-specific AI models. The company's growth over the past five years has attracted support from major tech supporters like NVIDIA, Salesforce, and Cisco, all of whom provided funding support for Cohere in a funding round in July. NVIDIA also participated in a funding round in May 2023, bringing approximately $270 million in funding to this AI startup.
"The NVIDIA AI Empire" continues to grow stronger
When NVIDIA invests in AI startups that focus on different application areas, these companies essentially allocate most of their investment funds to purchasing NVIDIA AI GPUs to establish or expand their AI training and inference infrastructure. AI startups need a significant amount of computational power to train their deep learning models, and NVIDIA's GPUs (such as H100, H200, and the upcoming Blackwell GPU) are industry standards in terms of performance, making them a natural choice.
NVIDIA CUDA is a highly optimized parallel computing platform and programming model that is deeply integrated with NVIDIA's GPU hardware. AI startups that receive investments from NVIDIA essentially invest heavily in the advanced version of the CUDA Accelerated Toolkit, further deepening their reliance on the NVIDIA ecosystem. This "lock-in effect" ensures that these startups continue to use NVIDIA's hardware and software tools almost inevitably when developing AI applications or iterating on large models.
In the future, when businesses use AI models or applications developed by these AI startups, they have to continue to rely on NVIDIA's full-stack ecosystem in the inference and deployment stages because they are built and iteratively optimized on NVIDIA's ecosystem. This allows NVIDIA to further expand its market share through these startups.
Furthermore, if the companies that actually use these AI models or AI applications from AI startups choose to deploy training/inference power on cloud computing platforms, this in turn requires AWS, Microsoft, and Oracle, and other cloud giants to continue to purchase NVIDIA's continuously iterating AI GPUs to build AI infrastructure. Many companies prefer to use cloud computing platforms such as AWS, Microsoft Azure, and Oracle OCI when deploying AI applications. If these cloud service providers' customers are using AI models and applications developed based on NVIDIA's software and hardware ecosystem, these cloud service providers will also need to continuously purchase large quantities of NVIDIA's latest AI GPUs and configure CUDA advanced acceleration tools and libraries to meet the huge computational power demand. The hardware and software full-stack ecosystem constituted by these factors collectively strengthens NVIDIA's "NVIDIA AI Empire".
Among them, the CUDA ecosystem barrier can be said to be NVIDIA's "strongest moat". NVIDIA has been deeply involved in the global high-performance computing field for many years, especially the CUDA computing platform it has built, which is popular worldwide and can be said to be the preferred software and hardware collaborative system for high-performance computing in areas such as AI training/inference. CUDA acceleration computing ecosystem is an exclusive parallel computing acceleration platform and programming assistance software developed by NVIDIA, which allows software developers and engineers to use NVIDIA GPUs for GPU-accelerated parallel general computing (only supports NVIDIA GPUs and is not compatible with mainstream GPUs such as AMD and Intel).
CUDA can be said to be the platform that generative AI applications such as ChatGPT are extremely dependent on, and its importance is no less important than the hardware system. It is critical for the development and deployment of large-scale AI models. With extremely high technical maturity, absolute performance optimization advantages and extensive ecosystem support, CUDA has become the most commonly used and fully popular collaborative platform in AI research and commercial deployment.
According to NVIDIA's official website, using NVIDIA GPUs for CUDA general accelerated computing programming and some basic tools is a free approach, but if it involves large-scale enterprise-level CUDA applications and support (such as NVIDIA AI Enterprise), or renting NVIDIA computing power on cloud platforms (such as Amazon AWS, Google Cloud, and Microsoft Azure), there may be additional fees for subscribing to advanced CUDA microservices for developing AI systems. In addition to the huge GPU hardware revenue brought by the tight integration of CUDA with AI GPUs, and the revenue brought by the large-scale enterprise-level CUDA applications, the software business derived from CUDA is also an engine for NVIDIA to achieve substantial revenue.
Based on the extremely powerful and high-penetration CUDA platform and the powerful AI GPUs, NVIDIA has been continuously escalating its layout in the software and hardware full-stack ecosystem recently. In March, NVIDIA officially launched a microservice called "NVIDIA NIM" at GTC, which charges based on GPU usage time. It is a cloud-native microservice that focuses on optimization, aiming to shorten the time to market for AI-based generative AI applications and simplify their deployment workloads on the cloud, data centers, and GPU-accelerated workstations, allowing companies to deploy AI applications based on NVIDIA's AI GPU cloud inference computing power and the acceleration foundation provided by the CUDA platform, seeking to establish a dedicated NVIDIA GPU system for the development of the AI application full-stack ecosystem.
Wall Street is calling the sell-off of NVIDIA stocks "excessive" and it is best to "buy on the dip" now.
This is also why Rosenblatt, a well-known Wall Street investment firm, is more optimistic about the core logic of the revenue growth of software based on CUDA compared to the revenue generated by NVIDIA's AI GPUs. Hans Mosesmann, a chip industry analyst at Rosenblatt, raised the firm's target stock price for NVIDIA from $140 to an amazing $200 per share in a research report, ranking it as the highest target price for NVIDIA on Wall Street.
Mosesmann stated that, based on nvidia's software business with CUDA as the core, the potential prosperity is expected, even though the stock price of nvidia, the dominant player in AI chips, has soared in the past year. As a result, besides the huge GPU revenue brought by nvidia's AI GPU firmly bound to CUDA, and the revenue brought by large-scale enterprise applications of CUDA, the software business derived from CUDA is also the engine for nvidia to achieve huge revenue.
Regarding the recent sharp drop in nvidia's stock price and the total market value evaporating by about $400 billion in the past week, top Wall Street investment giants such as Goldman Sachs have stated that cautious investors have 'over-sold' nvidia. Goldman Sachs analyst Toshiya Hari recently maintained a 'buy' rating for nvidia, stating: 'The recent stock performance of nvidia is not very good, but we still have confidence in this stock, and the recent sell-off is clearly excessive. First of all, the global demand for accelerated computing remains very strong. We are inclined to pay more attention to large-scale enterprises, such as Amazon, Google, Microsoft, and other global giants, but you will see that the range of demand is expanding to enterprises and even sovereign nations.'
With the further intensification of competition in the AI field by major technology companies such as Microsoft and Amazon, international bank UBS recently predicted that the overall AI capital expenditure of these technology giants in 2021 and 2025 may increase by 47% and 16.5% respectively, reaching $218 billion and $254 billion. However, UBS stated that the overall capital expenditure intensity (capital expenditure divided by sales) of large technology companies is still below historical peaks. UBS expected that as generative AI accelerates monetization, these technology giants seem to be on track to achieve close to 15-20% profit growth in the next few quarters. UBS predicted that the total free cash flow of large technology companies may increase from $413 billion in 2021 to $522 billion in 2025.
On Wall Street, the sentiment of 'buying on dips' is exceptionally strong. The bulls of U.S. stocks on Wall Street firmly believe that this round of correction has squeezed out the majority of the 'AI bubble', and those technology companies that can continuously profit from the AI wave are expected to enter a new round of 'primary uptrend' and soar, such as Nvidia, AMD, Taiwan Semiconductor, Advantest, and other popular chip stocks. Chips are an indispensable core infrastructure for popular generative AI tools such as ChatGPT, and these popular chip stocks can be called the biggest winners of the AI boom, especially the combination of 'CUDA ecosystem + high-performance AI GPU', which makes Nvidia's moat exceptionally strong.
In addition to Goldman Sachs, analysts from major banks such as Bank of America and Morgan Stanley are also optimistic about the trend of nvidia's stock price, and are exclaiming that it is a good opportunity to 'buy on dips'. Among them, Bank of America analyst Vivek Arya recently reiterated his 'buy' rating for nvidia, calling it the 'best industry choice', stating that the drop in nvidia's stock price has provided a good entry point, and raised nvidia's target stock price from $150 to $165, compared to Nvidia's close at $116.910 on Wednesday. The analyst at Bank of America emphasized that the market's skepticism about the potential of AI is unnecessary at least until 2026.
The current demand for AI chips can be said to be extremely strong and is likely to remain so for a long time. The management of Taiwan Semiconductor recently stated at an earnings conference that the advanced packaging CoWoS required for AI chips is expected to remain in a situation of supply shortage until 2025, and there is a possibility of a slight relief in 2026. Recently, industry insiders revealed that due to the extremely strong demand for Nvidia's upcoming mass-produced Blackwell architecture AI GPU globally, Nvidia has significantly increased its AI GPU foundry orders with chip giant Taiwan Semiconductor by at least 25%.
Huang Renxun stated at the meeting on Wednesday: 'The demand for AI GPUs can be said to be very strong, everyone wants to be the first to receive goods, and everyone wants to receive the most products.' 'We may have more emotional customers now, and this is natural. The situation is a bit tense now, and we are doing our best.' He added that the latest generation of AI GPU, the Blackwell GPU, has strong demand, and suppliers are catching up.
In addition, when asked whether large-scale AI expenditure provided clients with investment returns—this has always been a worrying issue in this wave of AI frenzy—Huang Renxun stated that, other than embracing 'accelerated computing', enterprises have no choice. He pointed out that Nvidia's technology accelerates traditional workload processing and can handle AI tasks that existing technologies cannot cope with.
Editor/ping