[News] A $500 Billion Investment Plan to Launch, Benefiting AI Chips

---

---

Under the influence of large AI models, AI infrastructure projects are booming worldwide. OpenAI has announced the launch of a $500 billion “Stargate” plan, while ByteDance is reportedly allocating significant funds to build and enhance its AI infrastructure.

Although there are differing opinions on such massive investments, the AI wave remains unstoppable, and AI chips are expected to benefit from it.

OpenAI Launches the “Stargate” Plan

Recently, OpenAI announced the launch of a large-scale project called “Stargate.” OpenAI will collaborate with companies such as SoftBank and Oracle to establish multiple AI data centers in the U.S. over the next four years.

OpenAI stated that it has already deployed $100 billion in the first round of investment for the project and plans to increase the total investment to $500 billion over the next four years.

To achieve this, OpenAI, SoftBank, Oracle, and the sovereign wealth fund MGX have jointly formed a venture called Stargate. Among them, SoftBank and OpenAI will be the main partners, with SoftBank taking on financial responsibilities. SoftBank CEO Masayoshi Son will serve as the chairman of the new venture, “The Stargate Project,” while OpenAI will handle operations.

In addition, major tech companies such as Oracle, Arm, Microsoft, and NVIDIA will join the project as initial technology partners.

OpenAI is actively pushing forward the “Stargate” plan, with construction set to begin in Texas before expanding to other U.S. states.

ByteDance Responds to a $12 Billion Investment in AI Infrastructure

Recently, foreign media reported ByteDance’s significant investment plans in AI, with the company potentially allocating over $12 billion for AI infrastructure. A portion of this investment would target overseas markets, while the rest would focus on the domestic market.

In response, ByteDance told domestic media that the rumored budget and plans were inaccurate. However, the company also emphasized its strong commitment to AI development and investment.

On the application side, ByteDance launched the “Doubao” AI chatbot in 2023, which has gained some attention in the Chinese market. As of December last year, “Doubao” had reached 71 million monthly active users, showing strong growth momentum.

Investment-Driven Growth, AI Chips Benefit

Industry experts point out that as major tech companies continue to increase capital expenditures in AI, capital spending is expected to remain optimistic through 2025, primarily directed toward servers, data centers, and network infrastructure. This trend is likely to boost the overall AI infrastructure supply chain.

AI chips play a crucial role in the AI industry by providing powerful computational support for AI applications. They also help optimize AI algorithms and models, drive technological innovation and application, promote industrial development, and enhance user experience.

From a technical perspective, AI chips are mainly categorized into GPUs, FPGAs, and ASICs. Functionally, they are divided into training and inference chips. Based on application scenarios, AI chips are classified into server-side and mobile-side categories or further divided into cloud, edge, and terminal deployments.

GPUs, as general-purpose chips, currently dominate the AI chip market. General-purpose GPUs are widely used for AI model training and inference tasks.

Currently, NVIDIA and its GPU products lead the global AI chip market. In recent years, domestic AI chip manufacturers have been rapidly emerging, achieving notable results in terminal applications and large-model inference. They are also making significant efforts in high-end GPU development and large-model training to narrow the gap.

(Photo credit: OpenAI)