Altman's new project is aimed at increasing global chip manufacturing capacity, and he is in talks with investors, including the United Arab Emirates**. Ultraman needs to raise $5 trillion to $7 trillion, while Microsoft and Apple have a combined market capitalization of about $6 trillion.
There are uncertainties, including how to find engineers to operate the new plant, how to get equipment to fill the plant, and how to get enough orders to justify the plant.
Nvidia controls about 80% of the AI chip market, and Sam Altman, the "father of ChatGPT," is trying to reshape the global semiconductor industry.
OpenAI CEO Sam Altman wants to seek $5 trillion-$7 trillion in funding for AI chip projects to improve global chip manufacturing capacity. He hopes to build dozens of chip manufacturing plants in the coming years. But raising trillions of dollars, or the easier part of Ultraman's chip plan, also faces challenges such as talent acquisition, cyclical markets and a lack of viable chipmakers.
And Nvidia CEO Jensen Huang quipped that $7 trillion "can obviously buy all the GPUs." A few days ago, at the World Summit in Dubai, Huang said that due to the expected advances in computing technology, the cost of developing artificial intelligence will not be as much as Altman wants to raise. "You can't take it for granted that you're going to buy more computers. "You also have to assume that computers are going to get faster, so you won't need as much of the total." ”
Raising trillions of dollars to reshape the global semiconductor industry
In the AI boom of the past year, Nvidia's GPUs (graphics processing units) have been in the limelight, powering large models for OpenAI, Meta, and other well-funded startups. Meta CEO Mark Zuckerberg said last month that the company plans to have about 350,000 Nvidia's flagship H100 processors by the end of the year.
Currently, Nvidia controls about 80% of the AI chip market, and Altman is trying to change that.
Altman has long been concerned about the supply and demand of AI chips, he has said that the limitations of AI chips hinder the development of OpenAI, often complaining that there is not enough GPU to support OpenAI to explore general artificial intelligence, and also complaining about the cost of Nvidia chips.
Altman said on February 8 that the world needs more AI infrastructure, including fab capacity, energy, data centers, etc., more than people currently plan to build. "Building large-scale AI infrastructure and resilient ** chains is critical to economic competitiveness. OpenAI will try to help. ”
Altman's new project, which focuses on increasing global chip manufacturing capacity, is in talks with investors, including the United Arab Emirates**. According to Wall Street**, a person familiar with the matter said that Ultraman needs to raise $5 trillion to $7 trillion. Altman has also met with U.S. Commerce Secretary Gina Raimondo and other U.S. ** in recent weeks to discuss his plans.
Such an investment would dwarf the size of the current global semiconductor industry, which last year sold $527 billion in chips and $100 billion in semiconductor manufacturing equipment. By the standards of corporate financing, Altman is also talking about ridiculously large amounts, even larger than the national debt of some major economies and the large sovereign wealth**. As the two most valuable companies in the United States, Microsoft and Apple have a combined market capitalization of about $6 trillion.
It is planned to build dozens of chip factories in the coming years
Altman is interested in chip manufacturing. In 2018, Altman personally invested in Rain Neuromorphics, an artificial intelligence chip start-up, which is located near OpenAI's San Francisco headquarters. In 2019, OpenAI signed a letter of intent to spend $51 million to buy RAIN's chips.
Today, Ultraman has even greater ambitions. In an effort to reshape the global semiconductor industry, some people familiar with the matter said that as part of the talks, Altman is pushing for a collaboration between OpenAI, investors, chipmakers and power companies, which will jointly fund the construction of chip foundries, which will then be operated by existing chipmakers. OpenAI will become an important customer for the new factory.
Among the Middle Eastern investors that Altman has wooed is Sheikh Tanu bin Zayed Al Nahyan, one of Abu Dhabi's richest and most influential figures. He is an advisor to the UAE, Chairman of the Abu Dhabi Investment Authority, Abu Dhabi's third largest sovereign wealth **ADQ (Abu Dhabi Development Holdings***), and Chairman of Abu Dhabi's artificial intelligence company G42, which has also established partnerships with Microsoft and OpenAI.
Altman has revealed his plans to Microsoft CEO Satya Nadella and CTO Kevin Scott. He also met with SoftBank CEO Masayoshi Son, as well as representatives of chip manufacturing companies such as TSMC to discuss cooperation. According to Wall Street**, people familiar with the matter said that during negotiations with TSMC, Altman had said that he hoped to build dozens of chip manufacturing plants in the next few years. His vision is to raise money from Middle Eastern investors for TSMC to build and operate the plants.
However, Altman's negotiations are still in their early stages, and the full list of potential investors is not yet known, and the negotiations may last for years and may not be successful in the end.
Facing challenges such as talent, cyclical fluctuations, etc
Raising trillions of dollars may be the easier part of Altman's chip plan, but he also faces challenges such as talent acquisition, cyclical markets and a lack of viable chipmakers.
Chip manufacturing is a capital-intensive industry, and it usually costs at least $10 billion to build a cutting-edge chip factory. But money is not the only factor for success. The chip industry is one of the most intricate industries in the world, with sharp cyclical fluctuations, and companies are cautious about aggressive expansion. Some chip companies faltered during the industry downturn, while others halted cutting-edge chip development due to high costs and risks. It took decades for the world's most advanced chipmakers to reach their current heights. According to Wall Street**, at present, there are only three companies in the world that can produce the most cutting-edge chips on a large scale, namely TSMC, Samsung Electronics and Intel. Many large companies, including Nvidia, design their own chips but outsource chip production to companies like TSMC.
Industry executives also said there are uncertainties, including how to find engineers to operate a new batch of factories, how to get equipment to fill factories, and some chip-making equipment has a lead time of about two years. In addition, there is uncertainty about how to get enough orders to justify these factories. Even if a large number of new chip factories are built, it will not necessarily solve the shortage of AI chips that Ultraman has encountered recently. According to Wall Street, the biggest bottleneck in the production of Nvidia's AI chips is packaging.
The challenges of manufacturing chips are different from those faced by Altman in his early days of entrepreneurship involving computers and software. Rand Corp., a semiconductor industry expert and a U.S. think tankJimmy Goodrich, a senior consultant, said that in software, anything is possible, it's just a matter of funding and programming. But in the field of hard technology, where the laws of physics have to be dealt with, and real-world and engineering challenges have to be considered, it's hard to do.
At the same time, chip industry insiders said that if Ultraman's plan is successful, the market may be oversupplied, causing companies to operate factories below capacity. Due to high fixed costs, this will sound the financial death knell for the chip industry.
*: The Paper.
Pay attention to *** for more information, "China Science and Technology Information" magazine.
Supervisor: China Association for Science and Technology Organizer: China Science and Technology News Society Content**Please contact WeChat: zkxxx1999**Submission platform: wwwcnkjxx.com submission**: 010-68003059 for reports and content cooperation, please contact WeChat: 15811564659 Magazine inclusion: "CNKI", "China Journal Core Journals (Selection) Database", "China Academic Journals Comprehensive Evaluation Database (CAJCED) Statistical Source Journals", "China Periodicals Full-text Database (CJFD)", "China Association for Science and Technology, China Library Association (Interpretation of the Scientific Outlook on Development Recommended Bibliography)".