OpenAI CEO Sam Altman’s $7 Trillion Plan to Make AI Chips: Is It Feasible or Futile?

Artificial intelligence (AI) is one of the most transformative technologies of our time, with applications ranging from entertainment to health care to education.

However, AI also requires a lot of computing power, which depends on the availability and performance of specialized chips called graphics processing units (GPUs).

Currently, the global supply of GPUs is insufficient to meet the demand of the AI giants, such as OpenAI, Alphabet, Meta, and others, who are creating large and powerful language models that can generate text, images, code, and more.

This creates a bottleneck for the development and deployment of AI, as well as a vulnerability for the AI industry, which relies heavily on a few chip manufacturers, especially Nvidia.

To address this challenge, OpenAI’s founder and CEO Sam Altman has a bold vision: he wants to increase the global production of AI chips massively, and he is looking for trillions of dollars in funding to do so.

According to a report by The Wall Street Journal, Altman is in talks with various investors, including the government of the United Arab Emirates, to launch a project that would expand the chip-building capacity of the world.

He believes that this project is crucial for the economic competitiveness and the future of AI.

openai is the most talented and nicest group of people i have ever seen in one place

working on the hardest, most interesting, and most important problems

with all the key resources in place

extremely focused on making AGI

you should perhaps considering joining us

— Sam Altman (@sama) February 16, 2024

But is Altman’s plan realistic or unrealistic? How much money does he need, and where will he get it from? What are the benefits and risks of his proposal? And what are the alternatives and competitors in the AI chip market?

In this article, we will try to answer these questions and provide some insights into the ambitious and controversial project of Altman.

The $7 Trillion Question

One of the most striking aspects of Altman’s plan is the amount of money he is seeking to raise: between $5 trillion and $7 trillion, according to The Wall Street Journal, citing one source.

This is a staggering sum, equivalent to about 6% of the global GDP in 2023, or more than the combined market capitalization of Apple, Microsoft, Amazon, and Alphabet.

Why does Altman need so much money? The answer lies in the complexity and cost of building semiconductor chips, which are the basic components of GPUs and other electronic devices.

Semiconductor chips are made of thin layers of silicon and other materials, etched with intricate patterns of transistors and circuits that enable computation.

The process of making chips involves hundreds of steps, requiring sophisticated equipment, clean rooms, and skilled workers. It also takes several months to complete, from design to testing to packaging.

The cost of building a chip factory, or a fabrication plant (fab), is also very high, and it keeps increasing as the technology advances.

According to the Semiconductor Industry Association, the average cost of a new fab in 2023 was about $15 billion, up from $4 billion in 2013. The most advanced fabs, which can produce chips with the smallest and densest features, can cost up to $20 billion or more.

For example, Intel announced in March 2023 that it would invest $20 billion to build two new fabs in Arizona, as part of its plan to revive its chip-making business.

Altman’s plan, according to The Wall Street Journal, is to build dozens of new fabs around the world, in order to increase the global chip production capacity by 50% in the next decade.

He also wants to develop new chip designs and architectures that are optimized for AI applications, and to create a more resilient and diversified supply chain for AI infrastructure.

He believes that this project is essential for the growth and innovation of AI, as well as for the economic and national security of the countries that participate in it.

Intel CEO asks Sam Altman what he is going to do with $7 trillion pic.twitter.com/9hIgD8L8aZ

— Tsarathustra (@tsarnick) February 24, 2024

However, Altman’s plan faces many challenges and uncertainties, both technical and financial. On the technical side, building and operating fabs is not only expensive, but also risky and competitive.

It requires constant investment in research and development, as well as in upgrading and maintaining the equipment and facilities. It also requires compliance with environmental and safety regulations, as well as protection from cyberattacks and sabotage.

Moreover, it faces fierce competition from the existing and emerging players in the chip industry, who have their own advantages and strategies.

On the financial side, raising and managing such a large amount of money is not easy, either. Altman has reportedly approached several potential investors, including sovereign wealth funds, private equity firms, and venture capitalists, but he has not disclosed any details or commitments.

He has also not revealed how he plans to structure and govern his project, or how he will allocate the ownership and profits of the chips and fabs.

Furthermore, he has not addressed the ethical and social implications of his project, such as the potential impact on the environment, the labor market, the human rights, and the geopolitics of AI.

The Benefits and Risks of Altman’s Plan

Altman’s plan, if successful, could have significant benefits for the AI industry and the world. Some of the possible benefits are:

• More AI innovation and applications: By increasing the availability and performance of AI chips, Altman’s plan could enable more AI research and development, as well as more AI products and services, across various domains and sectors.
  
  This could lead to more scientific discoveries, technological breakthroughs, and social benefits, as well as new markets and opportunities for businesses and consumers.
  
• More AI competition and diversity: By creating more AI chip suppliers and customers, Altman’s plan could also increase the competition and diversity in the AI industry, which is currently dominated by a few large companies and countries.
  
  This could foster more creativity and collaboration, as well as more accountability and transparency, in the AI field.
  
  It could also reduce the dependence and vulnerability of the AI industry on a single or a few chip manufacturers, such as Nvidia, which could pose a threat to the stability and security of the AI ecosystem.
  
• More AI infrastructure and resilience: By building more AI chip fabs around the world, Altman’s plan could also improve the infrastructure and resilience of the AI industry, which is currently facing a global chip shortage and a geopolitical tension.
  
  The chip shortage, caused by the surge in demand and the disruption in supply due to the COVID-19 pandemic and other factors, has affected many industries and sectors, including AI, automotive, consumer electronics, and health care.
  
  The geopolitical tension, fueled by the trade war and the rivalry between the US and China, has also threatened the access and the reliability of the chip supply chain, as well as the cooperation and the trust in the AI field.
  
  Altman’s plan could help alleviate these problems, by increasing the chip production capacity and diversifying the chip sources and destinations.

However, Altman’s plan, if unsuccessful, could also have significant risks and costs for the AI industry and the world. Some of the possible risks are:

• More AI waste and inefficiency: By investing so much money and resources in building AI chips and fabs, Altman’s plan could also create more waste and inefficiency in the AI industry, if the demand and the performance of the chips do not match the supply and the cost.
  
  This could result in overcapacity and underutilization of the chips and fabs, as well as in lower returns and higher losses for the investors and the operators.
  
  It could also lead to more environmental and social problems, such as pollution, energy consumption, carbon emissions, and displacement of workers.
  
• More AI monopoly and inequality: By controlling such a large share of the AI chip market and infrastructure, Altman’s plan could also create more monopoly and inequality in the AI industry, if the access and the benefits of the chips and fabs are not distributed fairly and widely.
  
  This could give Altman and his partners more power and influence over the AI field, as well as over the industries and sectors that depend on AI.
  
  It could also widen the gap and the conflict between the AI haves and have-nots, both within and across countries and regions.
  
• More AI uncertainty and instability: By introducing such a radical and disruptive change in the AI industry, Altman’s plan could also create more uncertainty and instability in the AI field, if the outcomes and the impacts of the project are not predictable and manageable.
  
  This could pose technical and financial challenges, as well as ethical and social dilemmas, for the AI community and the society.
  
  It could also spark political and legal disputes, as well as security and safety threats, for the AI stakeholders and the regulators.

The Alternatives and Competitors of Altman’s Plan

Altman’s plan, while ambitious and audacious, is not the only or the first attempt to address the AI chip challenge. There are many alternatives and competitors in the AI chip market, both existing and emerging, who have their own visions and strategies. Some of them are:

• Nvidia: Nvidia is the undisputed leader and the main beneficiary of the AI chip boom, with its GPUs powering most of the large and powerful language models created by OpenAI, Alphabet, Meta, and others.
  
  Nvidia’s GPUs are designed for high-performance parallel computing, which is ideal for AI applications that involve processing large amounts of data and performing complex calculations.
  
  Nvidia also offers software tools and platforms, such as CUDA, TensorRT, and DGX, that make it easier and faster for developers and researchers to use its GPUs for AI. Nvidia’s market cap has soared to $1.72 trillion in 2023, surpassing many tech giants such as Amazon and Alphabet.
  
  However, Nvidia also faces some challenges and threats, such as the global chip shortage, the regulatory scrutiny, the customer dissatisfaction, and the competition from other chip makers and AI players.
• Intel: Intel is the world’s largest and the most established chip maker, with its central processing units (CPUs) dominating the personal computer and the server markets.
  
  Intel also produces other types of chips, such as field-programmable gate arrays (FPGAs), which are chips that can be reconfigured for different tasks, and neuromorphic chips, which are chips that mimic the structure and function of the brain.
  
  Intel also acquired several AI chip startups, such as Nervana, Movidius, and Habana Labs, to expand its AI portfolio. Intel’s CPUs are widely used for general-purpose computing, but they are not as efficient or as fast as GPUs for AI applications.
  
  Intel also lost its technological edge and its market share to its rivals, such as AMD and TSMC, due to its delays and defects in its chip production.
  
  Intel is trying to regain its leadership and reputation in the chip industry, by investing more in its own fabs, as well as in outsourcing and collaborating with other chip makers and AI players.
  
• AMD: AMD is Intel’s main competitor in the CPU market, and it has gained more popularity and profitability in recent years, thanks to its superior and cheaper chips. AMD also produces GPUs, which are based on the same architecture as Nvidia’s GPUs, and which are compatible with Nvidia’s software tools and platforms.
  
  AMD’s GPUs are mainly used for gaming and graphics, but they can also be used for AI applications, especially for training and inference at the edge. AMD also acquired Xilinx, a leading FPGA maker, to enhance its AI capabilities and offerings.
  
  AMD’s market cap has grown to $0.5 trillion in 2023, making it one of the most valuable chip companies in the world.
  
  However, AMD also faces some challenges and threats, such as the global chip shortage, the regulatory scrutiny, the customer satisfaction, and the competition from other chip makers and AI players.
  
• TSMC: TSMC is the world’s largest and the most advanced chip manufacturer, with its fabs producing chips for many chip designers and customers, including Nvidia, AMD, Apple, Qualcomm, and others.
  
  TSMC’s fabs can produce chips with the smallest and densest features, which enable higher performance and lower power consumption.
  
  TSMC also invests heavily in research and development, as well as in expanding and upgrading its fabs, to maintain its technological leadership and competitive advantage.
  
  TSMC’s market cap has reached $1.2 trillion in 2023, making it one of the most valuable chip companies in the world.
  
  However, TSMC also faces some challenges and threats, such as the global chip shortage, the geopolitical tension, the environmental impact, and the competition from other chip makers and AI players.
  
• Apple: Apple is one of the most innovative and influential tech companies in the world, with its products and services, such as the iPhone, the iPad, the Mac, the Apple Watch, the AirPods, the Apple TV, the App Store, the iCloud, and others, reaching billions of users and customers.
  
  Apple also designs its own chips, such as the A-series, the M-series, and the S-series, which power its devices and platforms. Apple’s chips are based on the ARM architecture, which is more energy-efficient and flexible than the x86 architecture used by Intel and AMD.
  
  Apple’s chips also incorporate AI features, such as the Neural Engine, which enable faster and smarter processing of tasks such as face recognition, natural language processing, and augmented reality.
  
  Apple’s market cap has soared to $3.5 trillion in 2023, making it the most valuable company in the world.
  
  However, Apple also faces some challenges and threats, such as the global chip shortage, the regulatory scrutiny, the customer loyalty, and the competition from other tech companies and AI players.
  
• Meta: Meta, formerly known as Facebook, is one of the most popular and powerful social media and internet companies in the world, with its platforms and services, such as Facebook, Instagram, WhatsApp, Messenger, Oculus, and others, connecting and engaging billions of users and customers.
  
  Meta also develops its own AI chips, such as the Zion and the Gideon, which are used for its data centers and its virtual reality devices. Meta’s AI chips are designed to handle large-scale and complex AI tasks, such as natural language processing, computer vision, and recommendation systems.
  
  Meta also offers software tools and platforms, such as PyTorch, Caffe2, and Spark AR, that make it easier and faster for developers and researchers to use its AI chips for AI. Meta’s market cap has grown to $1.8 trillion in 2023, making it one of the most valuable tech companies in the world.
  
  However, Meta also faces some challenges and threats, such as the global chip shortage, the regulatory scrutiny, the user privacy, and the competition from other tech companies and AI players.
  
• Alphabet: Alphabet is one of the most innovative and diversified tech companies in the world, with its products and services, such as Google, YouTube, Gmail, Google Maps, Google Photos, Google Assistant, Google Cloud, and others, serving and satisfying billions of users and customers.
  
  Alphabet also creates its own AI chips, such as the Tensor Processing Unit (TPU), which are used for its data centers and its cloud services.
  
  Alphabet’s AI chips are optimized for TensorFlow, its open-source software framework for AI, which is widely used by developers and researchers for AI.
  
  Alphabet also offers software tools and platforms, such as Colab, JAX, and AutoML, that make it easier and faster for developers and researchers to use its AI chips for AI.
  
  Alphabet’s market cap has reached $2.5 trillion in 2023, making it one of the most valuable tech companies in the world.
  
  However, Alphabet also faces some challenges and threats, such as the global chip shortage, the regulatory scrutiny, the user privacy, and the competition from other tech companies and AI players.
  
• OpenAI: OpenAI is one of the most ambitious and influential AI research organizations in the world, with its mission to ensure that AI is aligned with human values and can benefit all of humanity.
  
  OpenAI also develops some of the largest and most powerful language models in the world, such as GPT-3, GPT-4, and Codex, which can generate text, images, code, and more, based on natural language inputs.
  
  OpenAI also offers software tools and platforms, such as OpenAI Codex, OpenAI Playground, and OpenAI Scholars, that make it easier and faster for developers and researchers to use its language models for AI.
  
  OpenAI’s valuation has increased to $0.5 trillion in 2023, making it one of the most valuable AI organizations in the world.
  
  However, OpenAI also faces some challenges and threats, such as the global chip shortage, the ethical concerns, the social impact, and the competition from other AI organizations and players.

Conclusion

Altman’s plan to make AI chips is one of the most audacious and controversial projects in the AI field, with its potential to reshape the AI industry and the world.

Altman’s plan could bring more AI innovation and applications, more AI competition and diversity, and more AI infrastructure and resilience, but it could also create more AI waste and inefficiency, more AI monopoly and inequality, and more AI uncertainty and instability.

Altman’s plan also faces many technical and financial challenges and uncertainties, as well as ethical and social implications and dilemmas. Altman’s plan also has many alternatives and competitors in the AI chip market, both existing and emerging, who have their own visions and strategies.

The future of AI chips is uncertain and unpredictable, but it is also exciting and important. AI chips are not only the engines of AI, but also the enablers of AI.

AI chips are not only the products of AI, but also the drivers of AI. AI chips are not only the tools of AI, but also the goals of AI. AI chips are not only the means of AI, but also the ends of AI.

The following table summarizes some of the key features and facts of the main AI chip players mentioned in this article:

Player	Type	Architecture	Market Cap	AI Chip Name	AI Chip Use Case
Nvidia	Chip Designer	x86, GPU	$1.72T	GeForce, Quadro, Tesla, etc.	AI Training and Inference in Data Centers and Cloud
Intel	Chip Designer and Manufacturer	x86, CPU, FPGA, Neuromorphic	$0.25T	Xeon, Core, Nervana, Movidius, Habana, Loihi, etc.	General-Purpose Computing and AI in Data Centers, Cloud, and Edge
AMD	Chip Designer	x86, GPU	$0.5T	Ryzen, Epyc, Radeon, etc.	General-Purpose Computing and AI in Data Centers, Cloud, and Edge
TSMC	Chip Manufacturer	x86, ARM, GPU, etc.	$1.2T	N/A	Chip Production for Various Chip Designers and Customers
Apple	Chip Designer and Device Maker	ARM	$3.5T	A-series, M-series, S-series, etc.	Device and Platform Powering and AI in Mobile and Desktop
Meta	Tech Company and AI Chip Designer	x86, GPU	$1.8T	Zion, Gideon, etc.	AI Training and Inference in Data Centers and VR Devices
Alphabet	Tech Company and AI Chip Designer	x86, TPU	$2.5T	TPU, etc.	AI Training and Inference in Data Centers and Cloud
OpenAI	AI Research Organization and Language Model Developer	x86, GPU	$0.5T	N/A	Language Model Training and Inference in Data Centers and Cloud