Summary of Key Points
The U.S. government is becoming one of the most aggressive investors in the field of artificial intelligence (AI) globally: Over the past year, it has invested $11.05 billion in 11 AI chip and upstream infrastructure companies, exceeding the $7 billion invested by Sequoia Capital’s flagship fund. Intel alone received $8.9 billion from these investments. This marks a shift in the U.S. government’s role from that of a “crisis firefighter” (such as bailing out banks and the automotive industry in 2008) to a “strategic industrial investor.” The primary goal is to control key aspects of the domestic AI supply chain, addressing the issue of being technologically advanced but lacking industrial security—by avoiding reliance on overseas suppliers (like TSMC) and coping with challenges such as chip shortages and Sino-U.S. technological rivalry. These companies cover core segments of the AI industry chain, including chip manufacturing, equipment, materials, storage, and security.
Detailed Analysis
1. Investment Scale: More Generous than Top Venture Capital Firms
What does the $11.05 billion invested by the U.S. government represent? Sequoia Capital, one of the world’s leading venture capital firms, has only raised $7 billion in its new flagship fund this year, yet the U.S. government’s investment already surpasses this amount. Even more noteworthy is that Intel received a single contribution of $8.9 billion, accounting for over 80% of the total. Moreover, this investment has already turned a profit: when the government purchased Intel’s shares, their valuation was around $89.9 billion; now Intel’s market value exceeds $500 billion, meaning the government has made a profit of over $40 billion—equivalent to several times the size of Sequoia Capital’s funds.
This is not a minor investment but a significant “heavyweight bet” at the national level, indicating that the AI industry chain has become a core strategic asset for the United States.
2. Role Transformation: From “Firefighter” to Proactive “Layout”
In the past, the U.S. government directly held shares in companies only during crises, such as bailing out banks during the Great Depression of the 1930s or General Motors during the financial crisis in 2008. However, the approach has changed: it no longer waits for industries to collapse before taking action but proactively invests in “bottleneck” areas like semiconductors and quantum computing.
Why this change? In the past, the U.S. believed that controlling design and capital was sufficient, leaving manufacturing to global supply chains. However, events such as chip shortages (e.g., a lack of automotive chips during the pandemic), Sino-U.S. technological tensions (e.g., Huawei’s chip restrictions), and the Russia-Ukraine war (disruption of key mineral supplies) have made it clear that without domestic production capacity, even advanced technologies can be vulnerable. For example, if TSMC in Taiwan experiences issues with its supply chain, the U.S. would lose its ability to produce AI chips. Therefore, the government must take direct action to control these critical industries.
3. Focus of Investment: Targeting Critical Nodes in the AI Industry Chain
These 11 companies were carefully selected, each targeting a key aspect of the AI industry chain:
- Chip Manufacturing: The Cornerstone
Intel is the only U.S. company capable of covering the entire chip manufacturing process, from design to packaging, similar to SMIC in China. By investing in Intel, the government aims to maintain an advanced manufacturing system that does not rely on TSMC. This ensures that the U.S. is not at risk if there are shortages in high-end chips.
- Data Transmission: Speeding Up the “Information Superhighway”
As AI chip performance increases, so does the demand for fast data transmission. Traditional copper wires are slow and inefficient. GlobalFoundries focuses on “co-packaged optics,” which involves attaching fiber optic cables directly to chips to enable faster data transfer between GPUs (like switching from electrical wires to fiber optics). The government has invested $375 million in quantum chip manufacturing and $300 million in silicon photonics technology for high-speed interconnectivity in AI data centers.
- Breakthroughs in Lithography Machines: xLight
Advanced chips require EUV lithography machines, which are monopolized by ASML. xLight aims to develop a large-scale light source that can be shared among multiple lithography machines, making them more cost-effective than the current individual light sources. The government has invested $150 million in their research to break this monopoly.
- Memory Bottlenecks: Overcoming Barriers
A common issue with AI chips is that they can process data quickly but wait long for data retrieval from memory. Kepler Computing uses 3D stacking technology to move memory directly next to GPUs, eliminating the need for data to travel back and forth, thus improving performance and energy efficiency. The government has invested $245 million in this innovation.
- Materials and Security: Invisible Guarantees
Thintronics develops low-loss materials to reduce signal degradation in AI servers; OBSIDIA provides digital identification for chips to prevent counterfeit products and track their distribution, which is crucial for high-end AI chips under export controls. The U.S. government needs to know where these chips are being used.
4. Behind the Logic: Industrial Security Over Market Efficiency
This investment is part of a “reindustrialization” strategy. While the U.S. once believed in the free market, it is now willing to use government funds to ensure industrial security. For example, Intel’s wafer manufacturing business may not be profitable, but the government still invests because it is the only advanced domestic option.
This reflects a new trend in global technological competition: AI and semiconductors are no longer purely market-driven but part of strategic rivalries between nations. Whoever controls the key aspects of the industry chain holds significant influence. The U.S. government’s goal is clear: to maintain absolute leadership and supply chain security in the AI era.
5. Potential Implications for Global AI Competition
The U.S. government’s actions will lead to several changes:
- A Stronger Domestic AI Industry Chain: The U.S. is strengthening its own capabilities from manufacturing to equipment and materials, reducing its dependence on overseas suppliers.
- Intensified Global AI Competition: Other countries (such as China and the EU) may follow suit by increasing government investment in the AI industry chain to avoid falling behind.
- Government Investment as a New Approach: Previously, AI relied on venture capital; now, with national capital entering the picture, it could accelerate breakthroughs in costly but critical technologies (e.g., quantum chips and EUV light sources).
In summary, the U.S. government’s investment is not just about spending money but about strategically positioning itself to secure its dominance in the AI era for the next decade.