Anna's Deep Dives

_{Just facts, you think for yourself}

A. U.S. Export Controls on China's AI: How They've Changed

In 2022, the U.S. started seriously limiting China's access to advanced AI tech. This was a big strategy change. The main goal was to slow down China's military updates and tech growth.

Early in 2025, the Biden administration made these controls even tighter with a new three-level system. Level 1 gave 18 allied countries mostly free access to U.S. AI tech. Level 2 put some limits on about 120 countries but allowed some access. Level 3, which included China, had very strict limits, basically stopping access to the top U.S. AI tech.

These steps came after more trade problems. By April 2025, taxes on some Chinese goods hit 145%. China fought back with taxes up to 125% on some U.S. goods. The U.S. also put about 140 Chinese companies on a special list to stop them from getting key U.S. tech and slow down China's AI work.

U.S. chip designer Nvidia saw this impact firsthand. The company expected to lose about $5.5 billion in sales. This was because of export rules on its H20 AI chip, which was popular in China. China used to make up about 15% of Nvidia's total sales. After the new controls, Nvidia's stock price dropped 5%.

But China is still trying to make its own semiconductors. Its "Made in China 2025" plan includes investing $150 billion by 2030 to do this. U.S. allies also reacted to the U.S. policy changes. By March 2025, countries like Japan and those in the European Union started changing their own export rules. They wanted to match U.S. rules more and stop important tech from getting to unfriendly countries.

Even with these efforts, some reports say groups are trying to get around these rules. They might be using cloud services or fake companies to hide who is really using the tech. The U.S. government is now trying to enforce the rules better to stop these tricks.

B. The "AI Diffusion Rule": What It Was, Why It Was Criticized, and Its Repeal

A big part of Washington's recent plan was the "Framework for Artificial Intelligence Diffusion." The Biden administration introduced this policy, often called the "AI Diffusion Rule," in January 2025. It was meant to control exports of advanced AI chips and other key tech.

It set up the three-level system for country access we talked about. Level One had 18 allied nations, like Japan and South Korea, with almost no limits. Level Two included countries like Brazil and India, which had some limits but could ask for more access. Level Three named unfriendly nations, like China and Russia, which were basically banned from getting advanced U.S. AI tech.

The AI Diffusion Rule had a clear goal. It tried to stop important U.S. tech from getting to groups that might use it against U.S. interests or its allies. The rule made companies follow new, often complicated, licensing rules. This often meant U.S. tech firms had to spend more money to operate.

The tech industry soon started criticizing it. Big companies, like Nvidia, worried that the strict rules could accidentally hurt U.S. companies in the important global AI market. Companies had to fully follow the AI Diffusion Rule by May 15, 2025. That deadline is today.

Some policy experts thought the rule might have an unexpected result. They said its limits could push some countries to work more closely with China for AI tech. But we might not see the full effect of the AI Diffusion Rule. The Trump administration announced plans to cancel this Biden rule, and the cancellation is happening today, May 15, 2025. This is another big change in U.S. AI export policy.

C. Biden's AI Rule: What It Tried to Do

Before it was canceled, the Biden administration's AI Diffusion Rule had clear goals. It aimed to control how American AI tech spread. A main goal was to stop advanced U.S. AI tech from getting to unfriendly countries, especially China and Russia.

The level system was the main way it did this. It tried to give different levels of access based on U.S. security and foreign policy. The plan was to help the U.S. stay the leader in AI. At the same time, it tried to handle the security risks from powerful AI spreading quickly. It used strict export controls and new licensing rules. The main idea was to slow down China's military and tech progress in AI.

D. Trump's New Plan: Simpler Rules or a Different Strategy?

There's a big policy change this week. The Trump administration is canceling the Biden "AI diffusion rule" today, May 15, 2025. This suggests they want to make the current complicated AI chip export rules simpler.

The new plan might get rid of the three-level system. Instead, the U.S. might make direct tech deals with certain partner countries. Countries like the UAE and Saudi Arabia have been named as possible early partners for this new plan.

The stock markets reacted quickly to news of the cancellation. Nvidia's stock price, which had been affected by the old rules, went up 3.1%. Other big U.S. tech companies like AMD and Intel also saw their stock prices rise. People who criticized the old Biden rule said it was too complicated. They argued it hurt American innovation and ability to compete. Industry groups like the Samsung Industry Association, and leaders like Nvidia's Vice President, had worried the old rules might make businesses look for tech help outside the U.S.

The new Trump administration's plan seems to be looking for a new balance. It wants to protect national security while also helping the U.S. economy grow and have strong tech partnerships with other countries. The stated goal is to make exporting easier. This would still try to stop important U.S. tech from getting to unfriendly countries, but maybe in different ways than the AI Diffusion Rule. This change could be a new strategy for handling AI chip exports.

E. The "Chip Security Act" Idea: Tracking Chips for Better Control

Adding to the U.S.'s changing strategy, Senator Tom Cotton introduced the "Chip Security Act" this month, May 2025. This new law idea aims to improve U.S. national security. It focuses on controlling the export of advanced AI chips, especially those going to China or that might end up there.

A key and controversial part of the Act is a rule that all advanced AI chips from the U.S. must have location tracking systems. The clear reason for this tracking is to stop chips from being smuggled and used by groups or for reasons that aren't allowed.

The Chip Security Act tries to add to the current CHIPS Act. The CHIPS Act gave $53 billion to help U.S. semiconductor production. Under this new proposed act, companies exporting these advanced AI chips would have to report any suspicious activity or if chips go to the wrong place. They'd have to start doing this within 180 days if the act becomes law.

This bill was introduced because of ongoing national security worries. There have been reports that top U.S. chips, like Nvidia's, have been smuggled into China even with current export rules. This illegal movement of chips could help China's military AI. It seems like both political parties are interested. Congressman Bill Foster is reportedly planning to introduce a similar bill in the House, which suggests both parties agree that stronger protection is needed.

F. How Location Tracking Could Stop Illegal Chip Shipments

The Chip Security Act mainly relies on location checks. This would add a new way to control where AI chips go. The proposal says AI chips must have built-in location tracking systems. This means they would need some way to always track their location.

This could allow for "geofencing." That means a chip might stop working if it's found outside an approved area or with someone who isn't allowed to have it. Businesses exporting these chips would also have strict new reporting rules. They'd have to tell U.S. authorities if they find out their products are being misused, tampered with, or sent to the wrong people.

The main goal is to greatly improve how the U.S. tracks and controls important semiconductor tech after it leaves the country. A key aim is to stop countries like China from illegally getting and misusing advanced U.S. tech for their military or other important plans. This deals with a real problem: reports say as many as 12,500 top Nvidia AI chips might be smuggled into China each year, getting around current export rules. Location tracking tries to make these illegal shipments much harder and easier to find.

G. Chip Tracking Problems: Privacy, National Rights, and Practical Issues

The idea of requiring location tracking in AI chips is already causing a lot of disagreement. Privacy is a big worry. Putting tracking tech in chips could create a lot of sensitive location data. This makes people ask who can see this data and how it's kept safe. It's natural to compare this to current privacy rules like Europe's GDPR. The GDPR, for example, says you need clear permission before tracking people. That idea might not fit with tracking all chips.

Issues about national rights also come up. Other countries might see U.S.-required tracking tech on chips used in their country as stepping on their rights. This could cause problems with diplomats and make important trading partners resist.

Besides policy, there are many practical problems. It's a big challenge to create a tracking system that's impossible to trick or get around. Enemies would surely try to turn off or fake these tracking systems. The cost to put in these tracking systems would also be very high. Chip makers would have to pay for it at first. This could affect chip prices and how well U.S. companies compete if other countries don't do the same. Finding a balance between security and these practical and ethical issues will be hard for lawmakers.

H. What's the U.S. Strategy? Stopping, Separating, or Just Slowing Down China?

Washington's many actions try to control exports of advanced AI tech, especially to China. This includes limiting access to key high-tech chips, advanced design software, and important manufacturing skills. The clear goals are: keep the U.S. ahead in AI, stop China's military from getting the newest AI, and improve overall U.S. national security.

But it's not clear if just trying to stop China (containment) will work. The quick rise of Chinese AI companies like DeepSeek shows China can innovate. DeepSeek made competitive AI models for about $6 million, while U.S. companies like OpenAI spent over $100 million on GPT-4. This shows China can get results with less money.

China's own semiconductor industry is also improving quickly. The country invested about 344 billion yuan (around $47.5 billion) in this area in 2022 alone. China says its goal is to make 70% of its own semiconductors by 2030. This local growth suggests U.S. export rules might not really slow down China's long-term tech progress in making chips.

China also produces a lot of AI research. Chinese researchers recently published 12,450 AI research papers, a bit more than the U.S. number of 12,030. The Chinese government invested about $184 billion in AI between 2000 and 2023, helping over 9,623 AI companies. This huge investment helps China make progress, even while U.S. rules try to limit its access to certain foreign tech.

The level-based access system in the Biden AI Diffusion Rule, which was just canceled, also made some people worry. Some experts thought it might push countries in lower levels to work more with China to get around U.S. limits. Washington's strategy seems to mix trying to stop China, partly separating tech, and trying to slow down China's AI progress. Whether these goals will work out in the long run is still being heavily debated.

I. Side Effects and Unexpected Results: How U.S. Policies Are Really Affecting Things

U.S. export controls on advanced AI chips, which started getting stricter in October 2022 and were detailed in the January 2025 three-level system (now being canceled), have caused major side effects. These policies tried to limit China's access but also ended up blocking many U.S. companies from the valuable Chinese market.

Nvidia, a top U.S. chipmaker, said it lost $5.5 billion directly because of these rules. It's not just one company; the whole U.S. semiconductor industry has been affected. Big companies like Intel and AMD reported money problems, and the industry's total market value has dropped by about $130 billion since the rules started in 2022.

An unexpected result might be that Chinese competitors get stronger. Experts think that U.S. rules, along with Chinese government money, could help Chinese companies speed up making their own competitive chips, maybe even cheaper ones. Companies like Huawei and Biren Technology are already showing quick progress and could challenge U.S. tech leadership in some areas.

Inside the U.S., there are challenges too. Proposed taxes on imported semiconductors, based on national security reviews (Section 232), suggested taxes could be over 100%. These taxes could make it much more expensive for U.S. companies that need imported parts to make things. Also, the U.S. still depends on other countries for important manufacturing supplies. About 60% of the special chemicals needed to make semiconductors are still imported, showing ongoing weaknesses in the supply chain.

The fact that very efficient Chinese AI models are appearing, like DeepSeek's V3 model (reportedly trained for under $6 million compared to OpenAI's GPT-4 at over $100 million), suggests that export rules might not be stopping Chinese innovation in AI software. Before the Biden rules were canceled, Nvidia had estimated they might cut down the available global market for certain advanced GPUs by as much as 80%. This shows how big the lost sales chances could be for U.S. companies.

Making sure these complex rules are followed is still a big challenge. Reports say that smuggling banned AI chips into China is becoming a bigger problem. Some estimate that as many as 12,500 top Nvidia chips might be getting into China illegally each year. U.S. Congressman Bill Foster is working on a law to track these chips to fight these illegal shipments.

U.S. export rules also make things complicated with friendly countries. Countries like Singapore, which used to send a lot of semiconductor-making equipment to China, now have tough decisions to make. U.S. allies might have to pay more to run their own industries and have less access to important tech if the U.S. thinks they might be a way for tech to get to the wrong places. 

Overall, Washington's export rules and taxes are changing the global semiconductor world. While they are mostly meant to protect U.S. national security, these policies might accidentally help China build up its own tech skills and create new money problems for U.S. industry.

We don’t take shortcuts, chase headlines, or push narratives. We just bring you the news, straight and fair. If you value that, click here to become a paid subscriber—your support makes all the difference.

Table of Contents

(Click on any section to start reading it)

I. AI Chips: The New Cold War

• Why AI Chips Are Key to National Power

• The "Silicon Curtain" and the New AI Cold War

• DeepSeek: A Chinese AI Company Shakes Things Up

II. AI Chips: What They Are, How They're Made, and Why There's Tension

• What AI Chips Are: A Quick Look at GPUs, ASICs, and More

• How Chips Are Made: The Complicated Global Supply Chain

• U.S. vs. China: Why They Don't Trust Each Other on Tech

• The AI Race: Why Being Top in Chips is a Must for World Leaders

III. U.S. Moves: Export Bans, Policy Changes, and Trying to Shape the Future

• U.S. Export Controls on China's AI: How They've Changed

• The "AI Diffusion Rule": What It Was, Why It Was Criticized, and Its Repeal

• Biden's AI Rule: What It Tried to Do

• Trump's New Plan: Simpler Rules or a Different Strategy?

• The "Chip Security Act" Idea: Tracking Chips for Better Control

• How Location Tracking Could Stop Illegal Chip Shipments

• Chip Tracking Problems: Privacy, National Rights, and Practical Issues

• What's the U.S. Strategy? Stopping, Separating, or Just Slowing Down China?

• Side Effects and Unexpected Results: How U.S. Policies Are Really Affecting Things

IV. Nvidia's Challenges, Industry Effects, and the Cost of Following Rules

• Nvidia: Stuck Between U.S. Rules and the Chinese Market

• Nvidia's Modified Chips for China (H20, etc.): New Ideas or Just Weaker?

• Weaker Chips for China: How This Affects Their AI Tech

• How Chinese Customers Are Adjusting

• The Cost: Billions Lost by Nvidia and Others Due to Export Rules

• How Export Rules Affect Chip Companies Worldwide

• Shaky Markets: Investor Worries, Stock Swings, and Rethinking AI Value

V. China's Answer: DeepSeek, Local Chips, and Aiming for AI Independence

• DeepSeek Shakes Things Up

• Smarter, Not Harder? DeepSeek's Efficient AI

• DeepSeek vs. Big Spenders: A New Way for AI?

• DeepSeek's Open Approach: Good for All or Smart Business?

• More Than DeepSeek: China's Own AI Chip Makers Are Growing

• Huawei, Alibaba, and Other New Chinese Chip Makers

• China's AI Chips: Catching Up or Doing Things Differently?

• China's Big Plan: Self-Sufficiency with Top Companies and Big Money

• Big Hurdles: Problems for China's Chip Independence Plan

• China's AI: Copying or Creating New Things?

VI. A Changed World: New Politics, Innovation Challenges, and What the AI Chip War Leaves Behind

• Broken Supply Chains: New Tech Groups and a More Split Digital World

• What's Next for Innovation: A Boost or a Bust from Competition?

• Tech "Islands": A Risk to Global Progress?

• Separate Work, Surprise Discoveries?

• Power Shifts: How AI Chips Are Changing Global Politics

• Ethics in the AI Race: Will Security Beat Safe Development?

• After 2025: What Could Happen with AI, U.S.-China, and World Peace?

• Lasting Effects: The AI Chip Cold War's Unavoidable Impact and Facing an Unsure Future

Baked with love,

Anna Eisenberg ❤️