USA and China Fight Over AI Chips

May 15, 2026, 04:50
ChinaTechnology
A2

USA and China Fight Over AI Chips

Introduction

The USA wants to be the best at AI. It is checking its rules about selling computer chips to China.

Main Body

Some people say the USA is faster than China. They want stricter rules. They do not want China to copy USA AI models. This keeps the world safe. Other people say China is slow. They say China does not have good data or fast computers. This makes the gap between the two countries bigger. The USA sells some chips to China but asks for more money. Some people try to steal chips or hide them. China cannot make enough chips in its own factories.

Conclusion

The USA wants to make money, but it also wants to stay the leader in AI.

Learning

⚡ Comparing Things

In the text, we see words that describe a difference in speed and quality. To reach A2, you need to know how to describe these opposites.

The Speed Scale

  • Fast →\rightarrow The USA is faster than China.
  • Slow →\rightarrow China is slow.

The Quality Scale

  • Best →\rightarrow The USA wants to be the best.
  • Good →\rightarrow China does not have good data.

💡 Pro Tip: The "-er" Pattern When we compare two things, we often add -er to the word.

Fast →\rightarrow Faster

Example: "The USA is faster than China."

🔑 Key Action Words These words tell us what is happening in the fight for AI:

  • Sell: To give something for money.
  • Steal: To take something without asking.
  • Copy: To make the same thing as someone else.

Vocabulary Learning

chip (n.)
A small piece of semiconductor material used in electronics
Example:The factory makes chips for computers.
rule (n.)
A regulation or law that people must follow
Example:The new rule limits who can buy AI chips.
copy (v.)
To make a duplicate of something
Example:China wants to copy the USA's AI models.
safe (adj.)
Protected from danger or harm
Example:These new rules keep the world safe.
data (n.)
Facts or information collected for analysis
Example:The company needs more data to improve its AI.
factory (n.)
A building where goods are made
Example:China cannot make enough chips in its own factories.
leader (n.)
A person or country that is in charge or ahead
Example:The USA wants to stay the leader in AI.
money (n.)
Currency used to buy goods or services
Example:The USA sells chips to China for more money.
sell (v.)
To exchange something for money
Example:The company will sell its chips to China.
gap (n.)
A space or difference between two things
Example:The gap between the two countries grows larger.
B2

Different Views on US Semiconductor Export Controls and AI Leadership

Introduction

The United States is currently reviewing whether its restrictions on semiconductor exports are effective in keeping a technological lead over China in the field of advanced artificial intelligence.

Main Body

There are conflicting opinions about the current technological gap between the two nations. For example, the company Anthropic claims that the US has a short window of time to secure a lead of one to two years. They emphasize that the government must implement stricter chip controls and stop 'distillation attacks,' where smaller models are trained using outputs from advanced AI. Anthropic asserts that if both countries reach the same level of capability, global safety could be at risk because companies might release untested models too quickly. On the other hand, experts like Zhang Chi from Peking University suggest that China's progress is slowing down due to a lack of high-quality data and hardware, meaning the gap may actually be widening. Meanwhile, the US government is struggling to balance security and business. Although the Trump administration has banned many Nvidia and AMD sales to China, a partial exception was made in August to allow H200 chip sales, provided a 25% government tax is paid. During a recent visit to Beijing, US Trade Representative Jamieson Greer stated that chip controls were not the main topic of discussion. He argued that the decision to import H200 chips is a sovereign choice for China, while emphasizing that the US must balance national security with the economic benefits of selling to foreign markets. However, this situation is complicated by illegal smuggling and the fact that Chinese factories are struggling to produce enough chips, leading to a shortage of computing power for Chinese AI models.

Conclusion

The United States continues to adjust its export policies to find a balance between making money through trade and preventing China from reaching the same level of AI technology.

Learning

⚡ The 'Weight' of Words: Shifting from Simple to Sophisticated

At A2, you describe things using basic verbs: "The US says..." or "China thinks..." To reach B2, you need Reporting Verbs. These change the 'flavor' of the sentence and show exactly how someone is speaking.

🔍 Contrast Analysis

Look at how the article avoids using the word "say" repeatedly. Instead, it uses these precise tools:

A2 Level (Simple)B2 Level (Precise)What it actually means
say / thinkClaimTo say something is true, even if others don't believe it yet.
say / wantEmphasizeTo give special importance to a point.
say / believeAssertTo state something with strong confidence.
say / thinkArgueTo give reasons to support an opinion.

🛠️ How to apply this to your speaking

When you move to B2, stop using "say" for everything. Try this mental map:

  • If you are 100% sure →\rightarrow Assert
  • If you are trying to persuade someone →\rightarrow Argue
  • If you are highlighting a danger →\rightarrow Emphasize
  • If it's a theory or a possibility →\rightarrow Claim

💡 Pro-Tip: The 'Balance' Phrase

Notice the phrase: "struggling to balance security and business."

Instead of saying "It is hard to do both," use "struggling to balance [X] and [Y]." This is a high-level structure that allows you to discuss two opposing ideas in one elegant sentence.

Vocabulary Learning

implement (v.)
To put into effect or carry out a plan or policy.
Example:The company will implement new safety protocols next month.
distillation (n.)
A process of separating a substance from a mixture by heating.
Example:Distillation is used to purify liquids in chemical labs.
capability (n.)
The ability or power to do something.
Example:Her capability to solve complex problems impressed everyone.
safety (n.)
The condition of being protected from harm or danger.
Example:Road safety is a top priority for the city council.
progress (n.)
Forward or onward movement toward a goal.
Example:The project made steady progress over the summer.
high-quality (adj.)
Of superior standard or excellence.
Example:They only sell high-quality handmade furniture.
balance (v.)
To keep something in a stable or equal position.
Example:She must balance her workload with her personal life.
banned (adj.)
Prohibited from being used or allowed.
Example:The use of plastic bags is banned in many cities.
sovereign (adj.)
Independent and self-governing.
Example:A sovereign nation has the right to make its own laws.
shortage (n.)
An insufficient supply of something.
Example:There is a shortage of skilled workers in the tech industry.
computing (adj.)
Relating to computers or the processing of information.
Example:Computing power has increased dramatically in recent years.
administration (n.)
The group of people running an organization or government.
Example:The university administration announced new tuition fees.
C2

Strategic Divergence Regarding United States Semiconductor Export Controls and Artificial Intelligence Hegemony

Introduction

The United States is currently evaluating the efficacy of its semiconductor export restrictions to maintain a technological advantage over China in the domain of frontier artificial intelligence.

Main Body

The discourse regarding AI leadership is characterized by conflicting assessments of the current technological gap. Anthropic posits that the United States possesses a transient window to secure a twelve-to-twenty-four-month lead in frontier capabilities, provided that the administration implements more stringent chip export controls and mitigates 'distillation attacks'—the process by which smaller models are trained using advanced AI outputs. The organization asserts that a parity in capabilities would jeopardize global safety and governance, as competitive pressures might incentivize the premature release of insufficiently tested models. Conversely, academic perspectives, such as those provided by Peking University's Zhang Chi, suggest that China's progress is impeded by a deficit of high-quality training data and advanced hardware, implying a widening rather than narrowing gap. Institutional responses to these dynamics remain complex. While the Trump administration has maintained prohibitions on Nvidia and AMD sales to China, a partial derogation was established in August, permitting H200 chip sales subject to a 25% government levy. Despite the presence of high-profile executives like Jensen Huang during President Trump's recent diplomatic visit to Beijing, U.S. Trade Representative Jamieson Greer indicated that semiconductor export controls were not a primary subject of bilateral deliberations. Greer characterized the decision to permit H200 imports as a sovereign determination for China, emphasizing the necessity of balancing national security imperatives with the economic benefits of overseas market penetration. This tension is further exacerbated by clandestine procurement efforts, exemplified by the relabeling of Nvidia chips to circumvent customs, and the struggle of Chinese domestic fabrication plants to achieve scalable output, which has resulted in the rationing of computing power for Chinese AI models.

Conclusion

The United States continues to calibrate its export policies to balance commercial interests against the strategic objective of preventing Chinese AI parity.

Learning

The Architecture of Institutional Nominalization

To transition from B2 to C2, one must move beyond describing actions and begin conceptualizing processes. This text is a masterclass in Nominalization—the linguistic strategy of turning verbs (actions) and adjectives (qualities) into nouns. This shifts the focus from 'who is doing what' to 'what systemic phenomenon is occurring.'

⚡ The 'C2 Shift': From Action to Concept

Compare these two modes of expression found in the text:

  • B2 Logic (Verbal/Active): The US is trying to decide if its export restrictions work so it can stay ahead of China.
  • C2 Execution (Nominalized): "The United States is currently evaluating the efficacy of its semiconductor export restrictions to maintain a technological advantage..."

Analysis: By replacing "if it works" with "efficacy" and "staying ahead" with "technological advantage," the author removes the human agent and creates an objective, analytical distance. This is the hallmark of academic and diplomatic discourse.

🔍 Dissecting the 'Abstract Density'

Observe how the text clusters nouns to create high-density meaning. This is known as Noun Phrase Expansion. Note these specific instances:

  1. "Sovereign determination" →\rightarrow Instead of saying "The government decided for itself," the author uses a noun phrase to frame the decision as a matter of state right and legality.
  2. "Clandestine procurement efforts" →\rightarrow Instead of "People are secretly buying things," we have a conceptual category of activity.
  3. "Strategic objective of preventing Chinese AI parity" →\rightarrow This collapses a complex geopolitical goal into a single, formal object.

🛠 Sophisticated Collocations for the High-Level Learner

C2 mastery is not about 'big words,' but about lexical precision. The text utilizes specific pairings that signal authority:

B2 ApproximationC2 Precision (from text)Nuance Added
Temporary gapTransient windowSuggests a fleeting, critical opportunity.
Rule changePartial derogationLegalistic term for a specific exemption.
BalancingCalibrate its policiesImplies precise, incremental adjustment.
Getting aroundCircumvent customsFormalizes the act of evasion.

The C2 Takeaway: To elevate your writing, stop focusing on the actor (the subject) and start focusing on the concept (the noun). When you describe a situation, ask yourself: "Can I turn this action into a phenomenon?"

Vocabulary Learning

efficacy (n.)
the ability to produce a desired or intended result
Example:The efficacy of the new policy was demonstrated by a measurable decline in emissions.
semiconductor (n.)
a material with electrical conductivity between that of a conductor and an insulator, used in electronic components
Example:Semiconductor manufacturers are investing heavily in research to improve chip performance.
restrictions (n.)
limitations or prohibitions imposed on something
Example:The new restrictions on data transfer have slowed international collaboration.
technological (adj.)
relating to or using technology
Example:Technological innovations are driving the rapid growth of renewable energy.
advantage (n.)
a condition or circumstance that puts one in a favorable or superior position
Example:Having a superior design gave the company an advantage over its competitors.
domain (n.)
a particular sphere or field of activity or interest
Example:She is a leading expert in the domain of machine learning.
frontier (n.)
the extreme limit or border of a region or field, especially one that is unexplored or developing
Example:The quantum computing frontier promises unprecedented computational power.
discourse (n.)
written or spoken communication or debate on a particular topic
Example:The academic discourse on climate change has evolved over the last decade.
characterized (v.)
described or depicted in terms of particular qualities
Example:The project was characterized by meticulous planning and rigorous testing.
conflicting (adj.)
presenting opposing or contradictory views or interests
Example:The conflicting reports made it difficult to determine the true cause.
assessments (n.)
evaluations or judgments of something
Example:The assessments of the program revealed significant gaps in funding.
gap (n.)
a space or opening, or a difference between two levels
Example:There is a widening gap between the rich and the poor in many societies.
posits (v.)
puts forward or suggests as a premise
Example:The researcher posits that increased funding will accelerate development.
transient (adj.)
lasting only for a short time; temporary
Example:The transient nature of the signal required rapid sampling.
window (n.)
a period of time during which something can be done
Example:There is a narrow window of opportunity to negotiate a treaty.
secure (v.)
to obtain or protect something
Example:They will secure the data by encrypting it before transmission.
stringent (adj.)
strict or severe in enforcement or application
Example:Stringent safety regulations were implemented after the incident.
mitigates (v.)
makes something less severe or harsh
Example:Proper ventilation mitigates the risk of overheating.
distillation (n.)
the process of separating components by heating
Example:Distillation of crude oil yields multiple fuel products.
attacks (n.)
aggressive actions or assaults
Example:Cyber attacks on the infrastructure have increased in frequency.
process (n.)
a series of actions or steps taken to achieve a result
Example:The process of manufacturing the device takes several weeks.
advanced (adj.)
highly developed or sophisticated
Example:The advanced sensors can detect minute changes in temperature.
outputs (n.)
results or products produced by a system
Example:The outputs of the experiment were recorded for analysis.
asserts (v.)
states or declares something confidently
Example:He asserts that the results are statistically significant.
parity (n.)
equality or equivalence in status or performance
Example:The company aims to achieve parity with global competitors.
jeopardize (v.)
to put at risk or endanger
Example:Failure to comply could jeopardize the company's license.
governance (n.)
the act or process of governing or managing
Example:Effective governance is essential for sustainable development.
competitive (adj.)
involving rivalry or contest
Example:A competitive market forces firms to innovate constantly.
incentivize (v.)
provide motivation or reward to encourage a particular behavior
Example:The government incentivizes renewable energy projects with tax credits.
premature (adj.)
occurring before the usual or expected time
Example:Premature deployment of the software led to numerous bugs.
insufficiently (adv.)
not enough or inadequate
Example:The training dataset was insufficiently diverse.
academic (adj.)
related to education or scholarship
Example:Academic research often informs policy decisions.
impeded (v.)
obstructed or hindered
Example:The project was impeded by funding shortages.
deficit (n.)
a shortage or lack
Example:The budget deficit rose to record levels.
high-quality (adj.)
of superior standard or excellence
Example:The high-quality materials ensured durability.
hardware (n.)
physical components of a computer system
Example:New hardware upgrades improved the system's speed.
implying (v.)
suggesting or indicating without stating directly
Example:Her silence was implying a lack of confidence.
widening (n.)
the act or process of becoming broader or larger
Example:The widening gap between urban and rural areas is concerning.
narrowing (n.)
the act or process of becoming smaller or more limited
Example:The narrowing of the river channel caused flooding.
institutional (adj.)
pertaining to an institution or established organization
Example:Institutional reforms are necessary for transparency.
responses (n.)
reactions or replies to a stimulus
Example:The responses to the crisis were swift and coordinated.
dynamics (n.)
forces or processes that produce change or activity
Example:The dynamics of the market are constantly shifting.
remain (v.)
continue to exist or stay in a particular state
Example:The problem remains unresolved despite efforts.
complex (adj.)
consisting of many interconnected parts
Example:The complex architecture required specialized knowledge.
prohibitions (n.)
bans or restrictions
Example:The prohibitions on pesticide use were enacted to protect wildlife.