The U.S. and China have stepped back from tariff brinkmanship for now — but don’t mistake handshakes for harmony. The real trade fight is just beginning, and it’s moving into murkier terrain: microchips, AI infrastructure, and data sovereignty.
And neither side is going to back down from the fierce contest over who controls the future of technology.
While tariffs still loom large, both sides have agreed to a temporary ceasefire and a framework to bring sky-high levies back down to earth. But the next front in the trade war isn’t about crates of soybeans or textile factories. It’s about who owns the rails of tomorrow’s digital economy — and who gets boxed out.
The U.S. and China have been circling — and increasingly clashing over — critical AI technologies for roughly a decade. Washington has heightened controls on semiconductor exports, limiting China’s access to advanced AI chips and pressuring allies to follow suit. Beijing is doubling down on self-reliance, retaliating with restrictions on critical minerals and squeezing the parts of the supply chain it still dominates.
“There are a lot of issues [the trade framework] doesn’t address,” Susan Shirk, a professor at the University of California, San Diego School of Global Policy and Strategy, said in an interview. “The tech tensions are going to be a very high priority.”
The Trump administration recently rolled back the Biden-era AI diffusion rule, which regulated global distribution of advanced AI technology — a move praised by top CEOs that has cracked open a lucrative market for U.S. tech giants. Now, companies such as Nvidia (NVDA) are cleared to ship millions of high-performance GPUs to countries including Saudi Arabia and the United Arab Emirates, supercharging those nations’ ambitions to become regional AI powerhouses.
The moves into the Middle East contrast starkly with Washington’s approach toward China.
Since 2022, the U.S. has been methodically severing Beijing’s access to the most advanced semiconductors and the tools needed to manufacture them, targeting the building blocks of AI and next-gen military systems. Those export controls, first implemented under President Joe Biden, have continued to varying degrees in President Donald Trump’s second administration.
Now, the scope of restrictions is poised to widen.
Policymakers in Washington are weighing measures that go beyond simply blocking hardware sales and moving into the realm of research and collaboration. That includes monitoring cloud infrastructure partnerships, joint academic projects, and potential third-country workarounds.
China isn’t swinging back with blunt instruments like tariffs. It’s opting for precision hits. Late last year, Beijing imposed export controls on critical minerals such as gallium and germanium — both vital for chipmaking and EV production.
And behind the scenes, China is pouring billions into building its tech lifeboat. The country’s top chipmakers — such as Semiconductor Manufacturing International Corporation (SMIC) and Yangtze Memory Technologies (YMTC) — are quietly expanding capacity, even under the weight of U.S. export controls. China’s leadership knows that even if the country can’t leapfrog the U.S. in the near term, it can at least build a parallel tech ecosystem — one resilient enough to survive sanctions, export bans, or a full-scale decoupling.
No two companies better illustrate the stakes and asymmetries than China’s Huawei and America’s Nvidia.
Huawei has become the face of China’s tech ambitions — and of U.S. attempts to stifle them. Last week, the Commerce Department warned that Huawei’s Ascend chips are subject to sweeping export controls. Then, China warned Tuesday that the move “seriously undermined the consensus” that the two countries reached when they agreed to pause tariffs, suggesting China might retaliate.
Huawei, long seen as the face of China’s tech rise, was hit hard by U.S. sanctions in 2019. But, denied access to U.S. chips and software, the company didn’t fold — it pivoted. It partnered with SMIC to develop chips for its Mate 60 Pro smartphones and is now pushing its Ascend AI chips as homegrown alternatives to Nvidia’s. Huawei’s chips aren’t on par with Nvidia’s, but they’re part of Beijing’s “tech independence” push.
Meanwhile, Nvidia, whose high-end GPUs power almost every leading AI model, has become a geopolitical linchpin. The U.S. has repeatedly restricted its ability to sell to China, citing military concerns. Now, Nvidia walks a tightrope: maintaining global dominance while complying with U.S. restrictions. How Washington treats Nvidia is a window into the Trump administration’s broader strategy. The U.S. wants to preserve Nvidia’s status without letting China piggyback on it.
Huawei and Nvidia are mirrored chess pieces in this larger game.
The chip war is about hardware, but the software battle is just as critical. The U.S. leads in large foundation models, with OpenAI, Anthropic, and others training massive models on U.S.-based infrastructure. These are tightly controlled via APIs and export rules — a “controlled openness” approach designed to balance innovation with national security.
OpenAI is the crown jewel of the U.S. AI stack — not just because it makes ChatGPT, but because it helped kick off an AI arms race in which the rest of the world is scrambling to catch up.
But China is countering with open-source models, the most notable being DeepSeek, from China’s Moonshot AI, which rattled Silicon Valley earlier this year. By releasing training weights and model architecture, DeepSeek sidesteps export restrictions and helps build domestic capacity.
DeepSeek and OpenAI offer a compelling contrast in how China and the U.S. are positioning themselves in the next phase of the tech war: not just around chips and hardware, but around AI capabilities, open-source strategy, and control of foundational models.
When DeepSeek’s “R1” model dropped early this year, it stunned the market. Despite running on less powerful hardware, it achieved performance levels that were previously thought to be exclusive to U.S. firms. Nvidia stock dropped 18% in a day — a $589 billion market cap loss — reflecting investor fear that efficient models could reduce the demand for high-end GPUs.
But what makes DeepSeek especially important is that it shows China is adapting, not just reacting.
“If each nation’s current trajectory holds … the battle for AI supremacy will be fought not between the United States and China, but between high-tech Chinese cities like Shenzhen and Hangzhou,” Kyle Chan, a researcher at Princeton University, wrote recently in The New York Times.
As AI becomes the backbone of national competitiveness, who owns the data — where it’s stored, how it’s governed, and who gets to access it — is turning into a geopolitical fault line. For the U.S., the concern is clear: Sensitive data flowing to China is a national security risk. That’s why Biden, and now Trump again, pushed forward a suite of export controls and executive orders aimed at walling off U.S. citizens’ data from potential misuse abroad.
The logic is that data is power, and in the AI era, leaking it is leaking leverage.
But China was ahead of the curve. Since 2017, it has required all foreign firms operating in China to store data locally and undergo cybersecurity reviews. Cross-border data transfers are tightly restricted. In response, some multinationals have built parallel “data bunkers” in China to maintain compliance.
The result: a splintered internet. Data once flowed freely across borders, but it’s now increasingly fenced in by digital walls. For global companies, navigating this minefield means balancing privacy laws, national security concerns, and market access. For countries, it’s about who trains the best models, who sets the standards — and who stays in control.
The stakes go far beyond privacy.
In an age of foundation models and AI training loops, data is the new oil, and sovereignty is about control of the refinery. Both Washington and Beijing know that the side with the best data will train the best models — and that whoever controls the pipelines controls the future.
Both Washington and Beijing are playing to win.
The U.S. is pressuring allies to block China’s access to advanced chipmaking tools. China, meanwhile, is financing Huawei-built telecom infrastructure in “One Belt, One Road” countries (China’s global infrastructure aimed initiative at connecting Asia, Europe, and beyond), pushing Chinese data standards in international bodies, and expanding RMB-based digital payments.
Caught in the middle are U.S. allies that trade heavily with China but rely on Washington for security. Some are hedging, adjusting policies without picking sides. But in a world where every cloud server and chip fab is a potential battleground, neutrality is becoming harder to maintain.
This next phase of the U.S.-China economic conflict is less about what’s being traded — and more about who gets to write the rules of the digital future.
Because this isn’t just a trade war anymore. It’s a contest for the 21st century.
The post Forget tariffs: These are the next fronts in the trade war appeared first on Quartz.