A Dual Disruption and the Constant of Surprises: Trump 2.0 and DeepSeek

Photo: BRENDAN SMIALOWSKI/AFP/Getty Images
The biggest surprises are not in what happens, but in what we thought couldn’t happen.
If we ever needed a reminder that surprises are a constant, the dual disruptions between Trump 2.0 actions and DeepSeek’s AI breakthrough over the last fortnight have not disappointed. Whatever we had priced in prior to these events, many of us—technocrats, technologists, and investors alike—were left rethinking assumptions.
The tsunami of executive orders, followed by planned International Emergency Economic Powers Act (IEEPA) tariffs, deportations, and actions from the Department of Government Efficiency (DOGE)—surprising in scope and speed—may have far-reaching implications for the U.S. economy, government, and international relations. For now, the United States in the midst of a thick policy fog—and it is hard to discern how things will play out.
A Looming Trade War?
The “Day 1” executive order on an America First Trade Policy teased a deliberate approach to studying the root causes of the merchandise trade deficits, the effectiveness of export controls, and the like. By February 1, the president had announced his use of the IEEPA to place 25 percent tariffs on products from Mexico and Canada (10 percent on Canadian “energy resources”) and 10 percent on all products from China. Originally scheduled to go into effect on February 4, the administration has hit pause on the Mexico and Canada tariffs for a month. The ostensible goal of the planned tariffs is to achieve noneconomic, national security ends—an action to address the fentanyl crisis.
Whether or not tariffs on the United States will be eventually implemented, observers are assessing the likely impacts of these import taxes on different categories of goods (e.g., food, energy, and electronics) and different geographies such as border and heartland states with meaningful exposure. Before the pause in implementation, a growing number of industry associations, business leaders, and investors urged the administration to find an off-ramp to avoid a trade war.
Public Sector Reform
On the public management front, the administration paused disbursements of Biden-era infrastructure and energy transition programs as well as other federal assistance programs, only to be temporarily blocked by a federal judge and rescinded by President Trump. The new U.S. DOGE service is implementing a fast-moving, still somewhat opaque, reform agenda—equal parts cost-cutting, return-to-work, and AI-enablement actions. According to press reports and Elon Musk’s own X feed, DOGE is accessing Department of the Treasury payment systems and reportedly halting payments to federal contractors. It has closed the offices of the United States Agency for International Development (USAID) and placed several senior USAID officials on leave. The agency will either be placed under the Department of State or according to Musk, “shut down.” It is still too early to tell what DOGE efforts will be on the U.S. economy and the economies of other countries, or whether spending freezes and cuts will improve government effectiveness.
Unleashing Energy and Technology
Presidential actions also included the declaration of an energy emergency, sought to unleash U.S. energy through regulatory reforms, and reversed Biden-era climate policies. The president also turned the page on AI and digital assets, stepping away from the Biden-era focus on safety and regulation. This is part of a Trump-era technology policy that included a new charter for the President’s Council of Advisors on Science and Technology and the announcement of an eye-popping $500 billion Stargate Project, which aims to leverage the combined firepower of U.S., Asian, and Middle Eastern investors in the AI infrastructure for OpenAI in the United States.
Stargate is meant to signal U.S. leadership on “the built data center infrastructure landscape, from power and land to construction to equipment, and everything in between.” Notwithstanding some skepticism about whether the full scale of these investments will materialize, the project stands out as a bold private sector alternative to Biden-era public investments to build the U.S. technological and industrial base.
An AI Surprise
And yet, true to Nassim Nicholas Taleb’s admonitions about surprises, this apparent triumph of proprietary AI was upended by the curiously-timed release of DeepSeek V3, China’s low-cost, open-source disruptor—what the economist Oliver Blanchard called “probably the largest positive tfp [total factor productivity] shock in the history of the world.”
There are many questions about the DeepSeek event. Does its $6 million price tag include the full costs of the development of its V3 (including failures along the way)? Will its claims of more efficient training models without access to the most advanced NVIDIA chips stand up to scrutiny? Does it pose a fundamental challenge to the prevailing U.S. model for AI development? Is this a Jevons paradox moment along the way to a further exponential AI demand, reinforcing the logic of the Stargate Project?
As the technology fog lifts, one thing is already clear: Chinese tech has thrown down a challenge in a domain where the United States has consistently led—AI training processes. For context, before the announcement of DeepSeek V3, China was gradually narrowing the United States’ massive lead on AI training, as measured by the maximum training compute achieved by models.
But the DeepSeek-V3 disruption represents a delta not in complexity but in enhanced efficiency—more for less. Its efficiency gains suggest generating positive externalities for the AI industry as a whole, spurring a rethink on the part of OpenAI and other U.S. tech leaders worried about being on the “wrong side of history.”
Making Sense—Four Tests
With so much change on so many fronts, prognosticators of policy and markets will likely stay busy. Indeed, there will be ample opportunities in the coming months and potentially years for those with an edge in predicting when the next shoe will drop. Equally important is understanding, as best as we can, how policy and technology disruptions will flow through to the economy and society, ours and others. Specifically, in an era of heightened geopolitical and technology competition, how will it impact U.S. economic security—that is, U.S. prospects for broad-based, innovation-led growth while ensuring the security of markets, supply chains, assets, and people amid global threats?
To this end, it is worth applying four tests to economic policies and technology strategies:
- First, there is the innovation ecosystems test. Do policies and strategies limit exposure of the U.S. technology innovation ecosystem to mercantilist threats (for instance, dumping of subsidized electronic vehicles or chips by China) or malign threats (for example, cyberattacks that threaten U.S. infrastructure or fentanyl trafficking that threaten American people)? By the same token, do they ensure that the U.S. ecosystem provides strong incentives for U.S. innovators to out-compete and out-innovative others including in fiercely competitive technology markets?
- Second, there is a technology opportunity test. Nobel Laureates Daron Acemoglu and Simon Johnson, in their thousand-year sweep of technological progress, Power and Progress, remind us that “there is nothing automatic about new technologies bringing widespread prosperity. Whether they do or not is an economic, social, and political choice.” Do policies and strategies ensure that technology innovation brings broad-based growth? In the U.S. context, this means geographically dispersed growth, including what both presidents, Biden and Trump, have called “people and places left behind.”
- Third, the technology alliance test. Even as the United States ensures the security of its universities and markets from malign actors, the reality is that research and development, innovation, and commercialization take place in partnership with researchers and entrepreneurs across borders. No country has a monopoly on path-breaking ideas and technology ecosystems—to be on the frontier—cannot be autarkic. At the same time, not all partners are the same. Do policies and strategies strengthen technology alliances with trust partners—across government, business, and universities?
- Fourth and finally is the economic governance test. Investors, technology entrepreneurs, and internationally competitive STEM students react to incentives by government policies and behaviors. Policy reforms are an essential ingredient of economic evolution and growth. And yet arbitrary actions by the state, we know all too well, can undermine the confidence of those whose talents, resources, and know-how are required to win the global tech race. Do policies, strategies, and government behaviors strengthen or weaken the enabling environment for rapid technological innovation, investment, and growth?
It is not always clear how a particular set of policies or strategies will net out when applying these tests (or other tests that we can apply). But policymakers and technology strategies should consider applying them dispassionately and consistently—especially in the midst of the inevitable surprises—to ensure that the United States and its allies have a fighting chance in the global economic and technology race. Our prospects for growth with security will depend on it.
Navin Girishankar is the president of the Economic Security and Technology Department at the Center for Strategic and International Studies (CSIS) in Washington, D.C.
Research associate Andrea Leonard Palazzi and research intern Ahmed Shareef, both at CSIS, assisted with data analysis and visualization.