Top U.S. think tank CSIS's latest report: Four truths and one misconception about Chinese technology...

Author: Top Innovation Zone Blockchain Research Institute

In early March 2026, Scott Kennedy, senior fellow at the Center for Strategic and International Studies (CSIS), released a groundbreaking 147-page report—The Power of Innovation: The Strategic Value of China’s High-Tech Drive.

Scott Kennedy, whose official Chinese name is Gan Side, is a renowned American political scientist and top China expert. His position at CSIS is highly significant: among hundreds of think tanks in Washington, D.C., CSIS is recognized as one of the top-tier, often guiding U.S. government interventions worldwide.

Kennedy’s style is pragmatic and rational, with an in-depth understanding of China’s operational methods. Think tank scholars like him often serve as “unofficial diplomatic envoys” (second-track diplomacy).

Notably, in September 2022—during China’s strict pandemic controls—Kennedy was the first Western think tank scholar to visit mainland China for several weeks of on-the-ground research and face-to-face exchanges with Chinese political and business circles, demonstrating his extensive network and influence on China-U.S. relations.

This influential report explores several key questions:

How does China’s technological innovation translate into geopolitical power? Why do some industries surge ahead while others struggle? As “decoupling” debates fade, what is the future direction of global tech competition?

1. The Underlying Logic of Technological Leapfrogging

Over the past decade, China’s tech policies have shifted from “market-driven technology acquisition” to “import, digest, and innovate,” and now to “independent innovation” with “security first” as a core principle. Especially since 2019, when the U.S. imposed entity list sanctions on Huawei and other Chinese firms, external pressure has instead accelerated China’s technological self-reliance.

Kennedy’s report cites data:

In 2023, China’s R&D expenditure, measured by purchasing power parity, reached $1 trillion—making it the second-largest economy globally, investing over 2.6% of GDP into R&D. During the most subsidized years, various industry funds and policy incentives combined exceeded $250 billion annually—enough to acquire General Electric and still have change.

This “national effort” has yielded clear results:

First, the rise of innovation clusters:

In the 2025 Global Innovation Index (GII), China rose to 10th place, with 24 global top-100 innovation clusters (the Pearl River Delta ranks first worldwide).

Patent numbers look promising: 13.3 per 10,000 people.

But a visit to Yiwu’s small commodity market reveals that some “innovations” are just changing the color of screwdriver handles. Kennedy’s team subtly notes in footnotes: “Significant differences in patent quality exist.”

Numbers don’t tell the whole story.

However, Western think tanks are perceptive enough not to be entirely daunted by these large figures. The report points out that China’s tech ecosystem still faces significant structural weaknesses:

For example, total factor productivity (TFP)—a key indicator of genuine technological progress—is nearly stagnant in China. In other words, despite heavy investment, output efficiency has not improved proportionally. Large subsidies often lead to inefficient resource allocation and severe overcapacity.

Deeper issues include talent gaps: China graduates 4 million science and engineering students annually (a huge engineering dividend), but there are still gaps in frontier breakthroughs and rural/basic education.

And the perennial topic: intellectual property rights.

China’s innovation ecosystem excels at “scaling up” and “engineering iteration”—producing samples that can be replicated in a tenth of the time and at a tenth of the cost, often doing it better.

But when creating entirely new paradigms from scratch, requiring “extreme freedom for trial and error” and “top-tier interdisciplinary talent networks,” institutional inertia becomes a barrier.

Nevertheless, the situation is definitely improving.

2. The Four-Quadrant Framework

The report features a four-quadrant diagram—

which we believe is the most insightful part of the report.

Many people view China’s tech as a monolith—either rising comprehensively or collapsing soon—but this isn’t the real world.

The report proposes an “industry differentiation framework,” dividing China’s tech success or failure into four quadrants based on “domestic ecosystem completeness” and “global market coupling.”

Quadrant 1:

Disruptive Success

In 2024, BYD invested $21.9 billion in R&D, employing 110,000 engineers—more than the entire Detroit auto industry’s engineering workforce.

But money and people aren’t everything. What truly enables BYD to compete globally is China’s “meat grinder” electric vehicle market.

Kennedy’s team found in Shenzhen that a new model takes an average of 18 months from concept to mass production, compared to 36–48 months in Germany. In 2024, over 100 EV brands fought in China’s price war, often earning only a few hundred dollars profit per vehicle.

CATL’s story is similar.

They hold 38% of the global battery market share, not because of government mandates, but driven by market forces—building factories next to lithium mines, locating R&D centers next to automakers, forming near-obsessive vertical integration.

When you can turn raw materials into finished batteries in 24 hours, while competitors need two weeks, the game changes.

“The species that survive are those that evolve, not those that are designed,”

Quadrant 2:

Adaptive Success

In this quadrant, EVs are “overtaking on curves,” while biopharmaceuticals follow a different path—

“Deeply embracing globalization.”

In 2023, China accounted for 39% of global clinical trials, mainly because Chinese hospitals can recruit enough patients within three months, whereas in the U.S., it may take a year. Time equals money and patent duration in pharma.

Hengrui Medicine exemplifies this.

They didn’t try to invent a new anti-cancer mechanism—requiring breakthroughs in basic research—but instead recruited top overseas talent, aligned with FDA standards, and integrated into global innovation networks.

In 2024, China launched about 1,250 new drugs, most of which are “me-too” or “me-better,” not “first-in-class,” but commercially viable.

Quadrant 3:

Disruptive Failure/Obstruction

This is the most lamentable part. The National Big Fund invested hundreds of billions; SMIC and Yangtze Memory Technologies received unprecedented resources.

But by 2026:

China has significant capacity in mature processes (legacy chips, 28nm and above), but in advanced nodes below 7nm, TSMC and Samsung remain far ahead.

Currently, a 3nm EUV lithography machine costs about $200 million per set, produced only by ASML.

The problem: semiconductors are not a money pile; a single EUV machine has 100,000 parts from over 5,000 suppliers worldwide.

ASML’s success isn’t because the Dutch are particularly clever, but because they integrated German optics, American lasers, and Japanese materials.

This is a highly complex system requiring global cooperation among hundreds of “hidden champions.”

Quadrant 4:

Adaptive Failure/Inefficiency

An example here is the C919:

In an industry monopolized by Boeing and Airbus for half a century, protectionism and nationalism alone can’t produce a good aircraft.

In 2024, COMAC delivered only 16 C919s.

In contrast, Boeing delivered 348, and Airbus 735 in the same period.

Moreover, those 16 C919s rely heavily on imported core components: General Electric engines, Honeywell flight control systems, Rockwell Collins avionics—up to 90% import dependence.

(See chart of Boeing 787 Dreamliner component suppliers from Boeing and Reuters; not to scale.)

In reality, policy subsidies can only help get the project off the ground. Building a true moat requires either extreme engineering efficiency—like EVs—or an open attitude toward integrating into the global innovation network—like pharmaceuticals.

Closed-door innovation is a major taboo in tech.

3. Innovation as Power: Turning Technological Potential into Geopolitical Leverage

Technology is never neutral—this may sound obvious, but Kennedy devotes 30 pages to demonstrating it.

He argues that China’s deep strategic value in high-tech lies in its ability to reshape the international power landscape. This power spillover mainly manifests in two core dimensions:

Military-civil fusion (MCF) and soft power through international standards.

1. Military-Civil Fusion

From 2010 to 2024, China invested about $105.8 billion in military-civil fusion.

Where did this money go?

iFlytek’s speech recognition tech used for military intelligence; BeiDou navigation system evolving from civilian to precise-guidance foundation; DJI drones—originally consumer gadgets—now standard reconnaissance and strike tools on modern battlefields.

Commercial tech’s feedback into China’s military is real.

But this feedback is “supplemental,” not “transformative.” Institutional trust barriers and departmental interests limit the seamless transfer of disruptive civilian tech into the military-industrial complex.

Thus, China has gained asymmetric tactical advantages in AI and drones but has yet to fundamentally overturn U.S. military dominance.

Why?

Because of internal trust barriers—no need to elaborate here.

2. Standard-Setting Power

Standards Power

No Monopoly, Strong Influence

“The third-rate companies make products; first-rate companies set standards.”

This popular saying in Chinese business has another layer in tech diplomacy: whoever controls code and protocols controls the game.

By 2025, China participated in 780 ISO technical committees and led 19 working groups in 3GPP. Huawei’s 5G IP accounts for about 20%.

Meanwhile, China leverages its large domestic market (85% standard adoption rate) to influence international standards—like HarmonyOS on 36 million devices, and NearLink tech.

But there’s a delicate balance:

International standards organizations operate on “consensus-driven” principles.

Want to push a standard? You must persuade other member states. Past lessons include WAPI (China’s Wi-Fi standard) and TD-SCDMA (3G standard), which became costly relics due to incompatibility with global ecosystems.

Kennedy notes, “China has strengthened its ‘veto power’ and ‘agenda-setting’ in global tech governance,” but “lacks the ability to unilaterally set the rules.”

The implication:

China can block certain initiatives but cannot unilaterally dictate the game.

4. The Bankruptcy of the Full Decoupling Theory

By 2026, a fascinating phenomenon emerges:

Top international think tanks and policymakers have split into distinct camps, with profound shifts underway.

1. Hawkish/Restrictive Anxiety and Counterproductivity

Represented by some Congress members and early reports from ITIF, they see U.S.-China tech relations as a zero-sum game: China gets stronger, America weakens, so the response must be to shut down all access.

But increasingly, think tanks like RAND and Carnegie argue that broad export controls and “small yard, high wall” policies backfire—

Cutting off supply harms U.S. companies’ revenue (which could fund next-gen R&D) and, more critically, destroys Chinese illusions, pushing China to rapidly develop domestic alternatives (e.g., Huawei’s Mate series comeback).

2. Pragmatists’ Awakening: Managing Interdependence

This is the core of the CSIS report and a new consensus among mainstream think tanks: “Full decoupling” is prohibitively costly and unrealistic.

What happens if global supply chains are forcibly severed?

Severe inflation in the West—due to inability to buy cheap Chinese-made goods;

Delayed global green energy transition—since China produces 80% of solar panels and 60% of wind turbines;

Loss of insight into China’s technological evolution—if you stop doing business with your rival, you won’t know how far they’ve advanced.

3. The Third Voice of the Global South

A report by the Atlantic Council keenly points out that in many developing countries in Asia, Africa, and Latin America, China’s 5G networks, affordable EVs, and AI infrastructure represent “affordable development opportunities,” not “national security threats.”

If the West only promotes “security anxieties” without offering competitively priced alternatives, its narrative in the Global South will be utterly discredited.

5. The Way Forward: Precise Coupling

If “full decoupling” is poison and “unconditional embrace” is wishful thinking, where is the solution?

CSIS’s answer, from an American perspective:

“Calibrated Coupling.”

Internally: Strengthen the domestic innovation ecosystem (economic perspective).

America’s true strength isn’t in suppressing Chinese companies but in its unmatched “lighthouse effect”—attracting the world’s brightest minds, a deep venture capital network, and robust basic science research.

They suggest U.S. government subsidies should be precisely targeted at strategic nodes like semiconductors, rather than broad trade protectionism.

Externally: Establish “surgical” barriers (realist perspective).

Abandon blanket bans, instead tightly restrict key chokepoint technologies with direct military applications, while maintaining normal commercial and academic exchanges in consumer electronics, mature chips, and open-source AI models.

In standard-setting, Western governments shouldn’t withdraw from international standards organizations out of fear of China’s influence. Instead, they should participate more actively, shaping rules through alliances and consensus to foster an open system.

In international cooperation—climate change, AI ethics, global health (medical trials)—deep interdependence and collaboration can generate huge economic benefits and serve as “shock absorbers” to prevent major power conflicts from escalating into hot war.

6. Returning Technology to Human Welfare

This CSIS report, along with the intensive voices from major think tanks in 2026, sends a clear message:

China’s high-tech strategy is multifaceted.

It has gained momentum in certain sectors—electric vehicles, batteries, 5G, biotech—that can reshape global industry patterns; but in fundamental ecosystems—advanced semiconductors, aircraft engines, top-tier basic research—it still faces long-term, structural challenges.

The future global tech landscape will be a highly complex “compound competition and cooperation.”

Whoever can attract global talent with an open mindset,

whoever can create an inclusive ecosystem to spread technology to developing countries,

whoever maintains restraint and rationality amid competition, and upholds pragmatism and openness,

will truly win the next decade.