Dec 10, 2025, 1:51 PM ETInterview, Webinar

The US-China Tech Race

The technological competition is characterized as a strategic rivalry with four distinct arenas: leading technology creation, applying technology, installing digital infrastructure, and achieving self-sufficiency.
The United States currently leads in fundamental technology domains, including semiconductors, AI frameworks, cloud infrastructure, quantum computing, and global talent acquisition.
China holds advantages in application and installation, evidenced by using 12 times as many robots in manufacturing per income-adjusted employee compared to the US.
China dominates global digital infrastructure installation outside of US allied nations, particularly through Huawei's dominance in the "global south."
China is aggressively pursuing self-sufficiency via "dual circulation" strategies, including edicts restricting companies from buying Western chips and offering energy discounts for training models without Western hardware.

Advanced lithography (specifically Extreme Ultraviolet or EUV) remains a primary choke point, controlled almost exclusively by ASML, with feature sizes shrinking to 2 nanometers.
TSMC acts as a critical manufacturing choke point, producing over 90% of the world's most advanced semiconductor nodes.
US export controls since 2018 have prevented Chinese companies from purchasing EUV equipment, forcing China to develop parallel technologies.
Paul Triolo notes that while US controls have cut off advanced GPU access, this has a collateral effect of accelerating Chinese innovation in alternative chip design and manufacturing.
China faces significant hurdles in lithography due to the complexity of systems integration, requiring specialized software, optics, photo resists, and a highly seasoned workforce for maintenance.
US foundry reliance is concentrated on Intel (mature) and TSMC, with TSMC's primary investment focus remaining in Taiwan rather than the US.
China lacks the "patient capital" and mature capital markets that the US leveraged to develop its semiconductor industry in the 1960s and 70s.

Leading open-source and open-weight AI models (e.g., DeepSeek, Moonshot, Qwen) are now Chinese, whereas leading closed proprietary models (e.g., OpenAI, Anthropic) remain US-based.
The US advantage in chip scaling is contested; the effectiveness of the race depends on whether more compute power directly translates to superior model capabilities.
Chinese government policy actively supports AI deployment by building national data centers and providing power infrastructure to lower barriers for smaller firms.
The US faces energy constraints for AI data center growth, including supply chain bottlenecks for transformers, turbines, and nuclear components, and a lack of a unified grid.
The proliferation of cheaper, open-source Chinese models is driving rapid global adoption, potentially offsetting US advantages in proprietary model development.
Energy availability is identified as a larger gating function for AI than rare earth minerals, which are critical for defense and auto but less so for data center build-out.

China maintains leverage through rare earth mineral dominance, having issued a one-year reprieve on restrictive controls while the US attempts to secure alternative supply chains.
US allies (Australia, Canada) are deemed vital for securing non-Chinese rare earth supplies, though current US investment in these sectors remains reactive.
Trust is a contested asset: China's state-controlled cyber laws and interoperability issues reduce trust among international partners, whereas the US faces trust issues due to inconsistent reliability (e.g., cutting satellite access for Ukraine).
Paul Triolo suggests rare earth restrictions are a "constant irritation" rather than a decisive blocker for AI infrastructure compared to energy supply constraints.
The US government is facing criticism for reducing research university funding and making immigration harder for global talent, undermining the "best technology" pillar.
China is leveraging its "market creation" strategy to lock the Global South into its tech ecosystem, accepting sacrifices in near-term performance to secure long-term customer bases.

A potential outcome scenario exists where the US owns the "blueprints" (code/standards) while China owns the "buildings" (physical installations and hardware).
China's massive output of engineering and PhD graduates provides a quantitative scale advantage that may translate into quality advantages in application and installation sectors.
The US faces a risk of strategic dependency on China for certain supply chains and market access even if it wins the initial technology race.
Future US leadership depends on resolving energy grid instability, securing stable mineral supply through government-backed offtake agreements, and competing on reliability standards.
If China successfully develops a competitive domestic supply chain for advanced GPUs and AI models, the effectiveness of current US export controls will be reassessed.