US-China Chip War: Who Will Dominate the Future of AI?

Four years ago, the United States (US) intensified pressure on China’s technological ambitions by imposing export restrictions on advanced chips, known as semiconductors, used in artificial intelligence (AI), data centres, and national defence.

The Joe Biden administration aims to limit Beijing’s ability to develop technologies that could bolster its military and financial power, while narrowing the gap between the world’s two largest economies.

These restrictions have spurred Beijing to accelerate efforts towards chip self-sufficiency, a goal previously outlined in the Made in China 2025 plan. Since then, the Chinese government has poured hundreds of billions of dollars into building domestic semiconductor production.

Chips as a national security issue

Beijing has provided substantial subsidies, tax breaks, and various cost incentives to nurture local competitors to NVIDIA, the US company behind the cutting-edge Blackwell AI chips, as well as TSMC, the world’s largest contract chip manufacturer for advanced semiconductors and developer of N2 chip manufacturing technology.

SMIC, the backbone of China’s self-sufficiency plan, recorded record revenue of $9.3 billion (approximately Rp148.8 trillion) last year, while HuaHong, the second-largest foundry or third-party chip factory in mainland China’s semiconductor supply chain, operated at 106% capacity due to high demand, according to its Q4 2025 financial report.

However, despite China’s strenuous efforts to catch up with major US technology firms, Ryu Yongwook, an assistant professor at the National University of Singapore, assesses that such progress is often exaggerated.

“Beijing wants to achieve chip self-sufficiency, but the current level is still far from that,” he said.

China lags behind the US in research, design, and innovation, and also trails Taiwan and South Korea in production, according to Ryu.

Chinese chip makers rise in the value chain

Nevertheless, China has achieved significant breakthroughs in recent years. According to the Rhodium Group, the country now controls about 30% of the global market share for legacy chips, components essential for vehicles, industrial equipment, and consumer electronics.

These chips are not the fastest or most advanced, but they are produced on a large scale by Chinese companies, raising concerns among global competitors.

“China’s production expansion will pressure global chip prices and put stress on non-Chinese vendors,” said John Lee from East-West Futures.

This phenomenon is already evident in some sectors, such as silicon carbide wafers, a key material for high-power chips.

Breakthroughs in advanced chips

China has also made progress in more advanced chips, successfully producing 7-nanometre class processors now used in Huawei’s latest smartphones.

These chips are comparable to products released by TSMC in 2018 for Western clients, but still lag behind 3 nm and 5 nm chips in terms of speed, power efficiency, and production costs.

Tim Rhlig from the European Union Institute for Security Studies describes China’s chip ambitions as facing a “technology wall” and US sanctions.

“There are limits to how far you can advance without access to the most advanced US chipsets,” he said, adding that China may need “about a decade” to catch up.

Reflecting a shift in priorities, the latest Five-Year Plan of the Chinese Communist Party no longer emphasises chip dominance as strongly. The document instead highlights AI and introduces a “model-chip-cloud-application” framework, positioning chips as part of a broader computing ecosystem.

China’s Plan B sparks new rivalries

China is now focusing more on practical, task-based artificial intelligence or AI for industries that require lower computing power, something that can be handled by domestic chips.

Even if not at the forefront of technology, Chinese chip and AI systems offer strong performance at much lower costs. This is driving rapid adoption in the Global South, where governments and companies increasingly opt for Chinese solutions over Western ones.

Market intelligence firm TrendForce notes that Chinese AI platforms like DeepSeek and Alibaba’s Qwen have captured about 15% of the global AI model market by the end of 2025.

This poses a long-term threat to the global dominance of US technology giants like Microsoft and Google, which are projected to spend a record $700 billion (approximately Rp11.2 quadrillion) this year on AI infrastructure according to Goldman Sachs.

US advantages face real challenges

There are other hurdles for Silicon Valley’s ambitions to create AI that surpasses human intelligence. In January, market intelligence provider ICIS warned that US data centres, reliant on top-tier chips, could soon be constrained by a strained power grid.

In contrast, China’s rapidly expanding energy sector provides an additional advantage. With projected electricity reserve capacity reaching 400 gigawatts by 2030, China can build large-scale data centres even if its chips are less efficient.

“Cheap energy is a very important factor, not just for chips but also for AI and other advanced technologies,” said Ryu Yongwook.

ICIS sees three possible outcomes in the chip competition:

The US maintains its lead by improving its power grid. The US continues to dominate AI research with advanced chips, while Chinese AI systems spread in the Global South. If geopolitical tensions escalate, two separate AI ecosystems could emerge.

Although the finish line is still distant, the chip industry faces a future where Chinese competitors not only offer lower prices but are also rapidly closing the gap in sophistication and product reliability.