China's domestic AI chip makers are rapidly closing the gap with Nvidia, with Cambricon reporting $423 million in quarterly revenue and capturing significant market share as US export controls reshape the global semiconductor landscape. Two of China's most prominent GPU designers disclosed first-quarter 2026 earnings that reveal a dramatic shift in the world's second-largest AI market, signaling that Nvidia's once-unassailable position in the region is under serious pressure.

What's Driving China's AI Chip Boom?

Cambricon Technologies posted revenue of 2.89 billion yuan, approximately $423 million, for the quarter, representing a 160% increase compared to the same period last year. The company's net profit surged 185% to 1 billion yuan, marking a remarkable turnaround for a chipmaker that was still losing money as recently as early 2024 and had just lost Huawei as a major customer. For full-year 2025, Cambricon achieved 6.5 billion yuan in revenue with a net profit of 2.06 billion yuan, marking the company's first annual profit since its founding in 2016.

MetaX Integrated Circuits, which went public on the Shanghai Stock Exchange in December, reported first-quarter revenue of 561.9 million yuan, up 75% year over year, though the company still recorded a net loss of 98.8 million yuan. Founded in 2020 by former AMD engineers, MetaX attributed the revenue growth to higher GPU shipments but remains far smaller than Cambricon and has yet to reach profitability.

These gains reflect a broader transformation in China's AI hardware supply chain. Chinese chipmakers delivered 1.65 million AI GPUs in 2025, capturing 41% of the domestic AI server market, according to IDC figures. Huawei led domestic suppliers with approximately 812,000 units shipped, while Cambricon and MetaX are rapidly scaling production to meet surging demand from major technology companies.

How Are US Export Controls Reshaping the Market?

The shift away from Nvidia stems directly from US export restrictions that limit the sale of advanced semiconductors to China. Nvidia's market dominance in China has eroded dramatically, falling from a pre-sanctions peak of 95% to less than 60%. Bernstein analysts project Nvidia's China market share could fall to around 8% in the coming years as local alternatives scale up, though all of these domestic processors remain multiple generations behind Nvidia's current Blackwell architecture in terms of raw performance.

The US government has made strategic decisions about which Chinese companies can purchase advanced chips. Cambricon is one of only two companies on China's government-approved AI hardware procurement list, alongside Huawei; Nvidia, despite easing US restrictions, still is not. This regulatory advantage has given domestic chipmakers a protected market to build scale and improve their technology.

Steps to Understanding China's AI Chip Strategy

• Customer Concentration: Cambricon's top five clients accounted for 94% of the company's revenue in the first half of 2025, with the largest single customer contributing roughly 80%, identified by Chinese media as ByteDance, the parent company of TikTok.
• Production Scaling: Cambricon is targeting 500,000 AI accelerator shipments in 2026, up from an estimated 116,000 units in 2025, with success depending heavily on SMIC's ability to produce chips at its 7-nanometer-class node where yields have been reported at around 20%.
• Market Diversification: Alibaba is expected to become a significant buyer of domestic AI chips as it expands its own AI infrastructure, reducing reliance on any single customer and strengthening the entire ecosystem.

The manufacturing challenges are real. Cambricon's production depends on SMIC, China's leading semiconductor foundry, achieving higher yields at advanced nodes. Current yields of around 20% mean that four out of every five chips produced are discarded, a significant constraint on scaling production to meet the ambitious 500,000-unit target for 2026.

Despite these technical hurdles, the momentum is undeniable. Cambricon's transformation from a loss-making company in early 2024 to a profitable enterprise with triple-digit growth demonstrates how quickly the competitive landscape can shift when regulatory barriers protect domestic alternatives. The company's success also reflects enormous demand from Chinese technology giants like ByteDance, which are building massive AI infrastructure to compete globally and need reliable access to semiconductors without worrying about US export controls.

The broader implication is clear: Nvidia's dominance in China is no longer inevitable. While Nvidia's Blackwell architecture remains technologically superior, the gap is narrowing, and Chinese chipmakers are improving rapidly. For Nvidia, maintaining influence in China will require navigating complex geopolitical tensions and finding ways to work within US export restrictions while preserving its CUDA software ecosystem, which has become the industry standard for AI development. The next few years will determine whether Nvidia can adapt to a fragmented global market or whether it loses its position as the undisputed leader in AI semiconductors.