China rewards companies that abandon Nvidia with 50% energy discounts

Alberto Noriega     9 November 2025     5 min.
China rewards companies that abandon Nvidia with 50% energy discounts

The measure seeks to reduce dependence on US semiconductors following new export restrictions imposed by the US.

The Chinese government has launched its most aggressive measure to date to strengthen its semiconductor industry: energy subsidies of 50% for data centers that use artificial intelligence processors manufactured in ChinaThe plan, confirmed by official and regional sources, aims to decouple the national technology ecosystem from US chips, in particular from Nvidiawhose products remain under severe export restrictions.

Local governments of Gansu, Guizhou and Inner Mongolia—areas where the main cloud computing hubs are concentrated—have extended these discounts to Alibaba, ByteDance and Tencentthree of the country's largest technology conglomerates. Companies that adopt chips from Huawei o Cambricon they will be able to benefit from a half off your electricity bills, one of the highest costs of operating data centers.

According to sources consulted by Times of India, some subsidies are so extensive that They offset the energy expenditure of an entire year., which represents a indirect multimillion-dollar injection in favor of companies that support the technological self-sufficiency strategy promoted by Beijing.

The centers that still use Nvidia chips They are expressly excluded from the benefit. The measure responds to the ban imposed in September by the Chinese internet regulator, which prohibited the purchase of foreign AI processors after months of tensions with Washington.

From US sanctions to the rise of technological nationalism

Beijing's move coincides with a tightening of US restrictions. President Donald TrumpIn statements made on November 2, he affirmed that the Nvidia Blackwell AI chips—the most advanced on the market—will be “exclusive for US use“We can’t sell them to China. We can’t sell them to other people.”

Pexels Luis Gomes 166706 546819

China's new energy policy acts as direct strategic response to this measure. For months, the country's major technology companies have been forced to Replace the Nvidia A100 and H100 chips, which previously powered their generative AI models, by less efficient domestic alternativesThis change has led to an increase in electricity costsbecause domestic chips require more energy to match the same processing levels.

The discounts are looking for to compensate for that performance gap y to maintain the competitiveness of the national sector while the local semiconductor industry accelerates its development.Reducing operating costs is key to preventing energy efficiency from hindering the advancement of Chinese AI.,” stated an analyst from Dialogue Earth.

The incentive also has a political dimension: reaffirms the Communist Party's commitment to technological independence and sends a message to multinational corporations that loyalty to the Chinese supply chain will be rewarded with tangible financial benefits.

Chinese chipmakers gain traction

The state's support comes at a time of Explosive growth for domestic semiconductor manufacturers. The company Cambricon, one of the leading providers of AI processors, reported in the first half of 2025 a 44-fold increase in revenue, reaching 2.900 million yuan, driven by strong demand for its chip line Siyuan 590.

For its part, Huawei It continues to consolidate its role as a pillar of the Chinese technology ecosystem. According to estimates from JPMorganThe company will ship between 600.000 and 650.000 AI chips This year, marking its return to leadership in advanced hardware despite international sanctions.

Even so, the gap with Nvidia remains significant. Domestic chips currently reach around 80% of the yield of the A100 from Nvidia, according to industry experts. But government incentives are looking to to make the adoption of domestic hardware economically viablewhile continuing to invest in R&D to improve its efficiency. “The goal is not to match tomorrow’s performance, but create a resilient industrial ecosystem"where Chinese companies can innovate without depending on foreign components," the economist explained. Zhou Zhiwen from the Shenzhen Institute of Science and Technology.

Pexels Brett Sayles 2881229

State support also translates into a higher guaranteed domestic demandThis allows manufacturers to reinvest in research and large-scale production. With more than 20 new semiconductor industrial parks planned for 2026, China seeks to close the complete cycle of design, manufacture and consumption within their own borders.

National AI, cheap energy, and a horizon of technological self-sufficiency

The energy subsidy package is only a piece within a broader national strategy known as "AI Made in China 2030Under this plan, the government intends that by the end of the decade Over 70% of the country's AI infrastructure is based on domestic hardware..

The 50% reductions in electricity bills not only do they alleviate the economic burden on companies, but they also They are redirecting capital investments towards software research, neural networks, and foundational models.In this way, Beijing is trying to ensure that Future technological revolutions should not depend on access to foreign chips.especially in sensitive sectors such as defense, energy, or health.

However, international analysts warn that this policy may deepen global technological fragmentation, creating parallel blocks of innovation led by the United States and China.We are witnessing a race for computational sovereignty"Says Sarah O'Neill, an expert in technology policy of Council on Foreign Relations. "And that race isn't measured only in teraflops, but in geopolitical control over who has access to computing power.".

Meanwhile, Chinese giants like Alibaba, Tencent and ByteDance They seem ready to take advantage of every opportunity. If the combination of domestic chips and subsidized energy manages to sustain its performance, China could rebalance the global artificial intelligence landscape, making it clear that The future of computing is not only built with silicon, but also with energy and industrial policy..

Comments closed