China’s largest contract chipmaker, Semiconductor Manufacturing International Corporation (SMIC), expects the global shortage of artificial intelligence computing capacity and supporting semiconductors to continue into 2028. The company believes expanding data-centre construction and rising demand from businesses building their own AI infrastructure will keep pressure on chip supplies for at least another two years.

SMIC co-CEO Zhao Haijun shared the outlook during the company’s third-quarter earnings call on November 13, 2026. He warned that computing power and supporting chips could remain in short supply through the middle or second half of 2028. The forecast comes as Chinese technology companies face uncertainty over access to advanced Nvidia processors amid the possibility of tighter US export restrictions, potentially increasing demand for domestically manufactured chips.

SMIC Warns AI Computing Shortage Could Last Until 2028

Zhao said the AI industry was still in its early stages, making future demand difficult to predict. However, the company is seeing strong requirements for computing resources from customers across different industries.

He expects shortages of computing power and supporting chips to persist until at least the middle or second half of 2028.

The demand is coming from more than large technology companies operating massive cloud data centres. Businesses across different sectors are increasingly building smaller, private computing facilities to support their own AI applications.

According to Zhao, companies want greater control over computing resources rather than depending entirely on rented AI processing capacity. Renting computing power can involve fluctuating prices and availability, creating uncertainty for businesses that need reliable infrastructure.

This shift is expanding the market for semiconductor manufacturers such as SMIC, which supply chips used in a wide range of electronic and computing applications.

The forecast also suggests that increasing semiconductor manufacturing capacity may not immediately resolve the shortage because demand for AI infrastructure continues to grow.

Data-Centre Expansion Drives Demand for AI Chips

The rapid expansion of AI data centres is one of the main reasons behind the continuing shortage.

Technology companies are investing heavily in specialised processors, memory chips, networking equipment and other components required to train and operate advanced AI models.

These facilities require large numbers of processors working together to handle complex workloads. As AI applications become more widely adopted, companies need additional computing capacity to support model training, inference and enterprise automation.

The demand is also spreading beyond traditional cloud providers.

Banks, manufacturers, telecommunications companies and other businesses are exploring private AI infrastructure to process sensitive information and run internal applications.

This creates additional demand for semiconductors even when companies are not building large public cloud platforms.

SMIC believes the expansion of enterprise-level computing infrastructure will continue supporting semiconductor demand over the coming years.

Nvidia Export Restrictions Could Increase Pressure on China’s Chip Industry

Another factor influencing SMIC’s outlook is uncertainty surrounding China’s access to advanced American semiconductor technology.

Nvidia has traditionally been a major supplier of AI processors to technology companies worldwide. Its GPUs are widely used for training large AI models and operating data-centre workloads.

However, US export controls have restricted China’s access to several advanced Nvidia processors.

Further restrictions could make it more difficult for Chinese companies to obtain high-performance computing hardware from American suppliers.

This could accelerate investment in domestic alternatives and increase demand for chips manufactured by Chinese companies.

SMIC is particularly important to China’s semiconductor strategy because it is the country’s largest contract chipmaker. Its manufacturing capabilities support domestic technology companies seeking to reduce dependence on overseas semiconductor supply chains.

Nevertheless, domestic production does not automatically replace advanced Nvidia GPUs. Manufacturing high-performance processors requires sophisticated chip designs, advanced fabrication processes, packaging technologies and supporting software ecosystems.

The ability of Chinese companies to close these gaps will influence how quickly the country can develop a more self-reliant AI computing industry.

SMIC Reports 82% Surge in Third-Quarter Gross Profit

The company’s positive demand outlook comes alongside strong financial results for the third quarter of 2026.

SMIC reported an 82.3% year-on-year increase in gross profit, reaching US$953.2 million compared with US$522.8 million in the same period a year earlier.

Revenue also increased, supported by higher wafer sales volumes, improved factory utilisation and a more favourable product mix.

Financial metricQ3 2026Year-on-year change
RevenueUS$3.19 billion+33.9%
Gross profitUS$953.2 million+82.3%
Gross margin29.9%+7.9 percentage points

The improvement in gross margin indicates that SMIC benefited from stronger manufacturing utilisation and changes in the mix of products it produced.

The company expects revenue to increase by another 4% to 6% sequentially in the fourth quarter. However, it anticipates gross margin could moderate to between 26% and 28%.

This suggests that demand remains strong, although manufacturing costs, pricing and product mix will continue to influence profitability.

Semiconductor Equipment Delays Could Limit Production Expansion

Despite strong demand, SMIC faces challenges in expanding its manufacturing capacity quickly enough.

Zhao said some semiconductor manufacturing equipment orders now have delivery lead times exceeding 24 months. These delays could affect the company’s capital expenditure plans and the pace at which additional production capacity becomes available.

Semiconductor fabrication facilities depend on highly specialised equipment for processes such as wafer fabrication, deposition, etching and inspection.

Expanding production therefore requires more than investment in buildings and manufacturing facilities. Companies must also secure advanced equipment, qualified suppliers and technical expertise.

Longer delivery schedules can delay the installation of new production lines and limit how quickly manufacturers respond to rising demand.

SMIC also expects its depreciation expenses to increase by approximately 30% year-on-year in 2026, reflecting its investments in manufacturing capacity.

The company must balance the opportunity created by strong demand with the financial burden of expanding its factories.

Memory Chip Supply Could Remain Tight Through 2028

The shortage extends beyond processors used for AI computing.

Memory chips are also becoming increasingly important as AI data centres require large quantities of high-bandwidth memory and other storage technologies to support demanding workloads.

According to Reuters, Zhao expects memory chip supply and demand to move towards balance around the middle to the end of 2028.

The distinction is important because different semiconductor categories face different supply constraints. Expanding processor manufacturing alone will not solve shortages involving memory, networking components and other supporting hardware.

AI systems depend on these components working together. A shortage in one part of the supply chain can limit the deployment of an entire computing system.

The continuing investment in data centres is therefore creating demand across multiple segments of the semiconductor industry.

China’s Domestic Chipmakers Could Benefit From Rising AI Demand

SMIC’s outlook reflects China’s wider effort to develop domestic semiconductor manufacturing capabilities.

Companies including Huawei and other Chinese technology developers are investing in alternatives to foreign AI hardware.

Domestic chipmakers could benefit if US export restrictions become tighter or if Chinese customers increasingly prioritise locally manufactured components.

However, China’s semiconductor industry still faces technological and manufacturing challenges, particularly in producing the most advanced processors at scale.

SMIC’s role is significant because it provides manufacturing services to companies that design their own chips but do not operate fabrication facilities.

As domestic demand increases, the company could attract more orders from businesses seeking alternatives to overseas manufacturing suppliers.

Still, its ability to capture the full opportunity will depend on production yields, equipment availability, manufacturing costs and the performance of the chips being designed by its customers.

The Bigger Picture

SMIC’s warning highlights how the AI infrastructure boom is creating a prolonged supply challenge for the semiconductor industry. Demand is no longer coming only from a handful of major cloud companies. Businesses across different sectors are investing in private computing infrastructure, expanding the market for processors, memory and supporting components.

Geopolitical restrictions are adding another dimension. China’s limited access to certain advanced American processors could accelerate domestic semiconductor investment, but building competitive alternatives requires years of research, manufacturing investment and improvements across the technology supply chain.

The shortage also demonstrates that increasing AI computing capacity is not simply a matter of buying more GPUs. Manufacturers need access to memory, networking equipment, fabrication tools and advanced packaging technologies. Delays in any of these areas can slow the expansion of data centres.

Looking Ahead

SMIC expects demand for AI computing power and supporting chips to remain strong through 2028, with enterprise infrastructure investment providing additional support. Its third-quarter financial performance indicates that higher wafer sales and improved factory utilisation are already benefiting the business. However, long equipment delivery times and rising depreciation expenses could complicate efforts to expand production quickly.

The next two years will be important for China’s semiconductor industry as domestic manufacturers attempt to increase capacity and reduce dependence on foreign chip suppliers. Whether companies such as SMIC can meet the growing requirements of AI developers will depend on manufacturing expansion, technological progress and access to critical production equipment. For now, the company’s outlook suggests that the global AI chip supply challenge is unlikely to disappear in the near term.

Get the day’s top stories in your inbox

One concise email. No spam, unsubscribe anytime.