ByteDance, Alibaba and Tencent have placed orders for more than 3 million Nvidia RTX PRO 5500 graphics cards, with a combined list value exceeding $35 billion, according to a Reuters report published on October 20, 2026. The orders highlight the growing demand for computing hardware among China’s largest technology companies as they expand artificial intelligence services, develop AI models and build infrastructure for inference workloads. However, Nvidia has reportedly warned customers that it may be unable to fulfil most of the requested quantities because demand is far greater than planned production.
The companies are turning to the RTX PRO 5500 because US export restrictions have limited Chinese buyers’ access to more advanced Nvidia AI processors. The graphics card is designed for professional workstations rather than large data-centre clusters, but it is among the more capable Nvidia products still available to Chinese customers under the current restrictions. The orders do not represent $35 billion in completed sales or guaranteed deliveries. Instead, they demonstrate the scale of China’s demand for Nvidia hardware and the difficulty of meeting that demand under export controls and supply constraints.
Key Takeaways
- Order value: ByteDance, Alibaba and Tencent have placed orders with a combined list value exceeding $35 billion.
- Hardware: The orders cover more than 3 million Nvidia RTX PRO 5500 graphics cards.
- Price per unit: The cards are reportedly priced at approximately $11,400 each.
- Supply constraint: Nvidia may be able to fulfil only 5%–10% of the requested quantities, according to sources cited by Reuters.
- AI demand: Chinese technology companies are seeking additional computing capacity, particularly for AI inference.
- Export restrictions: US controls have limited access to Nvidia’s more advanced data-centre chips in China.
- Important distinction: The reported order value is not confirmed revenue for Nvidia, and delivery quantities remain uncertain.
Why Are China’s Technology Giants Ordering Nvidia GPUs?
China’s largest technology companies are investing heavily in AI models, cloud computing, digital assistants and enterprise automation. These services require substantial computing resources, particularly as more applications move from experimental development into everyday use.
ByteDance operates platforms such as Douyin and TikTok and develops AI models and services. Alibaba has expanded its cloud computing business and its Qwen family of AI models, while Tencent is investing in AI applications across its digital ecosystem, including messaging, gaming and enterprise services.
All three companies need computing capacity to train models, operate AI applications and respond to user requests. Training generally involves processing large datasets to develop a model, while inference is the process of using an already-trained model to generate answers, produce content or perform tasks.
The two activities have different computing requirements. Training frontier-scale models often depends on large, interconnected clusters of specialised accelerators. Inference workloads can use a wider range of hardware configurations, depending on model size, memory requirements, performance targets and software optimisation.
That distinction helps explain why the RTX PRO 5500 has attracted attention. Although it is not a direct substitute for Nvidia’s most powerful data-centre processors in every workload, it can provide useful computing capacity for certain AI applications.
What Is the Nvidia RTX PRO 5500?
The RTX PRO 5500 is a professional graphics processing unit designed for workstation workloads. Professional GPUs are commonly used in areas such as visualisation, computer-aided design, simulation, graphics production and other demanding computing tasks.
Unlike Nvidia’s specialised data-centre accelerators, which are designed for large-scale AI infrastructure, workstation GPUs are intended for powerful individual computers and professional systems.
Nevertheless, graphics processors are highly parallel computing devices, making them useful for some AI workloads. Depending on the software, memory configuration and performance requirements, workstation GPUs can support model inference, development and other computational tasks.
The RTX PRO 5500 has become especially relevant to Chinese buyers because access to more advanced Nvidia hardware is restricted.
| Feature | RTX PRO 5500 order story |
|---|---|
| Product category | Professional workstation GPU |
| Reported unit price | Approximately $11,400 |
| Combined reported orders | More than 3 million cards |
| Combined list value | More than $35 billion |
| Main reported use case | AI computing, particularly inference |
| Major constraint | Export rules and limited production capacity |
The reported price and quantities come from people familiar with the orders cited by Reuters. They should not be interpreted as a confirmed final transaction value or a guarantee that every ordered card will be supplied.
How US Export Restrictions Are Reshaping China’s AI Hardware Market
The orders illustrate how semiconductor export controls can influence the products available to technology companies in China.
The United States has imposed restrictions on exports of certain advanced computing chips to China, citing national security concerns. The rules have limited Chinese companies’ access to some of Nvidia’s most powerful AI hardware.
Nvidia’s Blackwell-based B300 data-centre chip is not permitted for sale to Chinese buyers under the restrictions described in the Reuters report. Exports of the Hopper-based H200 are subject to licensing conditions and case-by-case review.
These restrictions create a different hardware market for Chinese customers. Companies that might otherwise seek high-performance data-centre accelerators must consider which products they can legally acquire, what performance those products offer and how much supply is available.
The RTX PRO 5500 is therefore attracting demand not simply because it is a professional GPU, but because it offers a potential route to additional Nvidia computing capacity within the existing regulatory environment.
Export restrictions do not necessarily prevent every Nvidia product from being sold in China. The applicable rules depend on the product, technical specifications, destination and licensing requirements. Buyers and suppliers must evaluate those conditions rather than assume that all Nvidia hardware is either permitted or prohibited.
The situation also creates uncertainty for Nvidia. Even when Chinese companies want to buy a particular product, sales depend on regulatory compliance, production capacity and the ability to deliver the requested quantities.
Nvidia May Be Unable to Meet Most of the Demand
The scale of the orders is only part of the story. Supply constraints could determine how much revenue Nvidia ultimately earns from the demand.
Reuters reported that Nvidia had told customers and partners it could not guarantee fulfilment because demand from Chinese technology companies exceeded planned production. One source said customers might receive only 5%–10% of the quantities requested.
Another person familiar with the matter said Nvidia had indicated it could produce no more than 3 million RTX PRO 5500 cards globally in 2027. Orders from Chinese customers alone were reportedly above 6 million units, illustrating the gap between requested quantities and potential supply.
| Supply indicator | Reported figure |
|---|---|
| Orders from the three named companies | More than 3 million cards |
| Total orders from Chinese customers | More than 6 million cards |
| Potential global production in 2027 | Up to 3 million cards |
| Possible fulfilment of requested quantities | 5%–10%, according to one source |
These figures are attributed to sources cited by Reuters and represent reported orders and production expectations, not confirmed shipments.
If production remains constrained, Chinese companies may need to prioritise the workloads for which the cards are most useful, seek alternative hardware or optimise software to get more performance from their existing systems.
Nvidia could also face a difficult allocation decision if demand continues to exceed capacity. It must balance manufacturing constraints, customer commitments, regulatory requirements and demand from other markets.
The company has not guaranteed delivery of the reported quantities, and the ultimate value of completed sales could be materially below the headline order figure.
What the Orders Mean for China’s AI Ambitions
China has made AI development a major technology and industrial priority. Its technology companies are expanding AI models and services while domestic semiconductor suppliers work to increase the availability and capability of locally produced processors.
Yet demand for AI computing has grown quickly, creating a gap between the hardware companies want and what the market can supply.
A Financial Times report published in October cited industry estimates suggesting that Chinese AI-chip production remained below expected demand. The report said domestic supply could reach around 3.5 million units in 2026, including some chips manufactured overseas, while demand was expected to exceed 5 million units.
Those figures are estimates rather than a complete audited count of all AI processors available in China. They nevertheless indicate the supply pressures facing companies attempting to scale AI infrastructure.
Access to Nvidia products can help fill part of the gap, but reliance on restricted foreign hardware also creates strategic risks. Future changes to US export rules, Chinese import policies or Nvidia’s production plans could affect availability.
For Chinese technology companies, the longer-term challenge is to secure enough computing resources while reducing exposure to unpredictable supply conditions.
Why AI Inference Is Driving More Hardware Demand
AI inference has become increasingly important as companies move from developing models to deploying them for millions of users.
Every interaction with an AI assistant, automated customer-service system or content-generation tool requires computing resources. More complex models and longer interactions can increase the amount of computation needed per request.
At large scale, even modest increases in the cost of each interaction can have significant financial consequences. Technology companies therefore seek hardware that offers a suitable balance of performance, memory capacity, power consumption and operating cost.
The RTX PRO 5500 may be suitable for selected inference workloads, depending on the model and software configuration. However, professional workstation cards are not automatically interchangeable with dedicated data-centre accelerators.
Large AI deployments also require networking, storage, power, cooling, software and system management. A company purchasing millions of graphics cards would still need to integrate them into suitable computing environments before the hardware could deliver useful capacity.
The reported orders therefore indicate demand for additional computing resources, but they do not establish how much AI capacity the companies will ultimately deploy or how efficiently they will use it.
What the Orders Mean for Nvidia
The reported buying spree highlights continued interest in Nvidia’s technology despite restrictions on advanced chip exports to China.
For Nvidia, demand from China could represent a significant commercial opportunity if the company can supply products legally and economically. However, the size of the reported orders should not be confused with realised revenue.
The company must navigate a combination of export rules, manufacturing limits and competition from domestic Chinese chip suppliers. Even if orders remain strong, a shortage of production capacity could restrict the amount of hardware it sells.
Nvidia also faces a longer-term strategic challenge. Export restrictions can reduce access to a major market, while product restrictions may encourage customers to accelerate efforts to develop or adopt alternative processors.
At the same time, the reported demand suggests that Nvidia’s software ecosystem and established hardware remain important to many AI developers. Companies may continue to seek Nvidia-compatible systems because existing software, development tools and deployment processes can be difficult to replace.
The eventual business impact will depend on which orders become approved deliveries, how much capacity Nvidia allocates to the product and whether restrictions or production plans change.
The Bigger Picture
The more than $35 billion in reported Nvidia GPU orders underline the scale of China’s demand for AI computing and the effects of semiconductor export controls on purchasing decisions. ByteDance, Alibaba and Tencent are seeking additional hardware for AI services, but the RTX PRO 5500 is a workstation GPU rather than a direct replacement for every high-end data-centre accelerator. Its appeal reflects the limited set of Nvidia products available to Chinese buyers under current restrictions.
The story also demonstrates why orders alone do not tell the full financial picture. Production constraints could sharply limit deliveries, while changes in export rules could alter what can be shipped. The broader contest involves not only who can buy AI chips, but also who can manufacture them, build efficient systems around them and sustain AI services at scale.
Looking Ahead
The next important developments will be whether Nvidia can expand production, how it allocates available cards and whether the reported orders translate into actual shipments. Chinese technology companies may also explore alternative hardware and domestic processors as they try to secure reliable supplies for AI inference and development. The timing and scale of deliveries remain uncertain.
For Nvidia, the episode shows that demand for its products remains substantial even in markets affected by export controls. For China, it highlights the challenge of meeting AI infrastructure needs while access to some advanced foreign chips is restricted. The outcome will depend on manufacturing capacity, regulatory decisions and the speed at which alternative computing platforms become viable.
Frequently Asked Questions
1. Which Chinese companies ordered Nvidia GPUs worth more than $35 billion?
ByteDance, Alibaba and Tencent reportedly placed orders for more than 3 million RTX PRO 5500 graphics cards, with a combined list value exceeding $35 billion, according to Reuters.
2. What is the Nvidia RTX PRO 5500 used for?
It is a professional workstation GPU designed for demanding computing tasks. It can also support selected AI workloads, including inference, depending on the model, software and system configuration.
3. Why are Chinese companies buying this GPU instead of more advanced Nvidia chips?
US export restrictions limit Chinese access to some of Nvidia’s advanced data-centre processors. The RTX PRO 5500 is among the more capable Nvidia products reportedly available to Chinese buyers under the current restrictions.
4. Will Nvidia deliver all the ordered GPUs?
That is uncertain. Reuters reported that Nvidia may fulfil only 5%–10% of the requested quantities because demand exceeds planned production capacity. The company has not guaranteed delivery of the reported orders.
5. Does the $35 billion figure mean Nvidia has earned that amount?
No. It is the reported combined list value of the orders, based on a price of approximately $11,400 per card. Actual revenue depends on approved orders, final pricing, production and completed deliveries.
6. How do the orders affect China’s AI industry?
They show the demand for additional AI computing capacity and the difficulty of securing enough hardware under export restrictions. The purchases may help with selected workloads, but they do not remove China’s broader challenges around chip supply, system integration and domestic semiconductor development.
Sources and further reading
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
