Huawei’s new AI SuperPoD could be its biggest Nvidia challenge yet

Huawei launched its next-generation AI technology this week, with plans to challenge Nvidia’s dominance in the global AI market.

At Huawei’s annual summit in Shanghai this Thursday, chairman Wang Tao unveiled the new Ascend 960 SuperPoD cluster, which is scheduled to hit the market in 2027, despite US restrictions. The firm stated that the 960DT chip will be available in the first quarter of 2027, while the Ascend 960PR will arrive in Q3.

Tao also introduced an upgraded iteration of its UnifiedBus interconnect technology. Moreover, the company added near-packaged optics (NPOs) to its SuperPoD architecture to address the dual challenges of speed and energy consumption.

It stated that integrating its UnifiedBus interconnect and Hi-ONE optical systems can now tie 4,000 processors together in a single SuperPoD, creating deep, high-speed optical paths throughout the network. 

Huawei’s Guo said they intend to focus on increasing the value of its chips

The Shenzhen-based firm is aiming for total compatibility; supervisory board Chairman Guo Ping noted earlier this month that the ultimate goal is for the new Ascend chips to power every AI model globally. He argued that Huawei can narrow the technology disparity by rethinking chip architecture, much like how Apple overcame its early limitations in chipmaking by focusing on custom design and software efficiency. 

He insisted that moving forward, the firm will deepen its capabilities and maximize value in the areas it has prioritized. 

“We expect the company to achieve even better results, providing digital infrastructure solutions to global customers, offering diversified solutions for the development of artificial intelligence, and contributing to the rise of China’s electronics industry,” he contended.

Aside from the new launches, the company recently secured a massive commitment from DeepSeek. The AI startup is building a data center in Inner Mongolia that will run on a massive fleet of 160,000 Ascend 950DT processors. 

Right now, the Ascend 960 is the next step up from Huawei’s current Ascend 950 chips. Though the firm is planning to release another set, the Ascend 970, in 2028, keeping pace with a three-year pledge to challenge Nvidia by doubling processor performance with every iteration. 

Huawei said it will focus on boosting its volumes and networking

Previously, Huawei had acknowledged that its hardware falls short of Nvidia’s in terms of pure speed and power. Thus, to make up the difference, it asserted that it would prioritize leveraging sheer hardware volume, superior networking, and state policy support to win over consumers.

It emphasized that system-level adaptations and networking can help offset limitations in individual chip performance. 

At the time, Bernstein analysts led by Qingyuan Lin, seemingly pleased with the Shenzhen-based chipmaker’s approach, wrote, “Huawei’s willingness to publicly articulate its AI roadmap represents a strong signal of confidence in the resilience of its future local foundry supply.” 

Recently, Morgan Stanley analyst Charlie Chan also stated, “System-level competitiveness matters more than ever. The effective gap is narrowing through multi-die design, advanced packaging, rack-scale system architecture, optical networking, and software-hardware co-optimization.”

Huawei’s new chips face a challenging AI market

Huawei’s latest roadmap comes as demand for AI computing continues to grow, putting more pressure on Chinese technology companies to access advanced AI processors.

The company is looking to create an alternative AI hardware ecosystem that is less dependent on foreign vendors and more local than foreign suppliers, making it easier for Chinese developers to access domestic computing capabilities.

The Ascend 960 will be important not only for its processor performance but for how Huawei is interwoven with networking, optical technology, and software.

By linking thousands of processors, Huawei aims to build large computing systems capable of handling demanding AI workloads.

This is becoming more important as the US export restrictions limit Chinese companies’ access to Nvidia’s most advanced AI processors. Huawei’s strategy is to improve system-level performance instead of just individual chip speed.

The company’s progress will also depend on its ability to secure enough advanced manufacturing capacity and persuade AI developers to adopt its hardware and software ecosystem.

However, according to Bernstein analysts, Huawei’s next-generation Ascend 950 delivers only a fraction—around 6%—of the performance of Nvidia’s upcoming VR200 superchip. Huawei, nevertheless, says it is exploring ways to cluster up to one million chips to theoretically close that performance gap. It explained that its in-house high-bandwidth memory architecture can improve data transfer between a processor’s various components. 

For some time now, Chinese regulators have been limiting local tech giants’ ability to purchase Nvidia’s chips, including the H200 series. However, more recently, Chinese authorities have allowed a few batches of Nvidia’s H200 chips into the mainland, to ensure that local AI developers can continue training their models.

According to sources, ByteDance and Tencent recently got the green light to import about 10,000 H200 processors. Nonetheless, even without official authorization, Nvidia’s most advanced processors continue to reach the Chinese market, though in limited quantities.

Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.