Nvidia vs Huawei: Chip Sector Dominance Battle Heats Up with AI Demand

As demand for artificial intelligence (AI) applications continues to soar, the chip sector is witnessing a heated battle for dominance between two of the world’s leading tech giants: Nvidia and Huawei.

Both companies are vying for supremacy in the development and manufacturing of high-performance processors capable of powering the next generation of AI-driven innovations.

Rival Recognition

Nvidia‘s Chief Executive Officer, Jensen Huang, recognized Huawei as a potent contender among a league of “very formidable” rivals in the intense competition to develop cutting-edge AI chips. In addition to Huawei, Intel and a burgeoning cadre of semiconductor startups are presenting a “formidable challenge” to Nvidia’s established supremacy in the artificial intelligence accelerator market.

Huang, addressing reporters in Singapore, emphasized the significant competition posed by Shenzhen-based Huawei, a company that has solidified its position as China’s de facto chip technology champion. Huawei has recently gained attention with its remarkably advanced domestically-produced smartphone processor, showcasing its capabilities on the global stage.

“We have a lot of competitors, in China and outside China – Huang said  -. Most of our competitors don’t really care where I am. They want to compete with us everywhere we go.”

Nvidia views Huawei as a strong adversary, particularly due to concerns about U.S. chipmakers facing challenges in establishing supply chains independent of China within the next decade. This vulnerability could have profound implications in the face of international tensions.

Given China’s historical significance, accounting for a fifth of Nvidia’s sales, the $7 billion market could swiftly shift allegiances to a local chipmaker, highlighting the strategic importance of this evolving competition.

Nvidia, with its expertise in graphics processing units (GPUs), has emerged as a frontrunner in the AI chip race. Its GPUs, renowned for their ability to handle complex calculations and parallel processing, are the backbone of many AI applications, including self-driving cars, virtual reality (VR) systems, and machine learning (ML) algorithms.

Huawei, on the other hand, has made significant strides in developing its own AI chips, known as Ascend processors.

Although they are not yet as widely adopted as Nvidia’s GPUs, Ascend processors are gaining traction for their energy efficiency and compatibility with Huawei’s proprietary hardware and software platforms.

Huawei Ambitions

In the wake of U.S. restrictions on Nvidia’s advanced artificial chips sales to China, Huawei is capitalizing on the opportunity to expand its market share, reportedly winning a substantial AI chip order from Chinese tech giant Baidu – worth 9888.HK – this year.

While globally recognized for its telecommunications and smartphone ventures, Huawei has been strategically developing its Ascend AI chip series for the past four years. Positioned as a direct competitor to Nvidia’s A100 chip, Huawei’s flagship product in this domain is the Ascend 910B.

Huawei unveiled the Ascend 910 in 2018, officially launching it in 2019 as part of its comprehensive AI portfolio strategy. Despite U.S. export controls, Huawei claimed the Ascend 910, manufactured on a 7-nanometer process, to be the world’s most powerful AI processor.

The chip achieved remarkable performance with 256 TeraFLOPS for half-precision floating-point (FP16) operations and 512 TeraOPS for integer precision calculations (INT8). Additionally, Huawei emphasized the chip’s efficiency, surpassing its initial power consumption target.

However, Nvidia’s A100 and H100 chips introduced in 2020 and 2022 solidified the company’s dominance in the global AI chip market.

Praise From Competitors

In August this year, the chairman of Chinese AI giant iFlytek praised  Huawei for producing a GPU comparable to Nvidia’s A100, revealing collaboration on hardware development. Subsequent reports identified this hardware as the Ascend 910B, previously undisclosed.

Documents related to the Ascend 910B, including driver and firmware upgrade guides, surfaced on Huawei’s website in August. During iFlyTek’s recent earnings call, Senior Vice President, Jiang Tao, reaffirmed that the Ascend 910B’s capabilities were “comparable to Nvidia’s A100.”

In August, Baidu met the demand for advanced AI chips by ordering 1,600 Huawei 910B chips for 200 servers. Analysts and sources acknowledge that while the 910B chips compete with Nvidia’s in raw computing power, there may be a slight lag in performance.

Nonetheless, these chips offer a sophisticated domestic alternative within China, establishing Huawei as a significant player in the dynamic field of AI chip technology.

Chinese Dominance Plans

With a burgeoning AI chip market projected to reach $7 billion in China alone, Huawei’s ambitions to challenge the dominance of U.S.-based Nvidia are gaining traction. Huawei’s CFO, Meng Wanzhou, has reiterated the company’s commitment to becoming a leading provider of AI computing power.

He stated in September that Huawei aims to establish a domestic computing base for China and offer the world a “second option” – a subtle yet firm declaration against the existing U.S. monopoly.

Although China’s AI firms presently use domestically produced chips, such as Huawei’s, which currently lag behind Nvidia’s performance, analysts foresee rapid advancements.

The substantial support and investment from the Chinese government in AI and semiconductor development suggest that Huawei could close the performance gap. This could position the company as a strong global competitor in the AI chip market.

The U.S., cautious about enhancing China’s tech capabilities, restricts the export of its advanced chips. Nvidia faces uncertainty, expecting a notable drop in Chinese sales in the coming months. The uncertainty lies in whether these sales will recover after the imposed ban.

Complicating matters, Huawei surprised the U.S. with a domestic smartphone processor that competes strongly against two-year-old chips from U.S. manufacturer Qualcomm. If China sustains its support for the local tech ecosystem, the motivation to shift business to American companies diminishes.

Who Else Can Undermine Nvidia?

Nvidia currently holds a dominant position in the GPU market, but several factors could challenge its supremacy in the future. Here are some potential contenders that could undermine Nvidia’s dominance, excluding Huawei:

AMD

Nvidia’s closest competitor in the GPU market, AMD  is offering competitive products at lower prices. AMD has been steadily gaining market share in recent years, and its recent advancements in its RDNA architecture could pose a significant threat to Nvidia.

Samsung Electronics

It is a major chipmaker with expertise in both CPUs and GPUs. Samsung  is collaborating with AMD to develop next-generation GPUs, which could further disrupt the market.

Microsoft

Bill Gate’s company  is developing its own custom ARM-based GPUs for use in its Surface devices and cloud infrastructure. These chips could potentially rival Nvidia’s offerings in certain segments, such as high-performance computing and edge computing.

Intel

The world’s largest chipmaker is also entering the GPU market with its discrete Arc graphics cards. While Intel ‘s initial offerings have faced challenges, the company has significant resources and expertise that could eventually translate into a competitive product.

Broadcom

It is a diversified technology company with a strong track record in developing high-performance semiconductors. Broadcom  has expressed interest in expanding its product portfolio to include GPUs, which could shake up the market if the company successfully enters the segment.

Startups

Several startups are developing innovative GPU technologies, including specialized AI accelerators and custom-designed chips optimized for specific applications. These startups could emerge as disruptors if their technologies prove to be superior to those of established players.