Key Highlights

  • Huawei introduced the Atlas 350 AI accelerator card, featuring its Ascend 950PR chip technology
  • The Atlas 350 reportedly achieves 1.56 petaflops of FP4 computing performance — representing a 2.8x advantage over Nvidia’s H20
  • Jensen Huang, Nvidia’s CEO, announced last week that H200 production for the Chinese market has resumed
  • KeyBanc projects potential Nvidia revenue from China at $30 billion this year, based on approximately 1.5 million H200 unit sales
  • After opening 1.6% higher in premarket trading Monday, Nvidia shares ended the session down 3.28%

Huawei has introduced a new AI chip card that the company positions as superior to Nvidia’s top-performing hardware approved for Chinese markets. The announcement comes just days after Nvidia resumed production of its H200 processors destined for China.

The Atlas 350 represents an accelerator card centered on Huawei’s Ascend 950PR chip architecture. Its primary function targets AI inference workloads — the computational process where trained models generate outputs. Applications span search result recommendations, large language model query responses, and multimodal content creation.

At Huawei’s China Partner Conference on Friday, VP Ma Haixu presented the card to attendees. Zhang Dixuan, who leads Huawei’s Ascend computing division, specified the performance figure at 1.56 petaflops of FP4 computing capability.

This represents a 2.8x performance improvement compared to Nvidia’s H20 — the chip Nvidia currently has authorization to distribute in China within US export control parameters.

Competition Intensifies for Nvidia in Chinese Markets

Jensen Huang, Nvidia’s CEO, revealed last week that the company had recommenced H200 processor manufacturing targeting Chinese customers and had begun receiving purchase orders. The H200 occupies a mid-tier position in Nvidia’s lineup, offering more capability than the H20 while trailing the latest Blackwell architecture.

Huang has characterized China as a $50 billion annual opportunity for AI infrastructure, expanding at a 50% yearly growth rate. John Vinh, an analyst at KeyBanc, projects Chinese enterprises could purchase approximately 1.5 million H200 chips throughout this year if regulatory approval permits large-scale transactions — translating to roughly $30 billion in prospective revenue.

This revenue potential holds significant importance for Nvidia, particularly as the stock remains confined within a trading range seeking momentum.

Nvidia shares opened 1.6% higher in premarket activity on Monday, supported by broader market optimism following President Trump’s comments about ongoing US-Iran diplomatic discussions. However, by market close, NVDA had declined 3.28%.

Evaluating the Competitive Landscape

Comparing chips directly presents challenges. Various system configurations can yield substantially different performance metrics, and Huawei’s earlier AI server implementations have faced scrutiny regarding power efficiency relative to Nvidia’s offerings.

The existence of a viable domestic alternative carries weight in the Chinese market. Organizations considering GPU purchases now have access to a locally-produced option — one insulated from US export policy uncertainties.

Huawei has been advancing this strategy methodically. The Ascend 950PR chip received its initial unveiling in September 2025 as part of a three-year development timeline. The company has constructed its AI chip ecosystem independently of American technology following the imposition of US sanctions.

Regarding storage infrastructure, Huawei announced planned enhancements to its OceanStor Dorado and Pacific 9926 platforms scheduled for this year, alongside the introduction of the FusionCube A1000 cabinet designed for small and medium enterprise AI deployments.

“While the first half of the AI era focused on computing power, the second half will be defined by data,” said Yuan Yuan, president of Huawei’s data storage product line.

Nvidia’s H200 production resumption for China and KeyBanc’s $30 billion revenue projection represent the latest tangible indicators shaping market expectations.