Tech war: Huawei’s AI chip capabilities under intense scrutiny after market leader Nvidia tapped as potential rival

The Ascend 910B follows on from the success of the Ascend 910, which Huawei released in August 2019, three months after its launch. Add to commercial blacklist. by the US Department of Commerce. According to Dylan Patel, chief analyst at San Francisco-based semiconductor research firm SemiAnalysis, the chip can rival Nvidia’s A100 in terms of powering AI algorithms.
A Chinese flag is seen near the logo above the headquarters of Semiconductor Manufacturing International Corporation in Shanghai. The country’s top contract chip maker is said to be the supplier of Huawei Technologies’ new artificial intelligence chipset Ascend 910B. Photo: Bloomberg
Since 2019, US sanctions have limited Huawei’s semiconductor development and dealt a severe blow. Smartphone Business But Shenzhen-based company has been quietly bolstering its chip business by partnering with various domestic suppliers, according to several people familiar with the matter, who declined to be identified due to the sensitivity of the matter.
A Kirin 9000s processor, developed by Huawei Technologies’ chip design arm HiSilicon, is pictured in the Mate 60 Pro 5G smartphone on September 3, 2023 in Ottawa, Canada. Photo: Bloomberg
At the domestic release of the Ascend 910, Huawei touted the chip as “the world’s most powerful AI processor” at the time, manufactured by the world’s top contract chipmaker. Taiwan Semiconductor Manufacturing Cousing a 7-nm process.
Huawei’s new AI chip seems to have arrived around the same time the Mate 60 Pro was released last August. Chinese online search and AI giant Baidu orders 1,600 of Huawei’s Ascend 910B chips In the same month, November 2023, according to a Reuters report that cited a source.
Two weeks before Reuters published the report, the Chinese AI company iFlytek Launched its Feixing One computing platform based on Huawei’s Ascend chips. it means. iFlytek Spark 3.0, the firm’s updated Large Language Model (LLM), may have been developed on AI chips. LLMs are technologies used to train creative AI services. Chat GPT.
WhatsApp Group Join Now
Telegram Group Join Now
Instagram Group Join Now

Huawei declined to comment on the matter.

Jensen Huang, co-founder, president and chief executive of American chip design firm Nvidia, attends the World Governments Summit on February 12, 2024 in Dubai, United Arab Emirates. Photo: Reuters
In an interview with tech media outlet Wired last month, Nvidia’s chief executive Jensen Huang Described Huawei as a “really, really good company”.

“They’re limited by whatever semiconductor processing technology they have, but they’ll still be able to put many of these chips together to make very large systems,” Huang said.

Amid increased focus on generative AI in the past year and tougher US sanctions, Huawei and SMIC have allocated more capacity to AI chips, according to a Reuters report last month.

A GPU distributor, who declined to be named because of the sensitivity of the matter, said the Ascend 910B is “available for order, but supply is really tight at the moment”.

Huawei, ZTE showcase AI capabilities at MWC Barcelona amid industry transformation

A server used for AI training and embedded with eight Ascend 910B cards costs around 1.5 million yuan (US$208,395), which is roughly in the same range as A100 server prices in black market channels. Joe also declined to be named, according to a separate person familiar with the matter.

Huawei has not commented on the Ascend 910B, despite reiterating that AI is a “key strategy” ahead of the firm’s display at the four-day trade show. MWC Barcelona Last week in Spain

Many analysts and industry professionals are reluctant to comment on the showdown between Nvidia and Huawei, although they point out that the American chip designer has depth in GPUs and its software ecosystem benefits from CUDA. , is a computing platform that allows developers to harness the full potential of semiconductors. .

“CUDA is sticky, Nvidia has done all the hard work themselves and is reaping the benefits,” said Brian Colello, technology equity strategist at Morningstar. “Huawei and its software partners will need to build a comparable ecosystem. [to Nvidia’s CUDA] When it comes to tools for building AI models.

SMIC, Huawei among top recipients of Chinese government funding this year

Despite lagging behind CUDA’s 2 million-strong list of registered developers, Huawei has its proprietary Compute Architecture for Neural Networks, a platform that combines Ascend hardware and software, to unlock AI computing power. is important for

Colello said Huawei will have to do the same. [big] Investing in mainland China to strengthen its software capabilities. He added that other companies will probably work on software libraries, while Huawei focuses on chip design.

“Huawei’s strength is not in the software stack,” said one Shanghai-based tech investor, requesting anonymity. “U.S. sanctions have limited chip performance and production output.”

WhatsApp Group Join Now
Telegram Group Join Now
Instagram Group Join Now

Leave a Comment