Amid market rumors that the Beijing authority has been advising local companies not to use NVIDIA’s H20, which is tailored for the Chinese market, tech conglomerate Huawei is said to initiate sampling of its latest AI accelerator, Ascend 910C, to Chinese customers, according to reports by the South China Morning Post and Tom’s Hardware.

According to the South China Morning Post, large Chinese server companies and internet firms have received the samples of the Ascend 910C, which is regarded as an upgraded version of the Ascend 910B.

Sources cited by a previous report of the Wall Street Journal noted that TikTok’s parent company ByteDance, search giant Baidu, and state-owned telecom operator China Mobile are in preliminary talks with Huawei to secure the Ascend 910C chip.

Citing remarks from Eric Xu, Huawei’s Rotating Chairman, the South China Morning Post indicates that as the AI chips embargo launched by the U.S. is unlikely to be lifted soon, the scenario gives Huawei an opportunity to step in.

Huawei’s Ascend 910B, which the company claims to rival NVIDIA’s A100, has been popular among multiple industries across the country for AI model training. According to the report by the South China Morning Post, the tech giant’s Ascend solutions were used to train roughly half of more than 70 of China’s top large language models as of last year.

According to the report, NVIDIA is projected to ship over 1 million H20 GPUs to China this year, generating around USD 12 billion in revenue. Initially, demand for the H20 was sluggish, but sales have steadily gained momentum in recent months.

In terms of the upcoming Ascend 910C, Tom’s Hardware notes that it may reportedly sell for roughly USD 2 billion together this year, as the launch time would probably fall in the fourth quarter.

However, it is worth noting that Huawei also stands to gain by locking Chinese companies into its Ascend 910C hardware-software ecosystem, the South China Morning Post suggests. Citing a server company employee, the report states that if a firm buys Huawei’s AI chips, it may also need to purchase other offerings, like network and storage solutions, which might cause some hesitation.

Read more

• [News] Huawei Rumored to Launch New High-End AI Chip, Potentially Rivaling NVIDIA’s H100
• [News] U.S. Reportedly Enhancing Chip Export Restrictions to China, while NVIDIA’s H20 Might Get Banned

(Photo credit: Huawei)

Rumors have been circulating that NVIDIA has stopped taking orders for its H20 chips customized for China since August. Now, according to the latest report by Bloomberg, regulators in China have been advising companies against buying H20, as part of the country’s strategy to bolster its semiconductor industry and respond to further US sanctions.

As the initiative aims to boost the market share of domestic Chinese AI chip manufacturers, Huawei and Cambricon Technologies, which are leading AI processor makers in China, may turn out to be the major beneficiaries, Bloomberg suggests.

Beijing’s approach has been more of a guideline than a strict prohibition, as the authority still hopes to support its own AI startups, the report notes.

However, it is indicated that in recent months, several Chinese regulators, including the Ministry of Industry and Information Technology, did issue the so-called “window guidance”—informal instructions that lack legal authority—to minimize the use of NVIDIA.

It is worth noting that China has a thriving AI sector amid US restrictions. Major tech player like ByteDance and Alibaba are making significant investments, while numerous startups are vying for dominance. According to an earlier report by The Information, it is rumored that ByteDance has ordered over 200,000 NVIDIA H20 chips this year for AI model training, costing it over USD 2 billion.

In addition, there are six rising stars in the country’s development of large language models, which are crucial for generative AI, including 01.AI, Baichuan, Moonshot, MiniMax, Stepfun, and Zhipu, Bloomberg notes.

According to Bloomberg, some companies are disregarding the Chinese directive to avoid H20 chips, hastily acquiring more before a potential US sanction by the end of the year. However, they are also purchasing domestic Huawei chips to appease Beijing.

As early as in 2022, the US government prohibited NVIDIA from selling its most advanced AI processors to Chinese clients to curb Beijing’s technological progress. In response, the AI chip giant launched a series of AI chips tailored for the Chinese market, including H20, L20 and L2. According to a previous report by Wccftech, H20 GPU has 41% fewer Cores and 28% lower performance versus H100.

NVIDIA declined to comment to Bloomberg’s report, neither did China’s Ministry of Commerce, Ministry of Information and Technology, and Cyberspace Administration respond, Bloomberg notes.

In a separate statement, NVIDIA CEO Jensen Huang noted in an interview with Bloomberg Television that he is focused on serving customers in China while adhering to US government restrictions.

Read more

• [News] Nvidia’s Custom H20 Chip for China Reportedly Halted

(Photo credit: NVIDIA)

The industry holds high expectations for photonic chips, which play a crucial role in data centers, especially in high-bandwidth and energy-efficient data transmission. With the growing demand for efficient data processing driven by the rise of artificial intelligence, cloud computing, and IoT devices, the research and development of photonic chips have become increasingly urgent.

China’s First Photonic Chip Pilot Line Launched in Wuxi

On September 25th, the first domestic photonic chip pilot line built by the Wuxi Photonic Chip Research Institute of Shanghai Jiao Tong University was officially launched. After the pilot line becomes operational, it is expected to reach an annual production capacity of 10,000 wafers. The first PDK will be released in the first quarter of 2025, providing external chip fabrication services.

The photonic chip pilot platform covers an area of 17,000 square meters, integrating research, production, and services. It is equipped with more than 100 top-tier international CMOS process machines, supporting a full closed-loop process for lithium niobate photonic chips, from lithography, thin film deposition, etching, wet processing, cutting, measurement, to packaging. The platform also supports other material systems like silicon and silicon nitride.

Jinan Achieves World’s First 12-Inch Lithium Niobate Crystal

in May of this year, Shandong Hengyuan Semiconductor Technology Co., Ltd. in Jinan successfully developed a 12-inch (300mm diameter) large-sized optical-grade lithium niobate crystal.

Shandong Hengyuan Semiconductor has been dedicated to the R&D of optoelectronic materials such as lithium niobate and lithium tantalate, as well as piezoelectric materials. Through technological advancements, the company has started mass production of 6-8 inch Z-axis and X-axis optical-grade lithium niobate crystals. Within three years, Hengyuan plans to increase its annual wafer production to 250,000 units.

Chinese Scientists Develop Mass-Produced New “Optical Silicon” Chips

In early May, the research team led by researcher Ou Xin at the Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, made a breakthrough in the preparation of lithium tantalate heterogeneously integrated wafers and high-performance photonic chips. They successfully developed a new type of “optical silicon” chip that can be mass-produced. The research results were published in Nature on May 8th.

Tsinghua University Team Releases AI Photonic Chip

In August, Tsinghua University announced that the research group led by Professor Fang Lu from the Department of Electronic Engineering and the team led by Academician Dai Qionghai from the Department of Automation pioneered a fully forward intelligent optical computing training architecture. They developed the “Taiji-II” optical training chip, enabling efficient and precise training of large-scale neural networks in optical computing systems. This research, titled “All-Forward Training of Optical Neural Networks,” was published in Nature.

The previously released Taiji-I, in April, achieved 879 T MACS/mm²area efficiency and 160 TOPS/W energy efficiency, marking the first time optical computing was applied to complex AI tasks such as natural scene recognition with thousands of categories and cross-modal content generation.

Chinese Team Successfully Develops Fully Programmable Topological Photonic Chip

In late May, the “Extreme Optics Innovation Research Team” from Peking University’s School of Physics, in collaboration with Researcher Yang Yan from the Institute of Microelectronics, Chinese Academy of Sciences, proposed and realized a fully programmable topological photonic chip based on large-scale integrated optics.

The chip built on a reconfigurable integrated optical micro-ring array, integrates 2,712 components in an area of just 11mm x 7mm. It successfully achieved the world’s first fully programmable optical artificial atomic lattice. The researchers also experimentally validated multiple topological phenomena on a single-chip platform, including dynamic topological phase transitions, multi-lattice topological insulators, statistical topological robustness, and Anderson topological insulators.

The wild journey of Intel has yet to end, as tech giants have been approaching the company for potential acquisitions. In addition to Qualcomm, UK-based Arm is also said to inquire the possibility of acquiring the struggling chipmaker’s product division, according to the latest report by Bloomberg.

However, the report notes that Arm was informed by Intel that the division is not for sale, according to a source familiar with the situation.

Around mid-September, Intel settled down plans for restructuring after the board meeting, and revealed schemes to transform its foundry business into an independent unit with its own board. The strategy will allow its foundry business to explore independent sources of funding.

In April, Intel disclosed the financials for the foundry business for the first time, with an operating loss of USD 7 billion in 2023, a previous report by CNBC stated.

Arm, according to Bloomberg, showed little interest in Intel’s manufacturing operations. Instead, it reportedly expressed the intention to acquire Intel’s product division, which sells chips for PCs, servers, and networking equipment, though the request was turned down afterwards.

The move did make sense. With an 88% stake owned by SoftBank, Arm generates a significant portion of its revenue from selling chip designs to smartphone-related clients, including Qualcomm, Samsung and Amazon.

According to Bloomberg, Arm CEO Rene Haas targets to gain a foothold in various applications, such as personal computers and servers, in which Intel still takes a lead. A report by Reuters in June notes that Haas aims to capture 50% of PC market in five years.

If the company were to partner with Intel, it would reportedly enhance its market reach, while accelerate the shift toward selling more of its own products, Bloomberg observes.

Unlike Intel, which remains a bystander in the AI boom, Arm is considered to be a main beneficiary of the wave. After being acquired by Japan’s SoftBank in 2016 in a USD 32 billion deal, it went public in September, 2023, on Nasdaq, with a market valuation around USD 54 billion. A year after, its stock price has nearly tripled, with a market value exceeding USD 150 billion.

Representatives for both Arm and Intel declined to comment, Bloomberg notes.

On the other hand, another report by Financial Times, cited by MoneyDJ, reveals that Intel and the U.S. government are on track to finalize the USD 8.5 billion subsidy in direct funding under the CHIPS Act by the end of this year.

According to the reports, the two parties are working to complete the technical negotiations that have been ongoing for several months, while Intel is also undertaking large-scale cost-cutting measures. An insider familiar with Intel indicated that it wouldn’t be surprising if the negotiation results were announced around the upcoming presidential election.

Read more

• [News] A Quick Roundup of Intel’s Five Core Businesses: Key Moves to Watch Next
• [News] Update on Potential Investors of Intel: Apollo Rumored in; Broadcom Temporarily out

(Photo credit: ARM)

Amid an operational crisis, Intel has abandoned its “5 Nodes in 4 Years” plan, shelving the Intel 20A process node to focus entirely on the more advanced Intel 18A. Now, there is finally good news regarding the 18A process.

During the recent Enterprise Tech Tour in Portland, Oregon, CEO Pat Gelsinger made the first public reveal of the Clearwater Forest Xeon server processor, produced using the Intel 18A node. This chip is a crucial element in determining whether Intel’s ambitious transformation plan will succeed, according to Tom’s Hardware.

The report highlights that Clearwater Forest is the first high-volume chip to be fabricated on the Intel 18A node—a process so pivotal that Gelsinger has effectively staked the entire company’s future on its success. Although the chip was showcased at the event, it won’t hit the market until next year.

Tom’s Hardware also notes that while Intel will manufacture a range of other processors on the 18A node, delivering the Clearwater Forest chips on schedule is critical to restoring confidence in Intel’s foundry business among potential customers. This node is key to Gelsinger’s larger turnaround strategy, representing the culmination of his bold yet desperate push to develop five nodes in four years to revive Intel’s fortunes.

At the same event, Intel also introduced its Granite Rapids Xeon chips for data centers, boasting core counts that finally match those of AMD’s EPYC processors—an achievement Intel has struggled to reach since EPYC’s debut in 2017.

(Photo credit: Intel)