Despite U.S. export controls aimed at preventing Chinese companies from acquiring advanced AI chips, small cloud service providers in China have reportedly found ways to obtain NVIDIA’s A100 and H100 chips. The cost of renting cloud services in China is even lower than in the U.S.

According to a report from the Financial Times, four small-scale Chinese cloud providers are offering servers equipped with eight A100 chips each, charging around USD 6 per hour. In comparison, similar services from U.S. cloud providers cost approximately USD 10 per hour.

As Chinese companies are reportedly bypassing U.S. export controls, industry sources cited by the Financial Times have further noted that the lower prices in China may hint at a robust local supply of NVIDIA chips.

Since the fall of 2022, the U.S. has banned NVIDIA from supplying A100 chips to China, and the more powerful H100 chips have not been approved for sale there.

However, industry sources and startups have revealed that these chips are still available in China. Ads for A100 and H100 have appeared on social media platforms like Xiaohongshu and e-commerce sites such as Taobao, with prices higher than those abroad.

At the Huaqiangbei electronics market in Shenzhen, reportedly, industry sources have revealed that the price of NVIDIA’s H100 is quoted at USD 23,000 to USD 30,000, while Chinese online sellers list it at USD 31,000 to USD 33,000.

Meanwhile, larger Chinese cloud providers such as Alibaba and ByteDance emphasize service stability and security in the local market. For servers equipped with A100 chips, they charge two to four times more than smaller cloud providers.

According to another source cited by Financial Times, large companies must consider regulatory compliance, which puts them at a disadvantage because they are reluctant to use smuggled chips. In contrast, smaller providers are less concerned.

The same report also indicate that after the US government tightened export controls in October last year, servers from Supermicro equipped with eight H100 chips were priced as high as approximately CNY 3.2 million. However, as supply constraints eased, the price has dropped to around CNY 2.5 million.

Several sources cited by the report claim that merchants from Malaysia, Japan, and Indonesia frequently ship Supermicro servers or NVIDIA chips to Hong Kong, from where they are then transported to Shenzhen.

In response to these issues, NVIDIA reportedly stated that it primarily sells chips to well-known partners, ensuring that all sales comply with U.S. export regulations.

NVIDIA also mentioned that its used products can be obtained through various channels and, although they cannot track products after sale, they will take appropriate action if they determine a customer is violating U.S. export controls.

Read more

• [News] China’s 1H24 Chip Equipment Purchases Exceed Taiwan, Korea, and US Combined, Reaching USD 25 Billion
• [News] China Contributes Over 40% of U.S. Chip Equipment Makers’ Revenue despite Export Controls

(Photo credit: iStock)

As per a report from Bloomberg citing sources, it’s pointed out that TSMC’s first US fab in Arizona began producing engineering wafers using the 4nm process in April, with yields reportedly comparable to those manufactured in its Southern Taiwan Science Park facility, showing promise in meeting its targets.

Addressing on the matter, TSMC reportedly stated that the project is proceeding as planned and is making good progress, without further commenting on the yield.

TSMC plans to build three plants in Arizona, each with cleanroom spaces twice the size of typical logic fabs in the industry. The first fab is expected to begin mass production in the first half of 2025.

TSMC’s second fab in Arizona will use 2nm process technology to meet strong AI-related demand, with production expected to begin in 2028. The third fab will employ 2nm or even more advanced process.

TSMC stated that once the fabs are fully operational, they will deliver the same level of manufacturing quality and reliability in Arizona as is provided in its Taiwan facilities.

The U.S. government officially announced in April that it will provide a USD 6.6 billion subsidy to TSMC, and confirmed that TSMC will build its third fab in Arizona, USA, with total investment rising to USD 65 billion.

Read more

• [News] TSMC’s Kumamoto Fab Influence Expands, Economic Spillover Effect Expected to Exceed JPY 10 Trillion
• [News] TSMC Receives USD 6.6 Billion for Chipmaking from the US Government, Commits to Constructing Third Fab

(Photo credit: TSMC)

Intel has outsourced the production of its Lunar Lake processors to TSMC. According to a report from Commercial Times, due to recent setbacks in Intel’s wafer foundry business, the company has decided to outsource all sub-3nm process manufacturing to TSMC.

The company is also said to be implementing a global 15% workforce reduction plan in an effort to reverse its decline. However, industry sources also reveal that the layoffs are primarily targeting the foundry business, while Intel’s Taiwan branch remains unaffected to maintain production partnerships with local chip manufacturers.

Intel remains committed to its wafer foundry business, as it reportedly released the 18A process design kit (PDK) to IC manufacturers in July.

However, recent reports indicate that Broadcom has expressed concerns about the feasibility of Intel’s 18A process, concluding that it is not suitable for mass production. A Broadcom spokesperson cited by the report has indicated that it is currently evaluating Intel Foundry Services’ products and services, though no final assessment has been reached yet.

Industry sources cited by Commercial Times further note that Broadcom has been collaborating with TSMC for many years, particularly in advanced processes below 7nm, positioning itself as a key player and securing a spot among TSMC’s top ten customers.

Looking at Intel’s latest quarterly report, its foundry business posted a loss of USD 2.8 billion, with an operating profit margin of -65.5%. The company acknowledged that the ongoing expansion of its Intel 4 and Intel 3 facilities in Ireland, along with increased R&D and startup costs for advancing its technology development, will impact profitability.

This has thus underscored the significant challenges Intel faced in achieving both technological breakthroughs and mass production in the semiconductor industry.

Intel is cutting costs and driving efficiency while actively pursuing transformation. The company aims to save USD 10 billion by 2025, even halted dividend payments—a move not seen in 30 years.

Additionally, its global expansion efforts have slowed. As per a recent report from Malaysian media outlet The Star, citing informed sources, it’s reported that Intel will temporarily halt its new chip packaging and testing project in Penang as part of cost-cutting efforts.

Read more

• [News] Amidst Rumors of Germany Plant Delay, Intel’s Penang Project Also Reportedly on Hold
• [News] Intel’s 18A Reportedly Runs into Trouble with Broadcom, while 20A Plan on Arrow Lake Cancelled

(Photo credit: Intel)

In late August, it is said that the Netherlands mulls to ban ASML, the country’s semiconductor equipment giant, from conducting equipment maintenance and providing related backup components in China. Now the latest regulation has been revealed, as the Dutch government announced last Friday the expansion of export restrictions on advanced semiconductor manufacturing equipment, which covers two of ASML’s DUV immersion lithography systems, according to a report by CNBC.

The new export control rule, which took effect on September 7^th, indicates that ASML will now have to apply for licenses with the Hague rather than the US authority for some of its machines, as the Dutch government regards these curbs as a critical measure for national security, and attempts to gradually take the initiative rather than following the U.S., a report by Bloomberg notes.

According to the official announcement made by ASML, the new rule will require its TWINSCAN NXT:1970i and 1980i DUV immersion lithography systems to obtain a license from the Dutch government before being exported. The Dutch export license requirement is already in place for ASML’s TWINSCAN NXT:2000i and subsequent DUV immersion systems.

Regarding ASML’s sales in lithography units in the second quarter of 2024, China emerged as the largest market, as it contributed 49% of the revenue, higher than South Korea’s 28% and Taiwan’s 11%.

Interesting enough, ASML refers to the updated license requirement as “a technical change,” and is not expected to have any impact on the company’s financial outlook for 2024 or for its longer-term scenarios.

The latest move from the Dutch government is less harsh than the rumors earlier, which indicated that the Netherlands might ban ASML from conducting equipment maintenance and providing related backup components in China. The measure, if implemented, would be a heavy blow to China’s semiconductor industry, especially on the development of advanced nodes.

In response, China has expressed “dissatisfaction” with the Dutch government’s decision to expand export controls on ASML chipmaking equipment, according to the statement by the Chinese commerce ministry on Sunday, cited by Reuters.

The ministry urged the Dutch government not to misuse export controls, to avoid actions that could harm Sino-Dutch cooperation in the semiconductor industry, and to protect the “shared interests of Chinese and Dutch enterprises,” according to Reuters.

Beijing has consistently been criticizing Washington’s approach of pressuring allies like the Netherlands and Japan to implement export controls aimed at limiting China’s access to advanced chips and chipmaking equipment.

Read more

• [News] ASML Claims the World Counts on China for Legacy Chips; So Do Chip Equipment Giants
• [News] ASML Might Not Be Able to Provide Advanced Chipmaking Maintenance Services to China as Early as 2025

(Photo credit: ASML)

TSMC’s fab in Kikuyo, Kumamoto Prefecture, Japan (Kumamoto Fab 1) is expected to begin mass production by the end of 2024, with plans for a second fab in the region. Thus, the influence of TSMC’s presence continues to expand, as per the latest estimates from local financial institutions.

Over the next decade, from now until 2031, the economic spillover effect of TSMC’s operations in Kumamoto is projected to surpass 10 trillion yen, marking a 60% increase from a previous estimate in August 2023.

According to a report from Nikkei citing Kyushu Financial Group (Kyushu FG), a new impact estimate has been released on September 5, showing that TSMC’s operations in Kumamoto Prefecture are projected to generate an economic spillover effect of approximately JPY 11.2 trillion over the next 10 years, until 2031.

This marks a 60% increase from the previous estimate of JPY 6.9 trillion published in August 2023. The projected impact on Kumamoto Prefecture’s GDP over the same period has also risen from JPY 3.4 trillion to JPY 5.6 trillion.

Reportedly, the previous estimate from Kyushu FG last August only considered the benefits of TSMC’s Kumamoto Fab 1. The latest report, however, includes the planned construction of the Kumamoto Fab 2 in its evaluation.

The upward revision is attributed to the expanded magnetic pull of TSMC’s Kumamoto operations (both fabs). The number of companies expected to set up or invest in the region has increased to 171, roughly double the previous estimate.

Initially, the first Kumamoto fab attracted strong interest from suppliers like Sony and Mitsubishi Electric. Following TSMC’s announcement in February to build a second fab, further investments are expected, not only from within Kumamoto but also from other prefectures and overseas suppliers, particularly from Taiwan.

Additionally, the economic impact is expected to extend to wage levels in Kumamoto Prefecture, with an estimated increase of JPY 380,000 in per capita annual income.

Meanwhile, as stated in an report from Bloomberg on May 11th, Kumamoto’s newly appointed governor, Takashi Kimura, once claimed that he would spare no effort to persuade TSMC to establish a third fab in the region.

In addition, a recent report from Kyodo News citing the interview with Taiwanese Minister of Economic Affairs J.W. Kuo has also hinted that TSMC plans to build a third fab in Japan, but with a projected timeline after 2030.

If the third fab is realized, the economic spillover effect is anticipated to expand further.

Read more

• [News] TSMC to Build Third Fab in Japan? The Timing May Be Reportedly after 2030
• [News] TSMC’s Kumamoto Fab 2 Reportedly Begins Land Preparation

(Photo credit: JASM)