According to a report from Tom’s Hardware, the U.S. is considering implementing new trade sanctions on China, looking to limit China’s access to advanced AI chip technology. This could result in a ban on NVIDIA’s HGX-H20 AI GPUs to China. If implemented, NVIDIA could potentially lose roughly USD 12 billion in revenue.

To comply with U.S. export regulations, NVIDIA introduced the HGX H20 GPU specifically for the Chinese market. Although it has reduced performance, it still offers powerful AI capabilities.

As per the report, the HGX H20 GPU features 296 INT8 TOPS/FP8 TFLOPS computational performance, 96 GB of HBM3 memory, and 4.0 TB/s memory bandwidth, making it competitive with the current entry-level AI chips on the market. Despite its downgraded performance, the HGX H20 outperforms Huawei’s self-developed Ascend 920 series AI chips in practical applications due to its better memory performance.

However, during the U.S. semiconductor export policy review in October, NVIDIA’s HGX H20 GPU might face a sales ban. The anticipated restrictions could take various forms, including product-specific bans, reduced computational power, or limited memory capacity.

Most Chinese AI companies have built their application ecosystems on NVIDIA’s CUDA computing platform, which makes switching to other platforms, like Huawei’s Ascend chips, both costly and time-consuming. Although the HGX H20 GPU’s computational performance is significantly lower than the H100, its full compatibility with NVIDIA’s CUDA computing platform makes it the preferred choice for many Chinese companies and applications, the report noted.

However, it is worth noting that despite the current export controls on China, Chinese companies still manage to acquire advanced NVIDIA GPU computing power for AI and high-performance computing through intermediaries and by renting cloud service servers from companies like Google and Microsoft. This is a primary reason which prompts the U.S. to tighten restrictions.

Additionally, the U.S. might extend export restrictions to other Asian countries, such as Malaysia, Indonesia, Thailand, and potentially overseas Chinese companies. However, due to the complexity of these measures, effective implementation poses significant challenges, according to the report.

TrendForce notes in April that the extension of export controls now includes not only the previously restricted AI chips from NVIDIA and AMD, such as the NVIDIA A100/H100, AMD MI250/300 series, NVIDIA A800, H800, L40, L40S, and RTX4090, but also their next-generation successors like NVIDIA’s H200, B100, B200, GB200, and AMD’s MI350 series.

In response, HPC manufacturers have quickly developed products that comply with the new TPP and PD standards, such as NVIDIA’s adjusted H20/L20/L2, which remain eligible for export.

Read more

• [News] NVIDIA Reportedly Facing Price Cut Pressure of H20 Chip in China Amid Competition with Huawei
• [News] NVIDIA’s China-Exclusive H20 to Begin Pre-sales Next Month

(Photo credit: NVIDIA)

As former U.S. President Trump has brought up the topic that Taiwan should pay for protection, raising concerns on cross-strait issues at the same time, Taiwan’s semiconductor industry and its foundry leader, TSMC, have once again draw market attention. Some scholars believe that the U.S. must realize that Taiwan is not a competitor but an important fabless partner, which could not only play a crucial role in advancing the development of next-gen technologies in the U.S., but help the nation to win the U.S.-China tech battle.

Citing Chien-Huei Wu, a research fellow of the Institute of European and American Studies at Academia Sinica, a report from Technews points out that while profits for foundries have significantly increased in the past five years, it is the IC design sector that gains the largest share of value in the entire supply chain. The benefits to foundries are not particularly significant.

Wu further notes that when NVIDIA, AMD, and Apple place orders, TSMC provides the best service, which indicates that the semiconductor industry in Taiwan is not the only major beneficiary, but U.S. IC companies and the American semiconductor industry as a whole.

U.S. Emerges as the Biggest Winner while Outsourcing Wafer Manufacturing to Taiwan

“The U.S. must recognize that without Taiwan providing better services and designing better processes to overcome limitations, its semiconductor industry would find it difficult to make progress continuously,” said Wu.

Wu believes that the relationship between Taiwan and the U.S. involves high-level cooperation in the supply chain, with each playing its respective role. This is why TSMC founder Morris Chang has repeatedly emphasized that TSMC is a service provider, a trustworthy and reliable partner.

Additionally, the outsourcing of the semiconductor industry is an inevitable trend. Look back in history: the industry had moved from Europe and the U.S. to Japan. Then the Reagan administration sought to suppress the Japanese semiconductor, leading to the signing of the U.S.-Japan Semiconductor Agreement in 1986, while TSMC was founded the following year (1987).

And the history seems to repeat itself. Even after more than forty years, the U.S. has struggled to prevent the industry’s relocation. Now, Taiwan and South Korea have taken the lead in the semiconductor manufacturing business, and there are reasons for this.

Wu further explains that the work cultures between the Europe, the U.S., and Asia are entirely different, and these differences in technology, personnel, and culture will all increase manufacturing costs. Furthermore, these hidden costs will have to be absorbed by American consumers, which will inadvertently increase U.S. inflation. The situation not only harms American consumers but also hinders the country in the competition of next-generation advanced technologies.

The U.S. May be More Concerned about TSMC’s Tech Reaching China than Whether the Wafers are “U.S.-made.”

Currently, the market is concerned that Trump might impose a 10% tariff on all imported products if he takes office. According to Wu, the purpose of the tariff is to protect American manufacturers, making consumers more inclined to choose American-made products at the same price level.

“But the premise is that the quality must be the same, and that there must be existing American manufacturers.” He believes that the U.S. currently lacks manufacturers of equivalent quality, as U.S. tech giants, including Intel, place orders with TSMC, which indicates that the tariff may not have the intended effect. If an additional 10% tariff is imposed, products from major companies like NVIDIA, AMD, and Intel will become more expensive, increasing U.S. inflation, which benefits no one.

If this happens, Wu suggests that TSMC could organize manufacturers into a lobbying group, including key players from both Taiwan and the U.S., to expand the impact of the tariffs to the American semiconductor industry.

Furthermore, he notes that the Taiwanese government must work more closely with the industry, acting as a mediator in times of conflict. With both sides maintaining a unified stance, the government should prioritize industry interests in its communications.

In addition to tariffs, the U.S. government is also very concerned about the share of TSMC’s high-end chips going to China, especially with the escalation of the U.S.-China trade war.

Regarding this, Wu comments that as Taiwan’s current technological controls on the flow of technology to China are still in early stages, it will be difficult to convince the U.S. that it is a reliable partner and service provider. Therefore, the Taiwanese government must enhance its efforts to control the flow of high-tech, talent, products, or intellectual property to China, as there is still significant room for improvement.

How Should Taiwan’s Semiconductor Industry Handle U.S.-China Geopolitical Tensions?

Regarding Trump’s views on cross-strait politics, Wu analyzes that he is different from typical Western politicians, as Trump is known for his transactionalism, unilateralism, and personal connections. Namely, he is more concerned with “What’s in it for me (Trump)?” or “What’s in it for the U.S.?”

From Trump’s perspective, maintaining order is no longer an U.S. obligation. Taiwan must protect its own country with its own efforts, including maintaining a certain defense budget.

Additionally, it is worth watching whether Trump’s chip policies aim to relocate the entire semiconductor supply chain from Taiwan to the U.S. Although this poses practical difficulties and is economically unfeasible, Taiwanese government and the semiconductor industry must prepare for the worst-case scenario, and starts policy planning in advance.

Read more

• [News] With Trump’s Odds of Winning Rising, Concerns Arise on Potential Adjustments in Chip Subsidies
• [News] TSMC Maintains Overseas Expansion Strategy, Unfazed by Geopolitical Disruptions

(Photo credit: TSMC)

According to a previous report from Bloomberg, Chinese 3D NAND Flash giant YMTC recently filed a lawsuit against American memory giant Micron in California, accusing Micron of infringing on 11 of its patents related to 3D NAND Flash and DRAM products. YMTC is requesting the court to order Micron to stop selling the infringing memory products in the United States and to pay patent royalties.

Established at the end of 2016 in Wuhan, YMTC is a major Chinese manufacturer of memory (DRAM) and flash memory (NAND Flash), supported by significant investments from the “Big Fund.” It has become a representative enterprise in China’s effort to build a local chip supply chain. However, in October 2022, the U.S. Department of Commerce added YMTC to the Entity List, preventing it from obtaining advanced equipment from U.S. companies to manufacture 3D NAND chips with 128 layers or more.

Before facing U.S. export controls, YMTC’s 128-layer 3D NAND chip products had already entered Apple’s supply chain and received technical and quality certification from Apple. At that time, Apple reportedly hoped to use YMTC’s chips not only for cost considerations but also to prevent flash memory from being overly concentrated in the hands of Samsung, SK Hynix, and Micron.

The report from Tom’s hardware states that YMTC’s current allegations assert that Micron’s 96-layer (B27A), 128-layer (B37R), 176-layer (B47R), and 232-layer (B58R) 3D NAND Flash products, as well as some DDR5 SDRAM products (Y2BM series), infringe on 11 of YMTC’s patents or patent applications filed in the United States.

Notably, last November, YMTC also filed a lawsuit against Micron and its subsidiaries in the U.S. District Court for the Northern District of California, accusing them of infringing on eight U.S. patents related to 3D NAND Flash. Additionally, per a report from South China Morning Post on June 7th of this year, YMTC filed a lawsuit in California, accusing the Denmark-based consulting firm Strand Consult, funded by Micron, of spreading false information that damaged YMTC’s market reputation and business relationships.

Industry sources cited by the Commercial Times also note that in recent years, China’s technological capabilities have significantly improved, and companies have been actively applying for patents domestically and internationally. With the support of the Chinese government, they have also started to frequently engage in patent litigation. Last year, Chinese courts received 5,062 technical intellectual property and monopoly cases.

Read more

• [News] U.S. Department of Commerce Introduces New Regulations to Restrict China from Training AI Using U.S. Cloud Services
• [News] New Focus in the US-China Tech War! Can China Overtake with Silicon Photonics?

(Photo credit: YMTC)

As geopolitical tensions rise, a significant increase in orders from China for TSMC was seen last quarter. According to a report from Economic Daily News, as the U.S. presidential election countdown continues, both party candidates agree on expanding semiconductor export controls to China. Consequently, Chinese companies are stockpiling chips, causing a surge in rush orders for TSMC.

Regarding concerns that Chinese customers seem to be increasing their orders in response to potential future export controls or tax issues, TSMC did not elaborate much during its previous earnings call. They only mentioned that the increase in orders from Chinese customers was mainly due to high-performance computing (HPC) applications.

Concerning U.S. export control issues, TSMC’s management reiterated throughout 2023 earnings calls that the company will comply with all rules and regulations while serving all customers.

Recently, the proportion of TSMC’s orders from China has risen rapidly. The company’s latest financial report shows that in Q2, North America remained the largest market by customer headquarters location, accounting for 65%. The Chinese market, however, surged to 16%, up from 9% in the first quarter and 12% in the same period last year, replacing the Asia-Pacific region as the second-largest regional market. The Asia-Pacific region’s share fell to 9%, Japan remained at 6%, and the remaining share came from the EMEA region.

The sources cited by the report further indicate that as semiconductor export controls to China tighten, many Chinese companies not yet blacklisted under these controls are proactively placing orders and stockpiling goods, especially for advanced processes below 5 nanometers, which are highly valuable.

The sources noted that the current market atmosphere is quite similar to when Huawei’s HiSilicon excessively stocked up before export controls were imposed. This reflects the active development of Chinese companies in the AI field, not only renting computing power from major U.S. companies but also stockpiling chips and equipment on a large scale.

Read more

• [News] US Reportedly Targeting China’s Mature Semiconductor Processes Next
• [News] Taiwanese Foundries and Networking Companies to Benefit amid Escalating US-China Tech War

(Photo credit: iStock)

TSMC continues to showcase its drive in the advanced packaging market. According to a report from MoneyDJ, TSMC has forecasted that CoWoS will remain in high demand through 2025, with potential for capacity to double in two consecutive years. TSMC is also entering the FOPLP (Fan-Out Panel-Level Packaging) space, a technology that OSATs and panel manufacturers have been developing for years, with a goal to launch it in three years.

TSMC’s stronghold on advanced packaging technologies has raised market concerns that OSATs may see their opportunities diminished as TSMC consolidates its market position.

During its earnings call, TSMC introduced “Foundry 2.0,” a new definition for the semiconductor manufacturing industry. This expanded definition now includes not only foundry services but also packaging, testing, photomask production, and other integrated component manufacturing, excluding memory production. TSMC further stated that this new definition will better reflect the company’s growing market opportunities and that the company will focus solely on cutting-edge back-end technologies.

Under this new definition, TSMC estimates the Foundry 2.0 industry will be nearly USD 250 billion in 2023, up from the previous estimate of USD 115 billion. With this new scope, the industry is expected to grow by 10% annually in 2024. Additionally, TSMC’s market share in Foundry 2.0 (logic semiconductor manufacturing) for 2023 is revised to 28%, with expectations for continued growth in 2024.

The sources cited by the report has pointed out that, observing TSMC’s process advancements, the 3nm process began mass production in 2022, with the 2nm process set for 2025, indicating a lengthening of the development cycle to three years.

Regarding the aforementioned nodes, advanced packaging may help enhance performance, reduce costs, and has the advantage of binding high-end products from top-tier clients. Additionally, the investment required for advanced packaging is significantly smaller compared to frontend technologies, making it a crucial area of focus.

The report continues to note that TSMC currently retains the majority of major CoWoS orders and collaborates with OSATs in the WoS segment. However, TSMC has reportedly yet finalized its CoW segment outsourcing orders.

Meanwhile, AMD and NVIDIA have reportedly turned to Amkor and ASE’s subsidiary, Siliconware, for CoWoS-related products, focusing on cost-sensitive high-performance products. Amkor is expected to supply about 70,000 to 80,000 units annually this year, while Siliconware can provide about 50,000 to 60,000 units.

Additionally, OSATs have been confined to mature IC FOPLP technologies for the past 7-8 years, whereas TSMC has announced it will launch its own FOPLP technology in three years.

TSMC’s planned FOPLP is a rectangular CoWoS-L concept, offering advantages in low unit cost and large-size packaging. However, it faces physical limitations such as issues in coating, spinning, and warping. TSMC, backed by a substantial equipment and materials supply chain, is well-positioned to address these challenges.

As for opportunities for other OSATs, the advanced packaging market is actually vast and can be broadly categorized into flip-chip, fan-out, fan-in, 2.5D/3D, and embedded die packaging.

These technologies can be integrated, and OSATs are not limited to TSMC’s top-tier Info, CoWoS, or SoIC technologies. OSATs can offer more cost-effective advanced packaging solutions compared to fabs, providing competitive alternatives in terms of cost and performance.

Read more

• [News] TSMC Pushes for FOPLP Mass Production by 2027, Reportedly Eyeing on Innolux’s Plant
• [News] TSMC Introduces “Foundry 2.0” to Include Packaging, Testing, Mask making and Others

(Photo credit: TSMC)