Starting from July 8^th, Samsung Electronics’ labor union, comprising over 28,000 workers, has initiated a three-day strike. However, as the dispute with Samsung escalates, the union declared earlier today that the workers plan to go on an indefinite strike, according to reports from Bloomberg and BBC.

On Monday, thousands of workers gathered outside Samsung’s chipmaking facilities south of Seoul to begin what was initially a three-day walkout aimed at securing improved pay. This marks the largest organized labor action in the conglomerate’s fifty-year history.

As the two parties are unable to reach a consensus, the three-day strike seems to further prolong. According to Bloomberg, the union announced on its website that as the management has shown no willingness to engage in dialogue, they have identified clear disruptions in production, and the company will regret this stance.

However, Samsung has challenged the assertion, saying that it will take necessary steps to prevent any disruptions in production lines. The company remains dedicated to conducting negotiations with the union in good faith, according to BBC’s report.

Regarding the impact of the Samsung strike, TrendForce analyzed that because semiconductor factories rely on automated production with low actual manpower requirements, the strike did not affect Samsung’s production. Even if the strike is extended, the current assessment is that there will still be no significant impact.

The strike does come at a sensitive time, though, as the tech giant has a busy schedule ahead. It is going to unpack its latest Galaxy lineup in Paris soon, while the company is also working on the qualification of its HBM3e products with NVIDIA.

Read more

• [News] Samsung, SK Hynix and Micron Ramp Up HBM Production, Reportedly Doubling Output Next Year
• [News] Samsung’s HBM3e Reportedly Passed NVIDIA’s Certification Though It Denies the Rumor

(Photo credit: Samsung)

According to a report from Commercial Times, SK Hynix, Samsung, and Micron, the world’s top three memory manufacturers, are actively investing in high-bandwidth memory (HBM) capacity expansion plans. Industry sources cited by the same report estimate that by 2025, the additional production will reach approximately 276,000 units, bringing the total capacity to 540,000 units, an annual increase of 105%.

Regarding the latest developments in HBM, TrendForce indicates that HBM3e will become the market mainstream this year, with shipments concentrated in the second half of the year.

Currently, SK Hynix remains the primary supplier for HBM, along with Micron, both utilizing 1beta nm processes and already shipping to NVIDIA. Samsung, using a 1alpha nm process, is expected to complete qualification in the second quarter and begin deliveries mid-year.

Regarding major memory players’ expansion plans on HBM, Samsung is gradually upgrading its Pyeongtaek facilities (P1L, P2L, and P3L) in South Korea to be used for DDR5 and HBM. Meanwhile, the Hwaseong facilities (Line 13, 15, and 17) are being upgraded to the 1α process, retaining only a small portion of capacity at the 1y/1z process to meet the demands of specialized industries such as aerospace.

SK Hynix produces HBM at its M16 production line in Icheon, South Korea, and is upgrading its M14 production line to the 1α/1β process to supply DDR5 and HBM products. Additionally, after receiving clearance from the U.S. government, its Wuxi plant in China is actively upgrading from the 1y/1z process to the 1z/1α process for producing DDR4 and DDR5 products.

Micron’s HBM production is conducted at its Hiroshima plant in Japan, with capacity expected to increase to 25,000 units in the fourth quarter of this year. In the long term, Micron plans to introduce EUV processes (1γ, 1δ) and build a new cleanroom.

In the short term, Micron will utilize its production capacities at the Linkou and Taichung plants in Taiwan, increasing the proportion of 1β process. By the end of 2025, the total production for HBM is expected to reach around 60,000 units.

According to Commercial Times, the HBM production volume of the world’s top three manufacturers will maintain high growth for two consecutive years, with the global total wafer production volume expected to reach approximately 540,000 units per month by the end of 2025.

Read more

• [News] Samsung Establishes New HBM Team to Advance HBM3, HBM3e and HBM4 Development
• HBM3e Production Surge Expected to Make Up 35% of Advanced Process Wafer Input by End of 2024, Says TrendForce

(Photo credit: SK Hynix)

After Samsung announced a major breakthrough in 2nm, securing the first batch of orders from Japanese AI company Preferred Networks, its rival TSMC is also advancing. According to reports from Wccftech and ET News, TSMC is set to begin trial production of 2nm chips next week, which would reportedly be used in the upcoming iPhone 17 lineup in 2025.

The reports note that the trial production will be conducted in TSMC’s Baoshan Plant in Hsinchu, northern Taiwan, as facilities have been brought in during the second quarter. The iPhone 17 lineup is rumored to be the first to feature TSMC’s 2nm chips. Following that, the chips will likely be used in the 14-inch and 16-inch MacBook Pro models.

According to a previous report by MoneyDJ, TSMC’s 2nm production bases are located in Hsinchu Science Park and Kaohsiung, southern Taiwan, while the mass production is expected to kick off in Hsinchu first, with an initial monthly capacity of approximately 30,000 to 35,000 wafers.

Apple and TSMC share a long history of partnership, as the smartphone giant’s A17 Pro, M3 and M4 chips are all manufactured with TSMC’s 3nm node. As TSMC reportedly plans to enter 2nm trial production next week for Apple’s M5 chip, the company’s target for 2nm to enter mass production in 2025 would be on schedule, Wccftech notes.

According to Wccftech, The M5 chip, compared to its predecessor M4, is expected to have performance increase of 10 to 15 percent and a power consumption reduction of up to 30 percent compared to current 3nm-based chips.

Regarding the progress of other semiconductor heavyweights on 2nm, Samsung is said to commence mass production of 2nm chips for mobile devices by 2025. The initial SF2 2nm process will be ready next year, followed by an enhanced version, SF2P, in 2026. Its latest 2nm process, SF2Z, has incorporated optimized backside power delivery network (BSPDN) technology, and will enter mass production in 2027.

On the other hand, Intel’s 20A manufacturing technology (2nm) is reportedly scheduled for launch in 2024, introducing two technologies: RibbonFET surround gate transistors and BSPDN.

Read more

• [News] The Imminent Arrival of 2-Nanometer Advanced Process
• [News] TSMC Reportedly Invests over USD 12.3 Billion in EUV, Advancing in 2nm Production

(Photo credit: Apple)

Samsung Electronics has received the first client for its 2nm process. According to the official press release from Samsung on July 9th, Samsung Electronics will provide turnkey semiconductor solutions using the 2nm process and the advanced 2.5D packaging technology Interposer-Cube S (I-Cube S) to Japanese AI company Preferred Networks.

Per a previous report by SamMobile, Samsung is set to commence mass production of 2nm chips for mobile devices by 2025. The initial SF2 2nm process will be ready next year, followed by an enhanced version, SF2P, in 2026. In addition, according to Samsung’s press release, its latest 2nm process, SF2Z, has incorporated optimized backside power delivery network (BSPDN) technology, and will enter mass production in 2027.

Preferred Networks was founded in 2014 and is in the field of AI deep learning development. The company has attracted significant investments from major Japanese industrial enterprises such as Toyota, NTT, and Fanuc. The order placed with Samsung’s foundry division for 2-nanometer AI chips also includes HBM and advanced packaging.

As per the official release, Junichiro Makino, VP and Chief Technology Officer (CTO) of Computing Architecture at Preferred Networks stated that as Samsung Electronics’ 2nm GAA process will significantly support Preferred Networks’ ongoing efforts to build highly energy-efficient, high-performance computing hardware that meets the ever-growing computing demands from generative AI technologies, especially large language models.

Driven by the strong demand from AI chips, Samsung expects the revenue of global chip industry to grow to USD 778 billion by 2028, according to Siyoung Choi, President and General Manager of the Foundry Business in Samsung, the report noted.

Read more

• [News] Samsung Unveils Next-gen 2nm Node Roadmap; SF2Z with Backside Power Delivery to Enter Production in 2027
• [News] Samsung Secures 2-Nanometer Order from Japanese AI Startup Preferred Networks

(Photo credit: Samsung)

As the demand for AI chips surges, orders for thermal compression (TC) bonders, which play a critical role in HBM (high-bandwidth memory) manufacturing, are also heating up.

To further gain market momentum, South Korean chip packaging equipment manufacturer Hanmi Semiconductor plans to launch 2.5D big die TC bonder models in the second half of 2024, while increasing its annual sales target for this year to 650 billion won (USD 471 million), according to the latest report by the Korea Economic Daily.

Citing Kwak Dong-shin, vice chairman and CEO of Hanmi Semiconductor, the report notes that the company eyes strong revenue growth in the next two years, projecting 1.2 trillion won (USD 870 million) in sales for 2025, and 2 trillion won (USD 1.45 billion) for 2026.

TC bonders play a pivotal role in HBM production by employing thermal compression to bond and stack chips on processed wafers, thereby significantly influencing HBM yield. According to the report, Hanmi plans to introduce several upgraded models in the next two years, including 2.5D big die TC bonders in the second half of this year, mild hybrid bonders in the latter half of 2025, and hybrid bonders in 2026.

Memory giants have developed their own ecosystems to secure TC bonders’ supply. The report notes that Hanmi has been providing its TC bonders to SK hynix, while the latter is a major HBM supplier to Nvidia. In addition, the company also entered into a 22.6 billion won agreement with Micron in April.

Whether in the near future, Hanmi Semicodutor would be able to finalize similar contracts with Samsung, another memory heavyweight, remains to be seen. For now, Samsung sources its equipment from Japan’s Toray and Sinkawa, as well as its subsidiary SEMES.

Hanmi Semiconductor produces TC bonders at its six factories located in Incheon, where its headquarters are situated. The report indicates that it aims to increase the capacity of its newest, the sixth factory from 264 units of TC bonders annually this year to 420 units next year, which makes it the largest annual capacity for TC bonder producers worldwide.

Read more

• [News] Rising HBM Production by Samsung and SK Hynix Energizes Korea’s TC Bonder Industry
• [News] ASMPT Reportedly Provided Demo TC Bonder to Micron for HBM Production

(Photo credit: Hanmi Semicondutor)