HBM4, the sixth generation of HBM, is poised to become the key to breakthroughs in computing power for next-generation CSPs (Cloud Service Providers). According to a report from Commercial Times citing Global Unichip Corp. (GUC), to support the development of HBM4, their semiconductor IP (Intellectual Property) is already prepared and awaiting CSP manufacturers to advance their manufacturing processes.

GUC pointed out that if future clients need to integrate general-purpose HBM4 into ASICs (Application-Specific Integrated Circuits), GUC can provide assistance.

GUC further emphasized that its IP is ready for HBM4 development, waiting for CSPs to advance their manufacturing processes. Currently, the ASICs being mass-produced by CSPs still use HBM2 or HBM2e, while HBM3 is in the R&D stage.

The company candidly acknowledged that it cannot play any role at the moment and needs to wait for CSPs to adopt HBM4 on a large scale, taking cost considerations into account. When that time comes, GUC expects to assist CSPs in designing their solutions.

Currently, SK hynix has the technological capability for the general-purpose base die used in HBM4. However, when moving to more advanced processes like 5nm or beyond, external design service providers will be required.

Industry sources cited by Commercial Times believe that the pace of advancements in computing power is accelerating.

For instance, Google’s sixth-generation TPU, expected to be launched by the end of this year, is already based on TSMC’s 4nm process and designed on the Arm architecture.

Similarly, Meta’s upcoming MTIAv2 is built on TSMC’s 5nm process. The trend toward developing in-house chips is characterized by lower power consumption and larger memory capacities.

Read more

• [News] TSMC Reportedly Secures Another AI Opportunity, Winning Order from SK Hynix
• [News] SK Hynix Disclosed Details Regarding HBM4e, Reportedly Integrating Computing and Caching Functionalities

(Photo credit: GUC)

According to a report from TechNews, after Innolux shut down its 5.5-generation plant last year, it initially planned to sell the facility to memory giant Micron. However, TSMC successfully acquired Innolux’s 4th Plant in Tainan and its associated facilities, with a transaction value of NTD 17.14 billion.

Despite this, industry sources cited by TechNews have hinted that Micron is still moving forward with plans to establish a facility in Tainan. They are reportedly negotiating with Innolux regarding the Tainan site and have begun subsequent planning.

Reportedly, it is known that Micron had previously approached AUO to inquire about the Tainan color filter fab, but this is still considered to be in the site-selection phase, with Longtan also mentioned as a possible location.

Given the high demand for Micron’s HBM products and persistent rumors about expanding in Taiwan, new facility construction seems necessary to accelerate HBM market penetration.

Additionally, Micron’s Taiwan Chairman, Donghui Lu, has publicly stated that Taiwan is a crucial part of Micron’s global advanced process and packaging strategy. Besides expanding in Taiwan and Japan, Micron is also considering further expansion in the United States.

Regarding inquiries about Micron, Innolux has stated that it does not comment on market rumors.

Industry source cited by TechNews have anticipated that Innolux’s continued reduction in capacity is expected, though the timeline for shutting down facilities remains undecided.

On August 15th, TSMC officially announced the acquisition of Innolux’s 5.5G manufacturing facility in Tainan, Taiwan, for NTD 17.14 billion.

Read more

• [News] TSMC Acquires Innolux’s Plant in Southern Taiwan to Expand Advanced Packaging Capacity
• [News] Innolux Targets to Begin FOPLP Mass Production by Year-End

(Photo credit: Micron)

According to a report from UDN, TSMC held a groundbreaking ceremony yesterday for its Dresden, Germany plant, offering a significant boost to the EU’s efforts to stabilize its chip supply.

TSMC Chairman C.C. Wei led a team of top executives at the event, joined by key officials including German Chancellor Olaf Scholz. European Commission President Ursula von der Leyen also attended, bringing with her the announcement that the EU has approved a EUR 5 billion subsidy for the Dresden plant.

TSMC announced last August that it would partner with Bosch, Infineon, and NXP Semiconductors to establish the European Semiconductor Manufacturing Company (ESMC) in Germany.

The joint venture will construct a 12-inch wafer plant, with TSMC holding a 70% stake, while Bosch, Infineon, and NXP each hold 10%. Construction is planned to start in the second half of this year, with mass production expected by the end of 2027.

The planned fab is expected to have a monthly production capacity of 40,000 12-inch wafers on TSMC’s 28/22 nanometer planar CMOS and 16/12 nanometer FinFET process technology.  TSMC will be responsible for the plant’s operations.

Following the U.S.-China tech war, the EU passed the “Chips Act” to fully support the development of the semiconductor industry, attracting key investments from companies such as TSMC, Intel, Belgium’s IMEC, GlobalFoundries, and GlobalWafers, all of which sought subsidies for their new European operations.

TSMC’s joint venture proposal, exceeding EUR 10 billion, stands as the largest global direct investment in Saxony’s history.

When C.C. Wei took the stage, he began by thanking the German government. He revealed that when he first met with the German Chancellor, he had prepared a polite speech to decline the offer of building a plant in Germany.

However, when the Chancellor mentioned that a budget had already been reserved for TSMC, Wei eventually found himself agreeing to the project.

C.C. Wei further highlighted that TSMC’s total investment in the German plant exceeds EUR 10 billion and is expected to create around 2,000 jobs.

He explained that the decision to locate the plant in Dresden was due to its proximity to TSMC’s customers and access to a large pool of talented individuals. Wei also pledged to continue recruiting and nurturing talent in the region, with the goal of making ESMC the most important semiconductor manufacturing hub in Europe.

Read more

• [News] TSMC Maintains Overseas Expansion Strategy, Unfazed by Geopolitical Disruptions
• [News] TSMC’s German Plant Reported to Break Ground in Weeks, Expected to Begin Production by Late 2027

(Photo credit: TSMC)

As global competition heats up in the AI sector, an emerging power has now joining the battlefield. Ola, an automotive manufacturer in India, plans to launch the country’s first in-house AI chip by 2026, which is based on ARM architecture, according to a report by Wccftech.

Though there are more details yet to be revealed, the report notes that Ola did highlight its key chip offerings, featuring the Bodhi series, which would be the nation’s first self-developed AI chips. The company’s product lineup also reportedly includes the Sarv-1 cloud-native CPUs and the Ojas edge AI chip.

When asked about the potential foundry partners in the future, Ola’s CEO Bhavish Aggarwal mentioned that the company plans to collaborate with a global tier I or II foundry, likely TSMC or Samsung, according to the report.

Ola’s AI lineup is expected to start with the Bodhi-1 AI chip, which is specifically designed for large-scale LLMs, with a focus on inferencing workloads, Wccftech suggests. Positioned as a low-to-mid-tier offering from Ola, the chip is said to be launched by 2026, followed by a more potent successor, the Bodhi-2, slated to be released in 2028.

According to Wccftech, it is worth noting that Ola also introduced an edge AI chip named Ojas, which is likely to be integrated into Ola’s next-generation electric vehicles. In addition, the Sarv-1, specifically designed for cloud computing, is expected to feature ARM Neoverse N3 cores, though this hasn’t been confirmed yet, the report states.

As the world’s fifth largest economy, India seems to be relatively slow in developing its own AI chips. China, the world’s largest developing country, has quite a long history in developing in-house AI chips.

Chinese tech giant Huawei is said to be testing its latest processor, the “Ascend 910C,” with internet companies and telecom operators recently. Reportedly, the company has informed potential customers that this new chip is comparable to NVIDIA’s H100 GPU, which cannot be directly sold in China.

On the other hand, Baidu’s foray into AI chips can be traced back to as early as 2011. After seven years of development, Baidu officially unveiled its self-developed AI chip, Kunlun 1, in 2018. T-Head, owned by Alibaba, introduced its first high-performance AI inference chip, the HanGuang 800, in September 2019.

Read more

• [News] Huawei Rumored to Launch New High-End AI Chip, Potentially Rivaling NVIDIA’s H100
• [News] Global Semiconductor Giants Eyeing India, Anticipating Three New Fabs Valued at USD 8-12 Billion

(Photo credit: Krutrim)

As per a report from Reuters, South Korea’s SK Telecom-backed AI chip startup Sapeon Korea and KT-invested startup Rebellions have officially announced their merger.

The combined entity is expected to be established by the end of 2024, with the merger projected to create a business valued at over KRW 1 trillion (roughly USD 750 million), hinting at a potential challenge to NVIDIA’s leading position in the AI chip market.

As competition in the AI chip market intensifies, Sapeon and Rebellions have formalized their merger. The merger, which was rumored in June according to the Korean Economic Daily, have now been officially agreed upon, with Rebellions’ co-founder and CEO Sunghyun Park set to lead the executive team of the merged entity.

Sapeon and Rebellions are two South Korean chip startups. Rebellions introduced the ATOM chip last year, Korea’s first NPU designed for large language model (LLM) data centers. Meanwhile, Sapeon launched its next-generation AI chip, the X330, in November, enhancing Korea’s competitiveness in the global AI semiconductor market.

The two companies previously noted that they view the next two to three years as a critical opportunity for South Korea to establish a presence in the global AI semiconductor market. Moreover, they also emphasized their commitment to accelerating the formation of the merged entity to capitalize on this pivotal period.

The report by TheElec, citing industry sources, noted that since SK Group will be the majority shareholder of the merged entity, it may prefer TSMC over Samsung Foundry, given that SK hynix and Samsung are rivals in memory chips. Currently, Sapeon uses TSMC as its foundry, whereas Rebellions collaborates with Samsung Foundry.

Read more

• [News] Korea’s First Semiconductor Unicorn in the Making: AI Chip Firms Sapeon and Rebellions Set to Merge

(Photo credit: Rebellions)