Recently, the semiconductor industry has seen three acquisitions.

Qualcomm announced the acquisition of Foundries.io, a leading provider of IoT and cloud development security services, with its flagship product FoundriesFactory being well-regarded in the industry. Microchip Technology acquired Neuronix AI Labs to enhance its capabilities in developing highly efficient AI edge solutions on Field-Programmable Gate Arrays (FPGA). Runjing Chemical, a developer of Tetramethylammonium Hydroxide (TMAH), announced the acquisition of two plants in China from Sumitomo Chemical, aiming to strengthen its competitiveness in the wet electronic chemical market in China.

• Runjing Acquired Sumitomo Chemical’s Two Plants

On April 16th, Runjing Chemical announced that it has acquired two plants in China from Sumitomo Chemical.

Runjing disclosed that it has reached an agreement on the full acquisition of Sumitomo Chemical’s subsidiaries, Sumitomo Chemical Electronic Materials (Hefei) Co., Ltd. and Sumitomo Chemical Electronic Materials (Chongqing) Co., Ltd., and both parties have completed the signing of the equity transfer contract.

Upon meeting the relevant items stipulated in the equity transfer contract, Hefei Sumitomo Chemical and Chongqing Sumitomo Chemical will become wholly-owned subsidiaries of Runjing.

Sumitomo Chemical, founded in 1913, has been engaged in the production of process chemicals for flat panel displays (FPD) in China since 2009. The two acquired subsidiaries focus on providing etchants, developers, and peelers with stable quality for downstream customers.

Founded in 2008, Runjing is a manufacturer of TMAH developer and offer products to major global panel manufacturers such as Samsung, LG, BOE, HKC, CSOT, and Tianma. In 2020, Runjing Technology established XINKE Electronic Materials in Hefei to produce high-purity semiconductor-grade products including hydrogen peroxide, ammonia water, and isopropanol, which has begun product introduction at its 12-inch fabs.

Runjing stated that through this acquisition, it will take over Sumitomo Chemical’s technological advantages and business network, enable it to quickly expand its product portfolio (Etchants, peelers, and CF developers, etc), and provide customers with a variety of comprehensive solutions, thereby enhancing Runjing’s competitiveness in the wet electronic chemical market in China.

• Qualcomm Acquired Foundries.io

Qualcomm recently disclosed its acquisition of Foundries.io in a press release regarding a WiFi product launch. Foundries.io is known as an open-source cloud-native platform provider, specializing in simplifying the complexity of developing Linux-based IoT and edge devices, and updating these devices.

Foundries.io was co-founded by executives and engineers. Headquartered in Cambridge, UK, it is recognized as a pacesetter in its field. The company’s  cloud-native DevOps products fitted for interconnected embedded security devices can expedite time-to-market and reduce costs for OEMs across industries.

Foundries.io reportedly established close relationships with ARM and its chip partners, adopting ARM SystemReady technology. It has developed Linux distribution that is the first to fully integrate Project Cassini and provide commercial support for it. Project Cassini is an open, collaborative, and standards-based program by ARM.

• Microchip Acquired Neuronix AI Labs

According to Microchip Technology’s press release, Microchip Technology recently acquired Neuronix AI Labs to enhance its capabilities in developing high-efficiency, AI-supported edge systems on FPGA. Neuronix AI Labs delivers neural network sparse optimization technology, which can reduce power consumption, size, and computation for tasks such as image classification, object detection, and semantic segmentation while maintaining high accuracy.

Microchip’s mid-range PolarFire® FPGA and SoC are already ahead in the industry in terms of low power, reliability, and security features. Acquiring such neural network sparse optimization technology will enable Microchip to develop large-scale edge deployment components with high cost-effectiveness and efficiency, which are specifically designed for computer vision applications on systems requiring low cost, small size, and low power consumption, thus enabling AI to exponentially increasing ML processing capabilities on low-end and mid-end FPGA.

The acquisition also gives our traditional Microchip MCU and MPU clients the ability to use FPGAs as accelerators through an easy-to-use compiler that will substantially improve their design productivity and system performance while shortening their time to market

Bruce Weyer, Vice President of Microchip’s FPGA Business Unit, said, “This acquisition will improve the efficiency of our FPGA and SoC deployed in intelligent edge systems utilizing AI/ML algorithms.” He added, “Neuronix technology, combined with our VectorBlox design flow, can enhance neural network performance efficiency and delivers outstanding GOPS/Watt performance in our  PolarFire FPGA and SoC with low power consumption. System designers can now build and deploy small hardware, which was previously difficult to achieve due to limitations of size, heat, or power.”

The neural network sparse optimization technology will allow non-FPGA designers to leverage the powerful parallel processing capabilities of industry-standard AI frameworks without deep knowledge of FPGA design processes.

Neuronix AI intellectual property, coupled with Microchip’s existing compilers and software PDK, can achieve AI/ML algorithms on customizable FPGA logic without the need for specialized knowledge of Register Transfer Level (RTL) or a deep understanding of underlying FPGA architecture. It also allows for dynamic updates and upgrades of CNNs without reprogramming the hardware.

Read more

• [News] Battle in AI PCs: Exploring Intel, AMD, and Qualcomm Chips’ AI Computing Power

(Photo credit: Qualcomm)

On April 15th, during the Member Representatives Symposium of the World Internet Conference, Jiang Tao, Vice President of Intel, and Mark Murphy, Executive Vice President and CFO of Micron, expressed their respective views on the theme “Interconnection, Mutual Benefit, and Common Prosperity — Building a Community of Shared Future in Cyberspace”, and discussed future development of international organizations and industry hot topics with other member representatives.

• Micron to Continuously Cultivate the Chinese Market and Jointly Shape the Future of the Global Semiconductor Industry

Having been rooted in China for over 20 years, Micron has established close cooperation with its customers.

On March 27, 2024, Micron announced that its new packaging and testing plant in Xi’an had officially broken ground, further strengthening the company’s unwavering commitment to operations, customers, and communities in China. Micron also announced at the groundbreaking ceremony that it will establish its first sustainable Center of Excellence (CoE) for packaging and testing in Xi’an.

In June 2023, Micron announced an additional investment of CNY 4.3 billion in Xi’an, inclusive of the construction of this new plant, introduction of new production lines, and the production of a wider range of product solutions, including but not limited to mobile DRAM, NAND, and SSD, with an aim to expand the existing DRAM packaging and testing capabilities in Xi’an.

Mark Murphy expressed that Micron is deeply rooted in China and has been committed to achieving multi-level development in China for a long time. Micron’s investments in China is a testament of its confidence in the Chinese market, strong semiconductor ecosystem, and high-quality talent. Down the road, it will continue to provide support for its customers and for the innovation in the broader semiconductor ecosystem in China.

• Intel: China Has many Conditions to Attract Global Enterprises

Currently, AI technology enjoys burgeoning growth, empowering various industries.

Intel’s VP Jiang Tao pointed out that China boast a vast market, resilient supply chain, numerous innovative application scenarios, and rich multi-level, high-quality talent resources, which are attractive to all global companies.

He expects that international organizations at the World Internet Conference can play a leading role in promoting AI technology and applications, consolidating industry consensus, and promoting full exchange among government, industry, academia, and relevant professionals in the latest developments and future trends of AI industry.

Read more

• [News] Following in NVIDIA’s Footsteps? Intel Reportedly Plans to Launch Chinese Version of AI Chips
• [News] Micron’s Financial Report Reveals High Demand for HBM in 2025, Capacity Nears Full Allocation

(Photo credit: Intel)

TrendForce releases the latest trends in memory spot prices. Due to sellers halting quotations, DDR3, DDR4, and DDR5 have all seen price increases. It is expected that prices will decline once quotations are fully resumed. On the other hand, with no strong signs of recovery in end-market demand for NAND Flash, inquiry interest remains subdued. Details below:

DRAM Spot Market:

In the spot market, sellers and module houses suspended quoting following Taiwan’s earthquake on April 3, leading to incremental price rises over several days. This upward trend applies to DDR3, DDR4, and DDR5 products. However, the spot market still lacks demand, and transactions have been limited in terms of quantity. TrendForce believes sellers will resume quoting very soon, and prices will shift down again as before. The average spot price of mainstream chips (i.e., DDR4 1Gx8 2666MT/s) rose by 0.93% from US$1.927 last week to US$1.945 this week.

NAND Flash Spot Market:

Suppliers are carrying on with their increase of contract prices for the mainstream 512Gb wafers by more than US$4, though inquiries have been sluggish as distributors are currently holding onto an excessive level of low-cost inventory, and that end market demand has also yet to resuscitate. TrendForce noticed that a number of suppliers are truncating with prices that are slightly below that of contract prices for March, which are generating some sort of pricing pressure. Spot prices of 512Gb TLC wafers have risen by 0.03% this week, arriving at US$3.765.

Apple currently uses TSMC’s 3-nanometer process for multiple chipsets. According to a report from wccftech, the iPhone 17 chipset will not adopt the 2-nanometer process. Instead, the A19 Pro chip, expected in 2025, is reportedly maintaining the 3-nanometer technology.

The same report suggests that the Apple A19 Pro chip is considering TSMC’s N3P process and may be featured in the iPhone 17 Pro and iPhone 17 Pro Max. TSMC aims to ramp up its 3-nanometer wafer capacity to 100,000 units by the end of 2024.

TSMC began trial production of 2-nanometer chips as early as June 2023. However, Apple’s A18 Pro chip for the iPhone 16 Pro and iPhone 16 Pro Max might use the N3E process. As for next year’s iPhone 17, its A19 Pro chip is likely to adopt TSMC’s N3P technology.

As per MoneyDJ’s report, it is currently expected that Apple’s iPhone 18 series, slated for release in 2026, will feature the first-ever 2-nanometer chip. Besides Apple, other 2-nanometer customers include Intel, with interest also anticipated from AMD, NVIDIA, and MediaTek. Looking at the process roadmap, this year’s iPhone 16 will use N3E, while next year’s model will adopt N3P. Thus, the first consumer product leveraging TSMC’s 2nm process is anticipated to launch in 2026.

Meanwhile, Apple is said to be striving to stay ahead of competitors and exploring alternative packaging technologies like 3DFabric with its foundry partner. A rumor from the account Yeux1122 has suggested that Apple is already delving into SoIC (Small Outline Integrated Circuit) packaging.

According to a report from South Korean media The Chosun Daily, Samsung Electronics is set to increase wafer input by approximately 30% this quarter at its NAND Flash production lines in Pyeongtaek, South Korea, and Xi’an, China. However, Samsung remains cautious about further production increases to avoid impacting NAND Flash price trends.

The report indicates that while Samsung’s NAND Flash production lines can exceed 2 million wafers in a quarter at full capacity, internal targets for wafer output in the second to fourth quarters are capped at 1.2 million wafers each, maintaining overall utilization rate at around 50%.

Market expectations cited in the same report suggest that Samsung Electronics will reaffirm its stance on reducing NAND Flash production during the first quarter earnings call later this month. In the previous earnings call, Samsung noted persistent high levels of NAND Flash inventory among major customers, necessitating continued aggressive production cuts.

The same report further indicates that SK hynix has set a quarterly cap of around 600,000 wafers for NAND Flash production, with overall utilization rate ranging between 50% and 60%.

As per TrendForce’s data, it has projected a strong 13–18% increase in Q2 NAND Flash contract prices, with enterprise SSDs expected to rise highest. Despite Kioxia and WDC boosting their production capacity utilization rates from Q1 this year, other suppliers have kept their production strategies conservative. The slight dip in Q2 NAND Flash purchasing—compared to Q1—does not detract from the overall market’s momentum, which continues to be influenced by decreasing supplier inventories and the impact of production cuts.

Read more

• [News] Memory Manufacturers’ Price Hikes Prompt Backlash, Module Suppliers Hesitate to Follow
• [News] Micron Expects 4-6% Quarterly DRAM Supply Impact Post Taiwan Earthquake

(Photo credit: Samsung)