American company, Marvell Technology said it has officially released a data center chip based on TSMC's 3-nm process. Here the company's industry-first silicon building blocks in this node include 112G XSR SerDes (serializer/deserializer), Long Reach SerDes, PCIe Gen 6 / CXL 3.0 SerDes and 240 Tbps parallel chip-to-chip interconnect.
According to Marvell, SerDes and parallel interconnects act as high-speed lanes in the chip for exchanging data between chips or silicon components inside the chiplet. Together with 2.5D and 3D packaging, these technologies will remove system-level bottlenecks to advance the most complex semiconductor designs. SerDes also help reduce costs by reducing pins, traces, and board space. Where a rack in a hyperscale data center may contain tens of thousands of SerDes links.
On April 20, GlobalFoundries accused IBM of illegally disclosing it's trade secrets to Intel and Japan company Rapidus.
"IBM unfairly received potentially hundreds of millions of dollars in licensing revenue and other benefits," GF said in a statement lawsuit. And GF will vigorously defend its investments in technology against those who violate them, as IBM has repeatedly demonstrated.
Since GF received "exclusive license and disclosure rights" to the unit's chip-making technology after acquiring IBM's microelectronics business in 2015.
On April 21, Esperanto Technologies announced that it has ported and is running a series of generative AI molds on its low-power RISC-V hardwares. Esperanto aims to provide access to researchers in the RISC-V community to help “democratize AI” and accelerate the development of RISC-V generative AI technologies.
It is a part of Esperanto's strategy to extend the benefits of RISC-V technology in AI and general-purpose applications from the cloud to the edge. For example, for large language models (LLMs), where using low-power hardware and reducing total cost of ownership compared to existing products are important.
On April 20, Tesla released its financial results for the first quarter of 2023. Total revenue was US$23.329 billion, an increase of 24% year-on-year (US$18.756 billion) over the same period last year, and net profit attributable to ordinary shareholders was US$2.513 billion, down 24% from the same period last year ($3.318 billion), which exceeded the expectations of many analysts in the market.
The reduction in profits is obviously related to Tesla’s previous price cuts. At present, The overall gross profit margin has dropped to 19.3%, which is lower than analysts’ average estimate of 22.4%. At the same time, the operating profit margin in the first quarter was only 11.4%, much lower than the 16% in the previous quarter and 19.2% in the same period last year. Faced with this situation, Musk said that Tesla will sell more cars at a lower profit and reap profits in the future. This may imply that a new round of price cuts by Tesla is coming.
Recently, Arm announced that it has reached a cooperation with Intel Foundry Services (IFS) for co-developing new processor IP for the Intel 18A process, and optimizing the power consumption, packaging area, performance and cost of the process. This also symbolizes that the foundry of Arm IP and chips has ushered in the next advanced technology participant after TSMC and Samsung.
Changxin Memory Technologies Inc (CXMT), which is currently China's largest maker of DDR4 and LPDDR4 memory chips, is seeking an IPO at a valuation of more than $14.5 billion, according to Bloomberg.
According to people familiar with the matter, Changxin Memory plans to list on the Sci-Tech Innovation Board at a valuation of no less than 100 billion yuan. It is in the process of picking underwriters and the size of the IPO has not been finalized. Also, the plan is preliminary and may change based on market fluctuations.
According to BusinessKorea, Samsung has begun to reduce the production of NAND flash memory, mainly focusing on its factory in Xi'an, China. According to market research firm TrendForce, Samsung will adjust its NAND production capacity to 620,000 12-inch wafers per month in Q2, 2023. Compared with the Q4 of last year, the figure was 5.34% lower.
According to Taiwan Business Times, it is expected that TSMC’s 2nm mass production plan will remain unchanged in 2025, and AMD will become an important customer of 2-nm process after Apple and Intel.
AMD officially launched the development of the new Zen 6 CPU core architecture in Q1, 2023. According to the recruitment information released by AMD on the recruitment software LinkedIn, it is confirmed that Zen 6 will adopt TSMC’s 2nm process, and the CPU launch schedule should be delayed after 2026.
©2012 - 2024 KUKE ELECTRONICS LIMITED All rights reserved
