Tom Spyrou

TSMC: Performance and Yields of 2nm on Track, Mass Production To Start In 2025 💡 In addition to revealing its roadmap and plans concerning its current leading-edge process technologies, #TSMC also shared progress of its N2 node as part of its Symposiums 2024. The company's first 2nm-class fabrication node, and predominantly featuring gate-all-around transistors, according to TSMC #N2 has almost achieved its target performance and yield goals, which places it on track to enter high-volume manufacturing in the second half of 2025. TSMC states that 'N2 development is well on track and #N2P is next.' In particular, gate-all-around nanosheet devices currently achieve over 90% of their expected performance, whereas yields of 256 Mb #SRAM (32 MB) devices already exceeds 80%, depending on the batch. All of this for a node that is over a year away from mass production. Meanwhile, average yield of a 256 Mb SRAM was around 70% as of March, 2024, up from around 35% in April, 2023. Device performance has also been improving with higher frequencies being achieved while keeping power consumption in check. Thanks again to Anton Shilov and AnandTech for the full article with more background and insights via the link below 💡🙏👇 https://lnkd.in/eHgpfX4k #semiconductormanufacturing #semiconductors #semiconductor #tech #technology #innovation #chip #chips #integratedcircuits #ic #foundry #taiwan #geopolitics #asml #thenetherlands #europe #euv #computer

125

Check out my new blog on using our #radhard #MCU to reconfigure #Space deployed #FPGA. Please reach out if you have interest and we can provide specific details on how to deploy

13

Great article with insights by Bernard Murphy at SemiWiki.com, "Arteris Frames Network-On-Chip Topologies in the Car." "... while there is plenty of complexity in evolving car architectures and the consequent impact on subsystem chip/chiplet designs, connected on-chip through both coherent and non-coherent networks, the intent behind these systems is quite clear. We just need to start thinking about total car system architecture rather than individual chip functions." #Arteris #NetworkOnChip #Automotive https://bit.ly/3UiKqls

25

Arm Is Rebuffed by #Intel After Inquiring About Buying Product Unit - Intel’s rapid decline has spurred takeover speculation - Arm is interested in product group, not factory operations Arm Holdings Plc approached Intel Corporation about potentially buying the ailing chipmaker’s product division, only to be told that the business isn’t for sale, according to a person with direct knowledge of the matter. In the high-level inquiry, Arm didn’t express interest in Intel’s #manufacturing operations, said the person, who asked not to be identified because the discussions were private. Intel has two main units: a product group that sells #chips for #personalcomputers, #servers and #networking equipment, and another that operates its factories. Combining with Intel would help Arm’s reach and kick-start a move toward selling more of its own products. The company currently licenses #technology and designs to customers, who then turn them into complete components. Its client list includes the biggest names in #technology, such as Amazon.com Inc., Qualcomm Inc. and Samsung Electronics Co. Thanks again to Ian King and Bloomberg News for the full article with more background and insights via the link below 🙏💡👇 https://lnkd.in/eCDvrC_h #semiconductorindustry #semiconductors #semiconductor #tech #it #chip #chips #innovation #technology #semiconductormanufacturing

130

6 Comments

a very nice article on chiplets, delving into physical connectivity (bridges like EMIB and CoWoS-R, standards BOW and UCIe), socket concept, thermal issues, interfacing it is a summary article, so doesn't dog deep but gives a comprehensive summary https://lnkd.in/gPfJFd6g

32

2 Comments

HBM (High-Bandwidth Memory), which is based on DRAM products, is a crucial building block for data centers. The following article addresses some complex issues in building such products. https://lnkd.in/gSYcAxuN

1

Efficient Power Conversion Corp (EPC), headquartered in El Segundo, CA, specializes in gallium nitride on silicon (eGaN) power FETs and integrated circuits for power management applications, with recent validation of their GaN intellectual property rights by the US International Trade Commission (ITC). The ITC upheld EPC's patents and found Innoscience Technology Co Ltd guilty of infringing EPC’s foundational patent, potentially leading to a ban on importing their products into the USA. This ruling is a significant validation of nearly two decades of EPC's hard work and research, reflecting the importance of their intellectual property portfolio. The ITC's final determination is expected on November 5. Innoscience EPC - Efficient Power Conversion #semiconductor #intellectualproperty #industry #lawsuit #technology #GaN https://lnkd.in/di792VMW

7

Wrapping up #DAC61, I am curious about the decision to rebrand the event as "DAC, The Chips to Systems Conference," instead of the traditional "Design Automation Conference." This is not just a simple change of words — consider the keynote speakers. Last year, they were all #EDA luminaries:  Mike Ellow (recently appointed CEO of Siemens EDA), Lip-Bu Tan (former Cadence CEO), Alberto Sangiovanni-Vincentelli (founder of both Cadence and Synopsys) and Walden C. ( Wally ) Rhines - long time CEO of Mentor Graphics. In contrast, this year, the keynote speakers were Jim Keller (Tenstorrent), Gary Patton (Intel) and Alan Lee (ADI) - all "Chip to Systems" executives. Is this a coincidence? Does it reflect EDA companies taking a lesser role at DAC? Siemens EDA (Siemens Digital Industries Software) was a notable exception as a Diamond sponsor, investing in a sizable booth and team, and hosting guests for a Giants' game at Oracle Park. (The famous Denali party is long gone...) Or is it because part of Design Automation is increasingly being done within "Chips to Systems" companies? Just look at Jim Keller discussing the importance of #opensource and Alan Lee explaining that "Product Design is Complex." What do you think? #Designautomation #chipdesign

55

2 Comments

Excited to lead this effort on AI for Wireless Chip Design, as one of the three anticipated awardees of the $30 million dollar inaugural proposal call from National Semiconductor Technology Center (Natcast.org), established through the US Chips and Science Act by the Biden-Harris administration. AI-enabled RFIC design started as a small effort from our group (nearly unfunded) about four years ago. Without any prior work, we went through quite a bit of sweat, tears, and failed ideas, before we demonstrated (in 2022) the first examples of how AI can discover new architectures for wireless chips, and allow rapid designs on demand. This has now evolved into a substantial effort across academia, industry and government. Expectedly, a highly multi-disciplinary effort--immensely proud of my students, who had to learn multiple areas from complex chip design to electromagnetics to AI, and build the foundations from scratch. Congrats to Emir Ali Karahan, Zheng Liu, Jonathan Zhou, and all the members of the group. Excited to partner with Hossein Hashemi, Ioannis Savidis, Xuan 'Silvia' Zhang, Mengdi Wang, Raytheon, Cadence Design Systems, and Keysight Technologies. https://lnkd.in/eZFfnyPY #AI #RFIC Princeton University, Princeton Engineering, Princeton ECE, Princeton Research

293

44 Comments

Semiconductor Industry Association Outlines Industry Recommendations for the National Semiconductor Technology Center 🇺🇸💡 SIA released a set of recommendations to promote the success of the National Semiconductor Technology Center (NSTC), a critical public-private research consortium established by the CHIPS and Science Act. SIA’s recommendations synthesize the most up-to-date industry consensus and priorities. The CHIPS & Science Act included the most significant federal investment ever made in chip research and development (R&D), committing a substantial $13 billion across the Departments of Commerce and Defense. The cornerstone of the CHIPS R&D program is the NSTC, a more than $5 billion initiative tasked to “conduct research and prototyping of advanced semiconductor technology and grow the domestic semiconductor workforce to strengthen the economic competitiveness and security of the domestic supply chain” (15 U.S.C. § 4656(c)(1). Since passage of the CHIPS Act, leading stakeholders have issued several valuable reports detailing recommendations and priorities for the initiation and operation of the NSTC. This body of work represents an emerging industry consensus that should form the guiding basis for the CHIPS R&D Office (CRDO) and Natcast—the new, purpose-built, non-profit entity created to operate the NSTC Consortium—in the establishment and operation of the NSTC. A big thank you again to Semiconductor Industry Association for the full article with more background and insights including the full report as FREE PDF Download via the link below 💡🙏👇 https://lnkd.in/e5aBA-K8 #semiconductorindustry #semiconductors #semiconductor #technology #tech #it #chips #innovation #chip #geopolitics #subsidies #usa #talent

33

Intel Corporation will now be spinning off it's Intel Foundry division into a wholly owned subsidiary and also consolidating all it's design operations into CCG (client computing group, or lpatop/PC division) and DCAI (datacenters and AI, the server group). Networking group NEX will remain for now, but it is too small now and I'm guessing it will be sold off. Qualcomm is already looking to acquire some of Intel's operations as per The Wall Street Journal; Qualcomm is most likely interested in the CCG piece. (https://lnkd.in/gYkgwxGq) Intel will likely break itself into three major pieces - Manufacturing/Foundry, CCG and DCAI will become separate companies. But let's talk about Intel Foundry : Intel consolidated all it's manufacturing and research into Intel Foundry which started reporting it's own P&L from Q1 of this year. But it wasn't a different legal entity which now Pat Gelsinger has announced will happen. But even then, without a much more massive US govt assistance, Intel can't pay to become competitive again. US govt is the only entity that can pay for, the R&D needed to catch to TSMC and match Samsung Semiconductor, pay for new fabs etc. However Intel's failure to get a customer (Broadcom most recently, and Qualcomm and NVIDIA long ago, dropped the 18A offering) created multiple problems - the US Commerce Department's conditons for funding were not being met. And since AMD never showed any interest, - any help to Intel Foundry, which has been a part of Intel (unlike help to TSMC and even Samsung Semiconductors - both of which have well established fab business and serve multiple customers), would be construe a help to wider Intel business which then competes with AMD, Qualcomm, Nvidia. The latest move to separate out the Foundry as a legal entity opens to path to solve multiple problems at month. In my speculation the following will happen before Q3'2025: (1) Intel Foundry will not only become a separate company (legal entity) as Pat Gelsinger has already announced, it will in fact drop the name "Intel" altogether and be names something like "American National Fab corporation" - National Semiconductor would have been a good name but it is already taken. (2) Intel (or whatever survives without Foundry) will have less than majority stake in the new entity. A consortium made of AMD, Qualcomm, NVIDIA, Texas Instruments, Analog Devices ... all the marquee names of US Semiconductor industry will also be "encouraged" by US Commerce department to aquire, between themselves, a majority stake. Even the likes of TSMC may be persuaded. (3) US Govt will subsequently open the floodgates of money to the foundry to make it a success. (4) The new owners will bring about management and technical ladder changes acutely needed. I hope this happens for the sake of the talented employees, the US and the world public.

8

Dual-core ARM Cortex-A55 SoM starting from $34 https://lnkd.in/eC4M4PUa

1

Our game changing inference performance will enable new applications

31

As the growth chart for gallium nitride (GaN) wafer demand is looking more like a hockey stick, it’s apparent that innovators who have long believed in the merits of GaN are being proven right. A private company in El Segundo, CA that predates some of the other startups in the GaN-based power IC space by more than a decade has been hard at work developing solutions in this space for over seventeen years. That company is EPC - Efficient Power Conversion and I got a firsthand look at some of their solutions when I was at the 2024 APEC Conference. They were co-founded in 2007 by a luminary of the power IC industry and CEO Alex Lidow. Lidow was at International Rectifier for over thirty years and had leadership positions spanning R&D and manufacturing. He was even their CEO for 12 years. As co-inventor of the silicon power MOSFET in the 1970s, his seminal work has shaped development in this sector. Lidow was flanked by Joe Cao and Robert Beach, both of whom have also made their own impact. Cao had over 15 years of experience and more than 30 patents in developing GaN and silicon power device technologies before deciding to embark on the quest with Lidow. Beach had founded multiple companies aimed at bringing GaN ICs to the marketplace before partnering with Lidow and Cao, making him a perfect complement to their triumvirate. What put EPC on the map was the development of their proprietary enhancement-mode GaN (eGaN) transistors which they released in June 2009. Unlike traditional MOSFETs, which used built-in channels between the drain and source, eGaN MOSFETS had no such channel. eGaN MOSFETs’ have small geometry, low power dissipation, and are easy to manufacture, making them desirable for a wide array of applications. The founders also had the foresight to manufacture using a foundry meant to produce silicon devices and were able to create a mature, efficient, and low-cost supply chain to bring new products to market quickly. Nowadays, newer companies in the space are gaining similar traction using this approach, but it’s safe to say EPC was a pioneer and began to reap the benefits decades ahead of competitors. This is another key element that differentiates them. In my experience, “mature startups” tend to be modest organizations, often letting their technology speak for itself and not over-indexing on marketing their brands to potential talent. But the hiring landscape is far more competitive than it was when they started and the allure of high-growth startups can be attractive for talented engineers. EPC’s goal is to enable companies to develop their technologies by leveraging the power of GaN and avoid being left behind by developing on silicon. That, combined with their lengthy market presence and leadership team, are strong elements to highlight and EPC would be well-served to fold these details into their talent attraction messaging. #semiconductorindustry #galliumnitride #powermanagement

22

As modern chip designs heavily rely on SRAM, its density plays a critical role in the scaling of advanced nodes. According to the ISSCC 2025 Advance Program, the SRAM density of Intel Corporation's 18A (1.8nm-class) process aligns more closely with TSMC’s N3 than its N2 (2nm-class) process. ◾ Key Metrics: ▪️ Intel 18A: SRAM bit cell size of 0.021 µm², achieving ~31.8 Mb/mm². ▪️ TSMC N2: SRAM bit cell size of 0.0175 µm², achieving ~38 Mb/mm². ▪️ Comparison: Intel’s 18A SRAM density marks a notable improvement over Intel 4 (0.024 µm²) but doesn’t match TSMC’s aggressive N2 scaling. ◾ Notable Features: ▪️ Both nodes incorporate gate-all-around (GAA) transistors, yet TSMC has aggressively reduced SRAM sizes compared to its FinFET-based predecessors. ▪️ Intel 18A introduces a backside power delivery network (BSPDN), enhancing power efficiency and logic density potential. ◾ Bigger Picture: While SRAM density is vital, logic density often has a greater impact on overall scaling and performance. However, comparative logic density data for Intel 18A and TSMC N2 remain undisclosed. As SRAM scaling becomes increasingly challenging due to design complexities and variability at smaller nodes, the focus shifts toward balancing SRAM density, power consumption, and logic performance. What’s your take on the SRAM vs. logic density debate in advanced process technologies? Thanks again to Tom's Hardware for the full article with more background and insights click the source link in the comments below. #Semiconductors #Intel18A #TSMCN2 #SRAMDensity #ProcessTechnology #AdvancedNodes #ChipDesign #TechInnovation

181

3 Comments

2024 Wafer Fab Equipment revenue will remain stable before a 2025 surge Wafer Fab Equipment (WFE) revenue is estimated to increase in 2024 by +1.3% year-on-year (YoY) to $108.1B. This is due to a WFE expenditure increase across advanced logic as well as #NAND and #DRAM memory, while legacy node investment is robust. Stronger 2024 growth in service and support revenue is expected, at + 6%YoY to $23.5B, due to increasing fab utilization rates worldwide. In the long term, WFE revenue is expected to have a +4.3% CAGR24-29, reaching $133.7B in 2029, while service and support revenue will be $27.6B in 2029, at a +3.3%CAGR24-29. The moderate growth rate is due to the large WFE revenue increase in 2021-2022, which is related to the large demand for semiconductor devices. As device revenue declined in 2023, the ratio of WFE revenue as a % of semiconductor device revenue reached as high as 19%. In particular, investments in Greater #China (domestic and foreign) ensured the overall WFE did not decline in 2023, and similar investment levels should continue in 2024. However, such high #investments by #chipmakers will not be sustainable in the long term and are expected to decline slowly due to continuous WFE buyouts, with capacity exceeding production needs, governmental incentives, and general increases in chip production costs. Thanks again to Yole Group for the full article with more background and insights via the link below 💡🙏👇 https://lnkd.in/ejy-Ve2k #semiconductorindustry #semiconductors #semiconductor #technology #tech #it #innovation #chip #chips #foundry #geopolitics #ic #ai #iot #ml #informationtechnology #datacenter #computer #hpc #data #cloud

52

6 Comments

If you try to be objective. AMD started with Su and Papermaster. They promised ryzen to be on par with lower tier intel. They kept promising and delivering. It is not rocket science, ppl are accustomed to announcements then delays and then barely matching the amd cpu that is at least 9 months old. To the outside the problem is very clear.

6

2 Comments