Marvell’s photonics buy; Onto’s process control stake; TSMC A14/13 and EDA flows; Apple shuffle; LPDDR6; SOCAMM2 chipset; MATCH Act progress; EUV resist crunch; India 3D packaging; digital test card for HPC; telemetry platform for power management; edge AI MCU for voice; new Google TPUs.
Onto Innovation is partnering with Rigaku, combining Onto’s analysis software with Rigaku’s CD-SAXS platform for advanced semiconductor process control. Onto also agreed to acquire a 27% stake in Rigaku for about $710M.
Tesla plans to use Intel’s 14A process for its Terafab AI chip manufacturing project in Austin, Texas, becoming Intel’s first customer for the new process node, reports Reuters.
Arteris is collaborating with GF’s MIPS to build physical AI computing platforms using smart NoC IP and SoC integration automation software.
TSMC North America Technology Symposium
Node number leapfrogging is in full swing, even though it’s not clear what those numbers actually mean anymore. Regardless, TSMC said this week that it will ship its A14 node in 2028, followed by an A13 optical shrink of that node in 2029. The foundry also said it will:
Allow customer access to COUPE co-packaged optics later this year;
Introduce backside power access at the A14 node;
Introduce a logic base die for HBM3E and HBM4;
Offer interposers up to 40X the size of today’s reticles, and
Synopsys announced successful silicon bring-up of low-power M-PHY v6.0 IP on TSMC N2P, tape-out of 64G UCIe IP, and 224G IP for next-gen AI systems, among ongoing collaboration on AI-powered digital, analog, and verification flows, agentic run assistance, and 3D-IC compiler improvements for TSMC’s CoWoS technology at 5.5x reticle.
Fig. 2: An advanced fab. Source: Siemens/TSMC
People
Apple announced that current CEO Tim Cook will become executive chairman and John Ternus, currently senior vice president of hardware engineering, will become the new CEO, effective September 1. Apple executive Johny Srouji will become chief hardware officer, effective immediately.
The European Semiconductor Industry Association elected Bosch executive Erik Rein as its new president, succeeding Michael Budde.
Memory
JEDEC previewed its LPDDR6 roadmap, targeting data centers and processing-in-memory. The LPDDR6 SOCAMM2 module standard is in development, as well.
Rambus announced a SOCAMM2 chipset, based on JEDEC’s forthcoming LPDDR6 SOCAMM2 standard. Aimed at low-power, high-performance LPDDR5X-based memory modules for AI server platforms, it will provide critical control, telemetry, and power delivery functions.
Kioxia released its BG8 Series SSDs for PC OEMs, built with BiCS FLASH generation 8 TLC-based 3D flash memory.
Fig. 3: A SOCAMM2 (small outline compression attached memory module) server memory module chipset. Source: Rambus
Geopolitics
The U.S. House Foreign Affairs Committeeadvanced a new bill titled “Multilateral Alignment of Technology Controls on Hardware (MATCH) Act,” aimed at limiting China’s access to advanced chipmaking equipment. The bill needs to pass a vote in Congress.
Asian chipmakers are facing photoresist supply shortages caused by the Iran war, per SCMP. The risk is acute for advanced EUV-based production.
3D Glass Solutions broke ground on India’s first advanced 3D semiconductor packaging unit, a greenfield ATMP facility in Odisha, India.
Samsung is facing labor pressure at its Pyeongtaek semiconductor complex, where thousands of unionized workers rallied for higher bonuses and threatened to strike as AI-driven memory demand lifts chip profits, reports AP.
Americas:
Advantest will joinApplied Materials’ EPIC platform in Silicon Valley, collaborating on links between front-end semiconductor manufacturing, advanced packaging, and back-end test, tying in with Advantest’s recently opened innovation center.
Resonac launched a new R&D center in Silicon Valley focused on next-gen semiconductor packaging.
EU/UK:
Austria’s Innio and the Net Zero Innovation Hub for Data Centers demonstrated 100% hydrogen-fueled backup power for data centers at the 3 MW scale.
Infineon is participating in three EU quantum pilot lines: SUPREME, CHAMP-ION, and SPINS.
In-Depth
Semiconductor Engineering published these articles this week:
Silvaco partnered with Taiwan’s ITRI to develop next-gen MCUs and support startup innovation.
Amazon to invest $5B in Anthropic, and up to $20B more in the future, with Anthropic securing up to 5 GW of Trainium chips for advanced AI models.
Fundings/listings
Australia-based Syenta raised US$26M in Series A for its advanced packaging technology for AI systems. Pat Gelsinger joined the board as the company expands into Arizona and pushes its localized electrochemical manufacturing technology.
Honeywell’s Quantinuum submitted a draft registration statement for its proposed IPO.
Forge Nano plans to list on NASDAQ through a merger with Archimedes Tech, helping scale its atomic layer deposition manufacturing technology for AI chips and defense battery applications.
U.S. Auto Decline Was Driven by Slow Adaptation, Not Just Cheap Labor (ITIF)
New Technologies
Advantest introduced a Pin Scale 5000B digital test card for its V93000 EXA Scale platform for advanced AI and HPC devices. The new card expands vector memory, supports concurrent testing across multiple IP cores with greater core-level visibility, and delivers data rates up to 5 Gbps.
Movellus launched a bench-to-in-field telemetry platform for data center power management with on-die Vmin sensing to capture voltage at nanosecond resolution and near-transistor granularity, enabling design teams to re-characterize silicon in the field against new workloads without going back to the bench.
Siemens expanded its portfolio of DC solutions, semiconductor circuit breakers, and solid-state switching devices for industry and infrastructure.
Synaptics launched its Astra SR80 series of edge AI MCUs delivering high-performance, voice-driven audio processing at ultra-low power for voice-first applications, including wearables, PCs, smart home, and office devices.
Expedera’s Origin Evolution NPU IP won the category of best edge AI processor IP at the Edge AI and Vision Product of the Year Awards.
Bolt Graphics taped out its Zeus GPU, claiming to reduce the total cost of compute by up to 17x across HPC, rendering, and emerging compute-intensive applications.
GigaDevice announced 32-bit MCUs with ultra-low power consumption for HMI and IoT edge solutions.
Robotics
Agibot unveiled five new robotic platforms and eight foundational AI models to support its “One robotic body, three intelligences” full-stack architecture.
DEEPX and Hyundai Motor Group Robotics LAB partnered on a physical AI computing platform for advanced robotics, co-developing AI architecture capable of real-time, large-scale generative AI models within robotic systems.
Lattice Semiconductor and Texas Instruments are collaborating on accelerating edge AI for robotics, combining TI’s sensing technologies with Lattice’s Holoscan Sensor Bridge solution.
Security
MIT researchers developed an ASIC for wireless biomedical devices that implements post-quantum cryptography and built-in defenses against physical attacks, claiming it is between 20x and 60x more energy-efficient than comparable PQC approaches.
Oversight
The UK National Cyber Security Centre announced a new device called SilentGlass that will block malicious or unwanted HDMI and Display Port connections.
The European Union Agency for Cybersecurity published an updated National Cybersecurity Capabilities Framework to assess the maturity of national cybersecurity strategies and strengthen national cybersecurity.
MITRE published a cybersecurity risk analysis for medical devices. The analysis focuses on cloud computing, AI/ML, and PQC.
Toyota announced an AI vision engine foundational model, bringing together visual, behavioral, and environmental data for its Toyota Woven City. The engine can be combined with the integrated ANZEN system to activate city behavior AI and drive sync assist, enabling people, mobility technologies, and infrastructure to operate as a single coordinated system.
Automotive industry news
Bosch released its third-gen SiC chips, delivering 20% higher performance and increasing overall drive electronics efficiency for electric vehicles.
China’s Nio established a new subsidiary dedicated to integrated circuit chip manufacturing, chip design, automotive R&D, and AI.
Humanetics launched a new ADAS testing platform, UFOnext, to meet the demands of automotive testing protocols. The new technology hopes to make ADAS testing faster and more reliable.
CounterPoint’s revised vehicle forecast showed a slowdown in EVs produced and purchased. This revision is the result of major automakers scaling back EV output due to a collective $70 billion loss from automakers like Ford, Stellantis, Honda, GM, and others.
Autonomous vehicles
Tesla and Pacific Gas and Electric Company announced that three Tesla products, the Cybertruck, Tesla Universal Wall Connector, and the Tesla Powershare Gateway, are approved for PG&E’s residential vehicle-to-everything program.
Waymo partnered with navigation app Waze to help cities find and repair potholes using user-reported information.
Lucid got new investments from the PIF and Uber, and expanded its robotaxi partnership with Uber.
Battery news
Long-awaited advances in battery chemistry and materials science are beginning to roll out, opening the door for higher capacity, faster charging, and a much lower likelihood of thermal runaway.
Nissan achieved real-world performance requirements for a solid-state battery, per Nikkei, for EV mass-production by the end of 2028.
CATL unveiled its Shenxing 3 battery, charging 10% to 98% in 6.5 minutes, per WSJ.
Magnachip announced two new Ultra Low-Ron 12V low-voltage MOSFETs designed for smartphone battery protection circuits.
Research
imec looked at how low-power wireless technologies, such as Bluetooth and ultra-wideband, are largely unexplored for enabling edge AI.
Recent paper
Strong ultrafast nonlinear optical response from megaelectronvolt electrons in semiconductors (Stanford)
Quantum
IonQ released a full-stack, buildable blueprint for scalable, fault-tolerant quantum computing.
Lawrence Livermore National Lab received $4.1M to lead a project as part of the Quantum Computing for Computational Chemistry (QC3) program, aimed at advancing commercial energy applications through quantum algorithms to accelerate chemistry and material science simulations.
Phoenix launched a Quantum Strategy aimed at becoming a national quantum hub, and partnered with Arizona State University on quantum workforce.
Trending Video
Why More CPUs Are Needed For Agentic AI: The shift from generative AI to agentic AI will significantly increase the amount of compute power needed in data centers. Queries to search for and analyze data from multiple sources will be performed simultaneously by agents and without human intervention, rather than a single request from a live person. Jeff Defilippi, senior director of product management at Arm, talks about the impact of running these queries 24 x 7.
InP and SiPho join CMOS as critical technologies. Lasers, CPO and OCS will be everywhere (indium phosphide, silicon photonics, co-packaged optics, optical circuit switch).
This website uses cookies to improve your experience while you navigate through the website. The cookies that are categorized as necessary are stored on your browser as they are essential for the working of basic functionalities of the website. We also use third-party cookies that help us analyze and understand how you use this website. We do not sell any personal information.
By continuing to use our website, you consent to our Privacy Policy. If you access other websites using the links provided, please be aware they may have their own privacy policies, and we do not accept any responsibility or liability for these policies or for any personal data which may be collected through these sites. Please check these policies before you submit any personal information to these sites.
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.
Leave a Reply