Tata-ASML deal; what happened to H200 chips for China? $1.5T IC industry by 2030; new funding; AI memory tools; MRC spec; embedding solar chips in windows; Ford launches battery biz; Waymo does the backstroke; UMC’s high-voltage finFET.
The U.S. created a licensing path for NvidiaH200 shipments in January and has since approved sales to 10 Chinese companies, but so far no shipments have been confirmed, reports Reuters.
With a looming end-of-year expiration, SIA, SEMI, and other business groups are urging Congress to extend the US semiconductor tax credit and expand it to cover semiconductor design and other activities.
TSMC said IC revenues will hit $1.5 trillion in 2030, up from $1 trillion this year, fueled by AI consumption of token and compute power, Taipei Times reported.
Siemens inked an agreement with the EU’s Chips JU program to provide participating companies access to EDA software.
Acquisitions and deals
Tata Electronics is partnering with ASML to equip and scale its planned 300mm semiconductor fab in Dholera, Gujarat, including support for lithography systems, workforce development, supply-chain readiness, and R&D.
indie Semiconductor plans to acquire the fabless CMOS image sensor group from ams OSRAM for €40M (~$47.1M). The deal is expected to close in Q3.
TSMC plans to sell 8.1% of its stake in VIS, but will continue outsourcing its interposer production and licensing of GaN technology to VIS.
Applied Materials and TSMC will collaborate to advance materials engineering, equipment innovation, and process integration technologies at Applied’s new Silicon Valley R&D center. Arizona State, RPI, and Stanford are also joining as research partners.
Startups
Fractile raised $220M to deliver frontier model inference chips where memory and compute are physically interleaved.
Photonic Inc. added $70M in funding for its distributed quantum architecture, combining silicon-based qubits with native photonic connectivity.
Startup AMP recently raised $1.3B to build an independent AI grid, a shared pool of specialized AI chips and compute capacity to give smaller AI developers a way to compete for resources.
Infineon launched its 2026 Startup Challenge, focused this year on humanoid robots. The company is inviting startups to develop hardware and system solutions for areas such as artificial sensing, sensor fusion, virtual feedback, and motor control. Applications are open until May 27, 2026.
Reports
Global semiconductor materials revenue increased 6.8% year-over-year to $73.2B in 2025, according to SEMI, with gains in both the wafer fab materials and packaging materials segments.
MIT and ORNL researchers found a way to deterministically move atoms repeatedly within a material’s 3D atomic lattice. “We can reprogram materials to create defects at will, realizing entirely artificial states of matter not found in nature with a wide range of potential applications, including sensing, optical, and magnetic technologies,” said the lead scientist in a university news release.
Imec demonstrated the first 3D implementation of a charge-coupled device for AI memory applications.
New protocols and specifications
OpenAIreleased the Multipath Reliable Connection (MRC) specification through the Open Compute Project. The new protocol defines an Ethernet-based transport for AI/ML systems designed to improve “GPU networking performance and resilience in large training clusters.” AMD, Broadcom, Intel, Microsoft, and Nvidia contributed.
An NTU Singapore team created near-invisible, ultra-thin perovskite solar cells that could be built into windows and glass façades to generate electricity while still letting light through.
CEA-Leti highlights its top semiconductor advances in its 2026 Scientific Report, including FD-SOI, e-memory, 3D integration, photonics, quantum, and on-chip learning and energy management.
Clemson University and Czech Republic partners developed a new carbazole-based polymer and used it to build memristors.
Cisco issued a critical vulnerability for its SD-WAN controller, prompting a CISA warning of active exploitation with significant risk.
AMD and Intel released a number of security alerts, including a critical alert for AMD’s ROCm ecosystem and one for Intel’s data center graphics driver.
CISA announced numerous industrial control advisories this week and also announced a joint release of Software Bill of Materials for AI – Minimum Elements.
Two exploits, named YellowKey and GreenPlasma, were found in Windows BitLocker that can bypass vulnerabilities, found by Nightmare Eclipse and released on GitHub.
New defenses and trackers
Mobilicom launched a wearable cybersecure SW-defined radio built for drones and autonomous systems that supports encrypted point-to-point, mesh, relay, and swarm communications.
OpenAI announced Daybreak, a new cyber defense combining OpenAI models with Codex as an agentic harness to help find, validate, patch, and prevent software vulnerabilities earlier in the development cycle.
Rambus introduced new chipsets for notebooks, desktops, and workstations that is designed for AI PC memory modules.
Imec implemented a block-addressable 3D CCD memory device with an IGZO channel that shows potential as a CXL type-3 buffer memory for AI use cases.
Micron Technology started sampling 256GB DDR5 RDIMM server modules.
Kioxia released new high-performance PCIe 5.0 client SSDs.
AI inference using modern models requires billions of parameters, and moving them to where they can be consumed requires time and energy. A new effort to standardize a high-bandwidth version of flash memory proposes to keep those parameters much closer: inside the package with the GPU.
Manufacturing and test
UMC released a 14nm embedded high-voltage (eHV) finFET process for display driver ICs, which provides a 40% reduction in power consumption and 35% reduction in chip area compared to its current 22nm display driver process.
Keysight introduced PCIe 7.0 128 GT/s receiver validation with stress calibration, enabling end-to-end validation of transmitters and receivers.
Design
Siemens unveiled AI-powered library characterization software designed to accelerate and improve the creation of SPICE-based Liberty files, addressing the growing complexity of new processes, tighter margins, more corners, and emerging formats such as LVF.
L&T Semiconductor will adopt Synopsys’ simulation software to accelerate design and optimization of next-gen power modules and intelligent power modules.
Power
Infineon released a new high-power module that boosts efficiency and power density in high-voltage systems.
Vehicles, Batteries
With automotive LPDDR4 prices rising 70%, the DRAM shortage has become a structural supply-chain issue for automakers with no quick fixes, says S&P Global. AI data center demand is pulling memory capacity away from automotive, driving up prices for older automotive-grade DRAM and increasing risk for cockpit and ADAS systems.
Ensuring automotive chips are reliable, defect-free, and secure adds a whole new dimension to design for testability (DFT). Here’s why it’s about to get even tougher.
Waymorecalled 3,791 autonomous driving systems after a vehicle entered a flooded high-speed roadway.
Batteries and charging
CISA is warning that copper theft from EV charging stations is creating infrastructure problems and potential dangers due to exposed electrical components.
Ford launched Ford Energy to provide US-made battery energy storage systems, with deliveries starting in late 2027.
SLAC and Stanford researchers found that adjusting the heating rate during lithium-ion cathode manufacturing creates more uniform nanoscale structures, limits cracking, and helps batteries retain nearly 93% of their energy after 500 cycles without added chemicals or coatings.
New CPU Memory Module: Benefits and questions surrounding a next-gen low-power standard for high-performance compute. Frank Ferro, group director for product management at Cadence, talks about the benefits of the next-gen modular low-power memory standard, how it compares with other types of DDR, and how it can be used to reduce heat inside these devices and potentially avoid active cooling.
Workforce, Education
TSMC and Arizona State University have jointly founded the ASU Foundations for Equipment Technician Program to prepare students for semiconductor equipment roles.
The University of Idaho has partnered with Hiroshima University on the Microchip Engineering and Security Alliance to give students a pathway into the semiconductor and AI industries.
CSIS contends that the AI workforce problem is fundamentally a math pipeline problem. As AI further automates routine coding and SW tasks, competitive advantage will depend more on workers who understand the mathematical and physical principles behind AI, robotics, engineering systems, and real-world deployment.
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