TSMC mature node capacity reduction; Photonics West highlights; Intel earnings; tariffs on memory makers; big fundings; Davos takeaways; ultrafast data transmission; DARPA’s IC heat program; new wafer smoothing tech; EV rankings; automotive hacking competition; side channel attacks utilizing memory reorderings; university blitz of hands-on IC training and labs.
TSMC is expected to reduce its Fab 14 mature-node capacity by 15% to 20% to free up resources for its advanced packaging technologies, reports Counterpoint. The foundry will likely rely on its VIS affiliate site in Singapore (operational in late 2026) and other overseas fabs to ensure continued supply for older nodes.
Memory
The U.S. threatened 100% tariffs on South Korean memory companies that don’t expand their U.S. production capabilities, reports Bloomberg.
The World Economic Forum was held in Davos, Switzerland this week, with some notable chip industry sound bites:
Anthropic CEO Dario Amodei told Bloomberg that selling advanced AI chips to China would be a big mistake. “It’s a bit like selling nuclear weapons to North Korea.”
Nvidia‘s Jensen Huang called AI the foundation of the ‘largest infrastructure buildout in human history,’ comparing it to a five-layer cake — energy, chips and computing infrastructure, cloud data centers, AI models, and ultimately, the application layer. Huang also said that “robotics is a once-in-a-generation opportunity” for Europe, given its strong industrial bases.
Big Fundings
Upscale AI raised $200M for commercial deployment of its synchronized AI engine, which unifies GPUs, AI accelerators, memory, storage, and networking.
Neurophos raised $110M to deliver its optical processing unit that integrates more than 1 million micron-scale optical processing elements on a single chip.
Ethernovia raised more than $90M to build Ethernet-based, packet processor-centric networking chips for SW-defined autonomy and physical AI.
SPIE Photonics West conference and related highlights this week included:
CEA-Leti demonstrated a co-packaged microLED and organic photodetector architecture that enables optical sensing functions, along with a hybrid architecture that integrates quantum cascade lasers with silicon photonic platforms for mid-infrared applications.
MIT demonstrated its new photonic chip with integrated antennas that cool trapped ions 10 times more efficiently than standard laser cooling.
Lumentum unveiled high-power laser platforms that support advanced manufacturing of electronics and semiconductors, as well as 3D sensing VCSELs.
Financial releases this week:
Intel beat Wall Street’s Q4 2025 expectations, despite reporting a quarterly revenue decline of 4% year-over-year and flat annual revenue compared to 2024. Guidance for Q1 was softer than expected due to low supply, but Intel expects better numbers in Q2.
JCET opened its Shanghai automotive grade IC packaging/test facility and announced a key milestone in CPO, delivering samples of its silicon photonics engine.
India’s Tata plans to invest $11B to build an “innovation city” in Mumbai.
In Saudi Arabia, KAUST established a quantum foundry for device prototyping and process development.
Europe
A €50M European test and production pilot line to make scalable and reliable photonic chips for quantum applications will launch this year.
The EU proposed a revised cybersecurity act, aimed at derisking the mobile telecom networks from high-risk third-country suppliers, including Huawei and ZTE. CEPA expresses skepticism about the new rules.
Americas
NETL has launched ClaiMM, a new digital platform dedicated to supporting U.S. research across the full critical minerals and materials (CMM) value chain.
CSIS‘ report, A Living Playbook for America’s Technology Long Game, asserts the U.S. “must (1) “play all the keys” by developing dexterity across technology types, (2) make speed-to-scale the organizing principle for enabling infrastructure and technology diffusion, and (3) defend its networks of innovators at home and abroad against mercantile and malign threats.”
Optalysys raised ~$31M to commercialize its silicon photonics technology.
AheadComputing raised $30M for its high-performance CPU microarchitecture for AI workloads.
Sensesemi Technologies raised INR 250M (~$2.7M) to develop edge AI chips for markets such as industrial automation and medical devices.
Noveon Magnetics raised $215M to expand its rare earth magnet capacity.
Quantum chip company SEEQC plans to go public via a SPAC merger with Allegro Merger.
Deals
France’s CEA-Leti and Canada’s C2MIinked an MoU to develop joint projects in areas such as microsystems, advanced packaging, and photonics.
Scintil Photonics, Presto Engineering, and CEA-Leti will work together on a roadmap for optical interconnects.
Keysight is providing satellite channel and user segment emulation solutions to support ground testing of 5G non-terrestrial network capabilities as part of the Airbus UpNext SpaceRAN demonstrator.
SEALSQ and Kaynes are planning a joint venture in Gujarat, India, that will produce up to 300M PQC chips per year.
D-Wave Quantum completed its $550M acquisition of Quantum Circuits.
Reports/Opinions
A McKinsey report argues that traditional estimates of semiconductor industry growth fall short by underestimating the impact of Chinese companies and OEMs with in-house design capabilities.
Total global server shipments are expected to grow ~13% YoY in 2026, with AI server shipments increasing ~28% YoY, per TrendForce.
Selecting the Right Low-Loss Materials for High-Frequency Circuits (IDTechEx)
Chips in Japan: Industrial policy, decline and renewal (RIETI)
Research
UC Irvine developed a new wireless silicon chip transceiver that achieves data transmission speeds of 120-Gbps, previously only possible through fiber-optic cables. “We call this technology a ‘wireless fiber patch cord’ because it offers the blistering speed of fiber optics without the physical cables,” said Payam Heydari, UC Irvine Chancellor’s Professor of EECS, in a news release.
DARPA provided an update on its Thermonat program, which drives advancements in modeling, predicting and verifying heat flow in advanced semiconductor devices. Highlights include UC Boulder-led efforts to bring high-accuracy thermal modeling tools for IC designers, as well as IBM’s and Ansys’ collaboration to model thermal behaviors of ICs at the atomic level.
Fraunhofer IIS highlighted advancements in hardware and software for spiking neural networks.
Cadence unveiled its sixth-generation HiFi iQ DSP, which uses a new architecture designed for next-generation voice AI and immersive audio applications to deliver over 40% performance uplift on numerous audio codecs compared to the previous generation.
Canon developed a wafer smoothing technology, called inkjet-based adaptive planarization (IAP). “Topographical irregularity is thereby reduced to 5 nm or less, facilitating the uniform layering of structures—essential for subsequent downstream processing,” according to the news release.
MediaTek debuted new chipsets based on 3nm and 4nm processes for flagship and premium smartphones.
Blue Origin introduced TeraWave, a satellite communications network designed to deliver data speeds of up to 6 Tbps “anywhere on Earth.”
SK hynix achieved ASIL-D certification on its latest LPDDR5X automotive DRAM.
Volvo released its new all-electric midsize SUV built on the “SPA3” electric architecture, with a range up to 400 miles, plus an ability to add up to 211 miles in 10 minutes using a 400kW fast charger.
Industry:
Germany unveiled a new €3 billion subsidy program for electric vehicles, including Chinese brands. Subsidies range from €1,500 to €6,000.
Canada is reducing its 100% tariff on Chinese EVs to 6.1% for up to 49,000 units annually. Tesla is expected to benefit, as it can now more easily export Shanghai-built vehicles to the Canadian market.
A Tutorial on 5G NR-V2X: Enhancements, Real-World Applications, and Performance Evaluation (Qualcomm)
Hydrogen storage technology that stores and uses negatively charged hydrogen inside a solid (Institute of Science Tokyo)
Security
At Tokyo’s Pwn2Own Automotive 2026 competition, Synacktiv researchers used a USB-based attack to successfully hack Tesla‘s infotainment system, and Fuzzware.io exploited a number of bugs on charging controllers and connectors. Find more results of the competition here.
Mitre rolled out an Embedded Systems Threat Matrix, a framework to safeguard embedded systems that power the U.S.’s critical infrastructure and defense technologies.
Morrison Foerster provided an alert on cybersecurity rules for the semiconductor sector, highlighting the laws that can apply directly and indirectly throughout the supply chain.
The SEMI Foundation has been approved by the U.S. Department of Labor as a National Apprenticeship Sponsor and will work directly with employers to design, register, and scale apprenticeship programs.
Universities are investing in hands-on semiconductor programs:
West Valley College in Saratoga, California, was awarded a $3.9M U.S. federal grant for a new semiconductor certificate program that includes a semiconductor fabrication lab for hands-on experience.
Texas Tech University was awarded a $12M grant that will go towards a $24M nanotechnology laboratory cleanroom at the Semiconductor Nanofabrication Center in Lubbock.
Purdue and Dassault Systèmes will build a new lab for virtual twin technologies for semiconductor processing and testing.
Trending Video
Workload-Specific Hardware Accelerators: What differentiates accelerators from other processing elements. Sharad Chole, chief scientist and co-founder of Expedera, talks about the role of neural processing units inside AI data centers, tradeoffs between performance and accuracy, and new challenges with chiplet-based multi-die assemblies.
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).
Reducing variation in manufacturing, monitoring behavior over time, and targeting specific workloads can have a big impact on power, performance, and area/cost.
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