Theft of Apple’s autonomous system IP; Meta’s new AI chips; AR laptop for travel w/100 inch screen; 5 minute EV battery-swap system; ML for edge acquisition; LPDDR flash in automotive; lithium-ion battery for sub-zero temps; hydrogen fuel cell advances; backdoor hacking; computer vision; 5G packaging.
Former Apple engineer Weibao Wang was indicted for stealing Apple’s autonomous vehicle hardware and software IP and giving the information to Chinese competitors. Among other items, authorities said they found source code for the project on the engineer’s personal laptop, which was seized at his home. Wang fled to China the same day a search warrant was executed. This is one of five cases brought by a recently created U.S. Disruptive Technology Strike Force.
Meta shared details about its first generation custom chip for running AI models, a new AI-optimized data center design with liquid cooling and the second phase of a 16,000 GPU supercomputer for AI research. The MTIA accelerator chip family specifically targets inferencing workloads. This MSVP ASIC solution is designed for processing video workloads.
San Francisco-based startup Ample showcased its five-minute drive-through battery-swap system, matching the time it takes to gas-up.
Miss your multiple monitors at home when you’re on the road? Startup Sightful demonstrated its augmented reality laptop called “Spacetop,” which it says adds the equivalent of a 100-inch screen with five-plus hours of battery life.
Infineon acquired Imagimob AB, a Swedish startup that provides a machine learning platform for edge devices. The TinyML and AutoML platform has a wide range of use cases, including audio event detection, voice control, predictive maintenance, gesture recognition, signal classification, as well as material detection, and will augment Infineon’s hardware/software platforms.
California-based Ethernovia raised $64 million to develop its automotive Ethernet technology. Investors include Porsche, Qualcomm Ventures, AMD Ventures and others.
Intel spinout Mobileye will supply Porsche with its SuperVision Level 2+ ADAS technology platform.
Big changes are ahead for EV manufacturers as they work out challenges such as hiring the right engineering talent, developing over-the-air software updates (OTA) capabilities, and getting factories to scale.
New automotive architectures are raising challenges for how to utilize memory effectively and efficiently. An LPDDR interface for flash allows different processors to utilize multiple banks of flash memory, which in turn reduces latency and interference. Here’s why memory architectures will need to shift to accommodate the new zonal controllers in automotive design.
Argonne and Lawrence Berkeley national labs developed an electrolyte for lithium-ion batteries, claiming it maintains high battery charging performance for future electric vehicles even at sub-zero temperatures. This is one of the key problems with BEVs today, because range is significantly reduced in cold weather even without running the heater.
Advent Technologies opened a new hydrogen fuel cell research, development, and manufacturing facility in Charlestown, Mass. The company is utilizing Los Alamos National Lab’s alternative material design, a composite polymer electrolyte with a power density of nearly 800 milliwatts/cm² at 160° C, a 60% improvement compared to phosphoric acid-based fuel cells.
Onsemi and Penn State announced an $8 million strategic collaboration that includes building a Silicon Carbide Crystal Center (SiC3) at Penn State’s Materials Research Institute.
Infineon rolled out a secure cold wallet application for private keys and cryptocurrencies, providing a security solution that simplifies payment transactions.
Check Point Research uncovered a malicious firmware implant for TP-Link and other routers/devices, allowing hackers to fully control the infected devices while evading detection. Dubbed “Camaro Dragon,” the hack included a customized backdoor to access compromised networks.
The Hardwear.io Security Trainings and Conference is happening in Santa Clara, CA on May 30-June 3 and includes in-person trainings, hardware security talks, workshops and a hardware hacking contest.
Traditional technology companies and startups are racing to combine machine vision with AI/ML, enabling it to “see” far more than just pixel data from sensors, and opening up new opportunities across a wide swath of applications.
The University of Southern California (USC) sees the broader computer vision market opportunity, as well, and proposed a “Processing-in-Pixel-in-Memory” (P2M). The goal is to reduce or eliminate the bottleneck in sending large amounts of amounts of data from the image sensor to the back-end processors.
Millimeter wave frequencies are essential for transferring more data more quickly, but they also require different packaging technology to minimize loss and drift. That opens up a number of tradeoffs around antenna in package, antenna on package, flexible circuits, and different substrates. There’s a host of new challenges ranging from over-the-air testing and cross-talk to impedance matching.
ETH Zurich students developed a lightweight carbon fiber smart cane for the visually impaired, featuring an integrated camera that provides haptic feedback, communication via vibrations, and real-time on-device machine learning.
The CHIPS Act, as well as the ongoing need for talent, are causing both industry and academia in America to rethink engineering education, resulting in new approaches and stronger partnerships.
Find upcoming chip industry events here, including:
Upcoming webinars are here.
 |
 |
 |
 |
 |
 |
Leave a Reply