By Bolaji Ojo

If a semiconductor market recovery is taking place right now, it must be the oddest upturn the industry has seen in its 60-plus years.

Veterans of the industry might, in fact, be forgiven for thinking the chips market has become unmoored from the fundamentals of its cycles. Unit shipments are down, inventories remain stubbornly high while average selling prices (ASP) are up. These facts do not align properly and are indicators of a market in data crisis.

By Bolaji Ojo

Memory semiconductor suppliers must be addicted to pricing turbulence. Or they are resigned to fate.

Having barely emerged from the last downcycle during which sales fell a stomach-churning 50 percent, memory IC makers are back at it again, adding new fabs and laying the groundwork for what could be the next punishing round of severe price swings.

They would say that is not the case, though.

By Junko Yoshida

What’s at stake:
At the height of autonomous vehicle hype, companies such as Nvidia, Qualcomm and Mobileye were front-runners dictating next-generation vehicle architecture. As more full-fledged software-defined vehicles (SDV) evolve, the competitive landscape is rapidly changing.

The path to software-defined vehicles (SDV) isn’t a straight line. It is branching into multiple forks. With transitions in E/E vehicle architecture happening at various speeds even within a single OEM’s fleets, the consolidation and integration of ECUs and software — from domain to zonal controllers and central compute — has been daunting.

Hence, for automotive chip suppliers seeking traction in SDV derby, solutions that simplify this transition are imperative.

Car OEMs are moving fast, not because SDV is fashionable but because they see SDV as the most logical path to control their future, secure the supply chain, and bring down vehicle costs over the long run.

By Ron Wilson

What’s at stake:
Intel has bet its chance to catch TSMC at the 1.4 nm node on being the first user of high-NA EUV lithography. It is a short road full of challenges, but early results look promising.

Lithography giant ASML announced this month that they have successfully printed a dense line pattern with 10 nm spacing, using the world’s first production high-NA EUV lithography system. In itself this is just another milestone in a long development schedule for ASML — similar patterns have already been printed using laboratory equipment. But in the global competition between Intel and TSMC for Angstrom-era semiconductor dominance, the announcement looms across the horizon like the first hints of dawn.

By Junko Yoshida

What’s at stake:
A lot rides on Japan’s first and only foundry company. Rapidus recently received an additional $3.9 billion in government aid. Included in the subsidies are a package that will help Rapidus grow into developing back-end packaging – in a country where no OSAT companies exist. Rapidus has yet to prove its 2nm process technology for front–end volume chip production. Is it trying to do too much?

Many factors contributed to the eventual downfall of the Japanese semiconductor industry. The Japanese industry languished in the1990s largely because of the prolonged US-Japan semiconductor trade war. A mixture of hubris and obsession drove Japan’s pursuit of higher quality DRAMs, while paying little attention to the demand by PC manufacturers looking for “good enough” memory. (Japan missed the tide of a growing PC market.) Finally, Japan, which dominated the global semiconductor market in the ’80s with DRAMs – Japan’s specialty then – never succeeded in converting its technology prowess to microprocessors and logic devices. 

Since then, the lessons Japanese government and semiconductor executives have learned could influence the outcome of Japan’s “renaissance” in chip manufacturing

By George Leopold

What’s at stake:
Fresh off receipt of U.S. government funding and tax breaks, Intel Corp. is spearheading the effort to revive American chip manufacturing. Will Intel Foundry’s internal sourcing strategy enable it to achieve a projected breakeven in 2027?

On the day it disclosed steep operating losses for its nascent foundry operations, Intel Corp. dropped the other shoe by unveiling a widely expected strategy that separates its fabless products unit from its fledgling foundry business.

The goal, Intel said, is establishing a “foundry-like relationship” among Intel Foundry, the chipmaker’s existing manufacturing operations, and its “product” business unit. The last is essentially the fabless portion of Intel Corp. that designs chips for data centers, networking and, the company hopes, future AI systems.

By Junko Yoshida

What’s at stake:
“Software-defined” is an inescapable trend even for automakers. Gone soon will be days when OEMs demanded tier ones to bring another (hardware) box with each new feature, so that they can pile up boxes in a vehicle that results in ever-growing complexity. While Software-Defined Vehicles (SDV) hold much promise, the trend demands that carmakers design vehicles using virtual ECUs in a virtual development environment.

For car buyers, software-defined vehicles (SDVs) could simply mean over-the-air updates or various apps running on in-vehicle infotainment systems that leave the underlying hardware unchanged.

For automotive design engineers, on the other hand, SDVs are both critical and problematic.