Skip to content


Two cars accident

PreAct Going Close-Up to Cut Lidar Cost

By Junko Yoshida

What’s at stake?
Elon Musk once famously called lidars a “fool’s errand,” a claim that has yet to be entirely disproven. Despite their high-precision performance advantages, lidars remain too costly to become an indispensable sensing modality in ADAS. Against the odds, a startup called PreAct sees a volume opportunity for “dirt-cheap” near-field lidars, not the long- or medium-range lidars most other players are pursuing.

Read More »PreAct Going Close-Up to Cut Lidar Cost
SiC laser cut boules

Laser-Cutting the Cost of Silicon Carbide Wafers

By Adele Hars

What’s at stake:
Investors are hot on the silicon carbide market, but the high cost of the wafers is an impediment to high volumes. Halo Industries, a startup just coming out of stealth mode, claims it has a solution that saves on the time, cost, and energy required to slice the SiC boules (ingots) into wafers. It’s got a big client and strong funding, but it has an unusual business strategy and is up against some much bigger players. Will Halo be able to scale to high volume fast enough to beat the competition?

Silicon carbide wafers are extremely expensive — anywhere from 20 to 50 times pricier than silicon. The drivers of the cost differential begin at the very beginning, with the weeks that the starting SiC boules spend in furnaces that are almost half as hot as the sun. The inefficiencies mount as the boules are cut with diamond wire saws, wherein up to 40% of the boule ends up as waste (or the more technical term, kerf), in the form of SiC dust lost in the sawing process. Yields are terrible, and the sawing process takes hours per wafer. Andrei Iancu, CEO and founder of Halo Industries, told the Ojo-Yoshida Report that his company’s laser systems are intended to replace saws, virtually eliminating waste and improving yield quality.

Read More »Laser-Cutting the Cost of Silicon Carbide Wafers
AlphaICs AI accelerator

AlphaICs: How to Stand Out in the AI Startup Crowd

By Junko Yoshida

What’s at stake?
The AI chip market is notoriously crowded, with many startups gunning for a share of the potentially large but highly fragmented edge AI segment. As the competitors court investors and some look down the road to an M&A exit, gaining market traction for an AI chip design will require more than a specsmanship game of “my TOPS are better than your TOPS.”

AlphaICs, an AI fabless chip startup based in Milpitas, California, is sampling an 8-TOPS edge AI inference co-processor that it says provides “the best frames-per-second (fps)/watt performance in the market for classification and detection neural networks.” The company will offer both the chip, called Gluon, and the software stack.

Read More »AlphaICs: How to Stand Out in the AI Startup Crowd