NXP to Nab Ambarella
The Financial Times is reporting that NXP is evaluating Ambarella as an acquisition. Although the FT describes $AMBA as a designer of camera chips for self-driving cars, the automotive business accounts for only 22% of the company’s sales. The rest is IoT, meaning a mix of industrial and consumer video applications. Best known among these is Insta 360, the company that ate GoPro’s lunch. As for NXP, 58% of its business is automotive and 19% industrial/IoT. The remainder is mobile (e.g., NFC and various small analog chips for cell phones) and comm infrastructure (e.g., power transistors for cellular base stations), two areas the company is deemphasizing.
NXP acquired Freescale in 2015. Before that deal closed, Freescale acquired Canadian company Cognivue, which had licensed video-processing technology that Freescale had integrated into an auto-focused S32V chip and had an exclusive license for using that tech in auto applications. I’m not aware of NXP having developed an S32V chip in the past four years, although the ones from then are still available for new designs.
Since then, NXP has acquired, developed, and licensed a variety of AI engines (NPUs), making it one of the first and most experienced companies in edge AI, although how lucrative AI-enabled chips have proven for the company isn’t obvious. Meanwhile, Ambarella was good at video processing, and it has diversified from classical computer vision approaches to AI-based ones. It’s also been becoming more of a developer-friendly processor company, with all of the enablement that entails, and less of an ASSP supplier. As a long-time microprocessor/microcontroller company, NXP excels at enabling customers.
NXP and Ambarella are a good match, serving similar markets with different products, now that the S32V has gotten stale. Any overlap in AI technology should be a nonissue given the proliferation of NPU designs within NXP—i.e., one more approach to AI within NXP won’t shake things up much. NXP’s investment in AI-related software should paper over differences for customers.
