2017 was a year to forget for chipmaker Ambarella (AMBA 3.63%), which produces image processing SoCs (system on chips) for action cameras, security cameras, dash cams, drones, and other devices.
The stock has stayed nearly flat for the year compared to a 45% gain for the Philadelphia Semiconductor Index. Let's look back at what went wrong for Ambarella this year, and whether or not it can bounce back next year.
Image source: Ambarella.
What the bulls believed...
Ambarella first captured the attention of investors as GoPro's (GPRO +4.09%) supplier of image processing SoCs. But after demand for GoPro's action cameras peaked in late 2014 and early 2015, interest in Ambarella faded.
Yet Ambarella bulls claimed that the chipmaker's fate wasn't completely tethered to GoPro's. For example, many GoPro clones (like Yi Technology's action cameras) also used Ambarella's SoCs, so GoPro's pain wasn't necessarily Ambarella's pain.
Moreover, Ambarella was still supplying SoCs to DJI Innovations and Hikvision, which were respectively the largest drone and security camera makers in the world. The bulls argued that both markets would grow and offset its softness in action cameras. They also claimed that Ambarella's "best in breed" reputation would prevent its competitors from gaining much market share.
What actually happened...
Unfortunately, Ambarella's top customers all started shopping around for cheaper chipsets over the past year. Even GoPro cut Ambarella out of the loop and hired Japanese chipmaker Socionext to create a custom GP1 SoC for its new Hero 6 cameras.
DJI's Spark drone. Image source: DJI.
Meanwhile, the Myriad 2 computer vision chip from Intel's (INTC 4.43%) subsidiary Movidius emerged as a major threat. DJI uses the Myriad 2 in its portable Spark drone, Hikvision uses the chipset in its AI-powered cameras, and Alphabet's Google uses it in its new Clips camera -- which automatically takes burst shots by learning and recognizing the faces in a user's life.
DJI previously used the Myriad 2 alongside an Ambarella SoC in its Phantom 4 drones. However, the Spark only uses the Myriad 2 for both computer vision and image processing tasks -- indicating that Ambarella could eventually be cut of the loop in its higher-end drones.
As a result, analysts expect Ambarella's revenue and earnings to respectively drop 5% and 35% this year. Those are dismal growth figures for a stock that still trades at 35 times earnings -- which is much higher than the industry average of 25 for semiconductor equipment makers.
But what about 2018?
2018 will likely be a make-or-break year for Ambarella. To make it, Ambarella must prove that its upcoming CV1 computer vision chip -- which has been in development for four years -- can counter Movidius' Myriad 2 and other similar chips.
Ambarella expects the first adopters of CV1 to be security camera and drone makers. Ambarella will demonstrate the CV1's capabilities at CES next January, but it doesn't expect the chipset to generate any revenue until later in the year. If Ambarella can leverage its "best in breed" reputation in image processing and convince OEMs to adopt the CV1 over Myriad 2, its growth might get back on track.
Another potential catalyst is a takeover. With an enterprise value of just $1.4 billion, Ambarella is a lucrative target for any company that wants to become the world's top maker of image processing SoCs. There's been buzz about Intel mulling a buyout, but investors should always take such rumors with a grain of salt.
Wall Street expects Ambarella's revenue to rebound 5% next year (fiscal 2019), but for its earnings to still slip 9%. This indicates that Ambarella's operating margins -- which have contracted over the past two years -- probably won't expand anytime soon as the chipmaker invests more heavily in CV1 and newer image processing SoCs.
Source: YCharts
The bottom line
I think the only real catalyst for Ambarella is a takeover, but I don't believe in buying unloved stocks simply based on buyout rumors. If Ambarella keeps going at it alone, it could struggle to match Intel's scale and retain its biggest customers. Moreover, the stock is still valued as a growth play, but the company simply isn't generating enough growth to support those valuations. Therefore, I think investors should avoid Ambarella and stick with better-diversified chipmakers instead.
