IBM Takes A Surprising Lead In Self-Driving Cars

| About: International Business (IBM)


IBM's Watson AI will control the first autonomous vehicle to begin regular passenger service in the US.

The ride-sharing service uses a vehicle built by Local Motors, which features extensive use of 3D printing.

IBM has shown that it can “skate to where the puck will be."

The first fully autonomous vehicle to take to public roads and provide regular passenger service will not have been built by Tesla (NASDAQ:TSLA), Alphabet (NASDAQ:GOOG) (NASDAQ:GOOGL), or any of the big three automakers. It will have been built by a heretofore obscure startup, Local Motors, with artificial intelligence by IBM's (NYSE:IBM) Watson.

Source: IBM


IBM issued the press release about the vehicle, called "Olli", yesterday. According to the release, Olli is already being used on public roads in the District of Columbia, and will be used in the Miami-Dade area and Las Vegas later this year.

Olli is a showpiece for IBM's Watson IoT cloud-based cognitive services. Olli carries more than 30 sensors for vehicle control and situational awareness, with Watson continually analyzing the data. Not only that, but also passengers can talk to Olli's Watson brain to give destinations and ask questions. IBM claims that passengers will be able to interact conversationally with Olli, asking questions about the vehicle, destination and routing decisions made by Watson.

This appears to be the right way to start an autonomous vehicle ride-sharing service. The geographic region for the service is limited and probably has been very carefully mapped prior to starting the service. Olli will always be on familiar ground.

Given the number of sensors and volume of data, it appears that a substantial amount of Olli's smarts are hosted locally in a computing device on the vehicle. It may be that Watson's cognitive services are strictly cloud based, with the vehicle system performing bulk sensor data processing and some other low-level functions such as position determination.

But this doesn't seem like quite the right answer. Many situations, such as pedestrian and obstacle avoidance, require split-second decision making, so a significant portion of Watson may be riding along with the vehicle. What the local computing device is, and what the exact division of labor is between it and any IBM cloud services is something that I'll try to investigate further.

Local Motors has opened a facility in Maryland to support vehicle operations, which features an open collaborative business model. Co-creators are welcome to experiment with new vehicle technologies and implement their designs on the company's large 3D printing system. Much of Olli itself is 3D printed.

First Among Equals, but a Mobile Outsider

With the appearance of Olli, IBM has suddenly been thrust into the position of first among equals in cloud-based AI services. Over the past few months, as I've analyzed the developer conferences of Microsoft (NASDAQ:MSFT), Google, and Apple (NASDAQ:AAPL), one of the very important themes has been the emergence of AI services as an augmentation of mobile devices. Now, IBM has emerged as a clear leader in AI, but not necessarily for mobile devices.

IBM Watson is not embedded in a mobile operating system the way that the AI services of Microsoft, Google and Apple are. Its services are available to any developer that wants to make use of them, so Watson could serve as the AI back end for any mobile app for any of the mobile platforms.

IBM has set itself up as a merchant vendor of Watson's cognitive services. This business model may have advantages, but it puts IBM into competition with the services integrated into the mobile operating systems of the big three. More importantly, the mobile platforms have these services integrated into app development tools provided to developers. Using the platform AI tools will necessarily be relatively frictionless compared to using IBM's Watson.

This is probably a significant barrier to widespread adoption of Watson, but there is an opportunity here as well. It's been pointed out that Apple's AI efforts are relatively weak compared to Microsoft and especially Google. It's a legitimate concern that Apple might lose the AI race.

Apple has released a set of APIs that will allow iOS and macOS developers to make use of Siri in their apps, something that both Google and Microsoft already provide. But Apple's Siri APIs are restricted to a very limited set of functions. Compared to the AIs of Google and Microsoft, Siri just isn't that smart.

Apple seems to want to pursue its homegrown AI, but this reminds me of AAPL's attempt to develop its own, homegrown multitasking operating system as a successor to the classic macOS. After Steve Jobs returned to Apple, he pulled the plug on these efforts in 2001. He brought with him the NeXTSTEP operating system, and NeXTSTEP became the basis of OS X.

Siri may be another example where the internally developed tech just doesn't cut it. Indeed, this seems to be the growing sentiment among developers as expressed by Marco Arment. IBM's Watson could fill this void. Apple and IBM are already partners, and it appears that AAPL has been willing to forego offering server-side programming tools (enterprise cloud services) to its developers in deference to IBM. Apple turning to IBM for Watson AI services would be a win-win for both companies.

Exactly how this would work remains to be seen, but developers would probably want the AI services to be integrated into the Xcode development environment. Apple might offer the Watson AI tools as part of its cloud services in a "powered by IBM's Watson" branding.

Investor Takeaway

I think the Olli project is going to be a tremendous boost for IBM, assuming that everything works, which I think is a reasonable assumption. I doubt IBM would be willing to have its name put on something like this that hadn't been thoroughly rung out.

Even though IBM has a reputation for being stodgy and slow, it appears that things are starting to turn around under CEO Ginni Rometty. IBM has definitely skated to where the puck will be on self-driving vehicles. Other initiatives such as its RISC-based Power9 server, and the incorporation of Apple's Swift as a server-side programming language indicate considerable forward thinking. Apple and IBM seem very complementary.

There are still headwinds. IBM's revenue fell 5% y/y to $18.7 billion in 2016 Q1. GAAP net income fell as well by 17% to $2.0 billion. IBM is still in transition, but the future looks very bright for its cloud services, including Watson. I rate IBM a hold.

Disclosure: I am/we are long AAPL.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.