Why Nvidia Has Pivoted to Software
The Santa Clara, Calif.-based company Monday announced a Maxine, a cloud-based set of developer tools to help businesses build better videoconferencing software.
That’s right, a company best known cutting edge hardware, also makes innovative software tools.
Software is the reason investors should continue to buy Nvidia shares.
It’s hard to explain but in 2007, at the beginning of the cloud computing revolution, Jensen Huang, Nvidia’s chief executive decided to go all-in on artificial intelligence. The choice baffled many pundits because AI had been a lost cause for big tech so far.
Since the 1970s technologists have been drawn to the idea of computers learning to process data efficiently like the human brain. Huang made the case succinctly in 2016 when he addressed a Deep Learning conference about the work Nvidia was doing with self-driving cars.
He told an audience about how easily humans quickly distinguish between flying insects hurtling toward our car windshield and dangerous projectiles. We don’t swerve out of the way for bugs. On the other hand, computers have to process the image, identify the threat level, then react accordingly, all in real time. They need vast amounts of bandwidth to make this happen.
AI could absorb a big portion of the data processing.
Nvidia engineers are using the same big ideas they brought to self-driving to tackle other software problems, like telecommunication networks, connected medical devices, and now videoconferencing. And they are having a huge amount of success.
Maxine is a set of software tools that uses AI to fix the nagging problems inherent in our new work-from-home software experience like poor resolution, lousy face lighting, noisy backgrounds and even gaze correction – because working at the kitchen table can be distracting for a variety of reasons.
Videoconferencing has become a huge business. No company better illustrates the trajectory of that growth better than Zoom Video Communications ((ZM) -Get Report). Prior to the worst of the global pandemic in March, the small San Jose, Calif.-based firm mostly offered videoconference services to enterprises trying to hook up with farflung offices.
At the end of August, company managers reported 370,200 business customers with more than 10 employees, up 458% year-over-year. The number of clients with annual billing of $100,000 or more surged 112%. And the people using the service, corporate and otherwise, rose to 300 million every month.
The biggest problem all of these popular services face is data consumption. High resolution video images pack a lot of pixels and getting the information to reliably transfer across unstable home WiFi connection is a chore at best.
The software engineers at Nvidia used AI to reimagine the process. Instead of streaming an entire screen of pixels, Maxine analyzes the facial points of each person on the call, then uses algorithms to reanimate the image for other participants, according to a corporate statement.
These algorithms running on Nvidia hardware in the cloud can reduce the amount of data flowing back and forth by 90%. It means smoother videoconferencing experiences.
The connection between hard and software is important.
Investors should not think of Nvidia as a hardware company that also does software. Huang is building a powerful AI ecosystem with throngs of third party developers eager to make new applications for everything from self-driving cars, telecommunication networks, and medical devices, to videoconferencing.
In the same way that Apple ((AAPL) -Get Report) has an App Store for iPhones, and Salesforce ((CRM) -Get Report) has App Exchange, new apps developed for the Nvidia ecosystem enhance the stickiness of the platform. Ultimately it means more Nvidia hardware sales at the hyperscale data centers that will run those applications.
When the company reported second quarter financial results in August, Colette Kress, chief financial officer, noted data center revenues climbed to $1.75 billion, up 167% year-over-year and 54% sequentially. She also reported the Nvidia ecosystem of developers reached 2 million members, up 100% since 2018.
Nvidia shares trade at 49.3x forward earnings and 26.9x sales. While these metrics are rich, they do reflect the market’s confidence in heady growth ahead for the AI juggernaut.
Long-term growth investors should consider buying shares into all significant declines.