Today, in Facebook’s latest earnings statement, the social networking giant disclosed that it spent $1.06 billion on research and development in the first quarter of 2015. That’s almost 30 percent of all the revenue that it brought in during the quarter ($3.54 billion).
The 30 percent figure might be freaking out financial wonks — since it’s higher than usual for staid tech companies. (And it would be fair for them to cite this quarter’s revenue dip.) But really, they shouldn’t worry. Facebook is spending to hire more people in order to make itself — or its technology, really — super-smart.
The spending will continue, at least through the year.
“In 2015, we plan to continue hiring software engineers and other technical employees to support our research and development initiatives,” as Facebook stated plainly in its annual report in January.
It’s probably not worth it to throw a fit over the spending because over time, all the investment might very well pay off.
In December 2013, Facebook undertook a major commitment to research, specifically around artificial intelligence, signaled by the hiring of Yann LeCun. A pioneer of convolutional neural networks, a key architecture for a trendy type of AI known as deep learning, LeCun has been rapidly bringing on talent, partly through organic hires, and partly through acquisitions. For example, in January Facebook announced it had bought Wit.ai, a small speech-recognition startup.
And last year Facebook also brought on board Vladimir Vapnik, known for his work on a popular type of machine learning algorithm, the support vector machine (SVM).
The arrival of people like Vapnik and LeCun suggests that Facebook intends to be a haven for researchers, not just coders who can maintain or make slight tweaks to existing infrastructure.
Facebook has been making advancements when it comes to mining videos, pictures, and text for information. And in the domain of speech, Facebook picked up formidable technology through the Wit.ai deal. Facebook has been improving its AI systems and even sharing them with the world.
All this activity takes money. Deep learning talent is highly sought after. So Facebook has been digging into its deep pockets for this sort of expertise. Ads that are better targeted to consumers as a result of highly intelligent algorithms will inevitably pay off in the form of increased revenue to Facebook while other companies scramble to catch up and improve their tech stacks.
Eventually, the R&D activity — specifically hiring — should settle down. And eventually the most staggering numbers in Facebook’s earnings statements will end up being the top line — and the bottom.