MARK ZUCKERBERG may have a history of copying others’ ideas, but when it comes to navigating the generative artificial intelligence (AI) hype cycle, he’s the one forging a smarter path.
In the year to date, Meta Platforms, Inc.’s shares have significantly outperformed its tech giant peers, rising 54% while the likes of Alphabet, Inc.’s Google and Microsoft Corp. have gone up 18.5% and 12% respectively. Why? Investors have become more skeptical in recent months about how quickly new AI models will economically benefit the customers of cloud giants, and their enthusiasm for Microsoft, Google and Amazon.com, Inc. has waned.
Zuckerberg’s company is different. Instead of merely pitching artificial intelligence as an economic boon, it has actually demonstrated how AI has made Meta’s own ad systems more effective and ultimately profitable. Zuckerberg’s clarity on those details in earnings calls has won him praise from analysts, and likely contributed to the stock bump. The lesson he’s offering his peers on AI is an oldie but a goodie: Show, don’t tell.
Zuckerberg, of course, is in a different position to Amazon, Microsoft and Google, in that he doesn’t sell AI tools as part of a market-dominating cloud service like they do. But that doesn’t mean the so-called hyperscalers can’t learn from his approach. One of the biggest factors currently weighing on sentiment about artificial intelligence is around utility. Cutting-edge AI models from OpenAI, Anthropic and Google can conjure humanlike prose, computer code and photorealistic images — but figuring out how to plug that into business processes is taking longer than many in the market would like.
Anthropic, a leading maker of generative AI models, told me that the three most popular use cases for its technology were chatbots, text summarization and coding. But it and other companies selling AI services haven’t been able to articulate more concrete examples of their technology’s return on investment.
Zuckerberg has, particularly in Meta’s last quarterly earnings call. “It was his best earnings call as a public CEO,” according to Gene Munster, managing partner of Deepwater Asset Management. “He explained the near-term benefits of AI, the long-term benefits and the timing of how all this is going to play out. And he did it in a compelling way.”
According to Meta’s latest earnings report, the company’s artificial intelligence tools and large-language-model technology boosted content personalization on Instagram and Reels, driving daily active users up by 6%, the most in the past eight quarters. Daily watch time across all videos including Reels went up 25%, and the short-videos feature is now expected to bring in $10 billion in revenue for 2024, which would be more than three times what it was three quarters ago, something analysts are attributing to the AI systems powering its recommendation feeds. Zuckerberg also said that Meta’s AI technology helped increase time spent on Instagram by 24% and led to more effective ad targeting.
All of this gives Zuckerberg the permission he needs to build out highly expensive AI technology, according to JPMorgan Chase & Co. equity analyst Doug Anmuth. Meta “continues to earn the right to spend big on genAI,” he said.
Spending on AI is a touchy subject for Big Tech at the moment. In the quarter ending in June 2024, Apple, Inc., Microsoft, Alphabet, Amazon and Meta spent a combined $55 billion, a capex figure that’s 55% than the previous year, according to data compiled by Bloomberg, numbers that have rattled investors.
Google’s explanation for why it should keep spending on AI hasn’t helped by being vague: “The risk of under-investing is dramatically greater than the risk of over-investing,” Alphabet CEO Sundar Pichai has said.
In his latest earnings call for the second quarter, Pichai highlighted how Google’s revenue for search and YouTube had grown by 14% and 13% year over year, respectively, but he didn’t explicitly attribute that growth to the use of AI to improve their financial performance, for instance, by enhancing recommendation feeds or making ad targeting more effective. Pichai should consider offering more concrete examples of how Google is eating its own dog food, and demonstrate with specifics from his own business about how AI can improve a company’s financial performance.
Zuckerberg learned all this the hard way, thanks to his disastrously expensive bet on the metaverse — Meta’s Reality Labs division lost $16 billion in 2023 — a business that shows little sign of paying off any time soon. He knows that to get backing for his latest gamble on generative AI, he needs to articulate how it is benefiting his own business. So far, he seems to be doing a better job at that than his peers.
BLOOMBERG OPINION
