Meta 2Q26: Avoiding TikTok 2.0

Research evidence excerpt

Meta 2Q26: Avoiding TikTok 2.0

Mark Shmulik +1 917 344 8508 mark.shmulik@bernsteinsg.com 29 July 2026

DETAILS

Meta reported 2Q26 results after the market close.

PORTFOLIO MANAGER'S SUMMARY

There’s not much to report back on today. Revenues… in line, margins (ex one-time legal costs)… in line, 2026 CapEx saw

the bottom end of the range come up by $5B but top-end unchanged… about as expected. And the AI story? Still a work in

progress.

The earnings call felt a lot like a good old-fashioned brainstorming session. Model APIs for businesses, subscription and

outcome based pricing for AI products, AI agents for businesses for customer service and sales, consumer AI agents that

operate 24/7, temporarily selling excess compute, and so on. Whatever AI business models you can imagine coming out, that’s

the opportunity ahead of Meta. But just like those McKinsey problem-solving sessions, ideas shared on PowerPoint decks or

earnings calls aren’t real. That is, until it is real.

What Meta is building isn’t something optimized for earnings calls. Earnings calls are for estimate revisions and

quantitative evidence of new TAMs, new product engagement traction, and revenue run-rates. We’re just simply not there yet.

But if you want to better understand why then is Meta taking such an expensive and unpredictable ROIC endeavor while others

AI players at least grapple with demand and commoditization risk, the answer lies at the beating heart of Meta: consumers.

Meta gets to count 3.6 billion daily users across their family of apps, and incredulously they’re still growing with Instagram

now surpassing 2B+ daily users, WhatsApp is seeing 30 million messages per second, and even Threads now has 500M+

monthly users. Every once in a while, something new captures the consumer zeitgeist.