During Nvidia’s (NASDAQ:NVDA | NVDA Price Prediction) earnings call a quarter ago, the management revealed something that even Meta (NASDAQ:META) did not. Nvidia’s CFO said, “For example, at Meta, advancements in their GEM model drove a 3.5x increase in ad clicks on Facebook and more than a 1% gain in conversions on Instagram…”
This explains a lot about what is happening in the background and what kind of impact AI is having on these businesses. The strength in Meta’s advertising business in the past few years isn’t just because advertisers are seeing booming demand.
During Meta’s earnings call, the company’s management talked about “AI-driven” improvements in ad performance. The “GEM model” is likely what they are talking about.
Is AI driving ROI after all?
Investors have been worried about AI’s return on investment. Will the hundreds of billions being spent on AI amount to significant cash flow figures for these hyperscalers once the buildout is done? No one knows just yet, but we can look into Meta’s advertising business to see if it has grown considerably.
We can look at the Family of Apps segment for that. If you strip away the metaverse story and the AI rhetoric, Meta is basically an ad machine.
Global advertising revenue was at $113.6 billion for all of 2022. It was already growing significantly, as 2019 ad revenue came in at $69.5 billion. In 2025, Meta reported $196.17 billion in advertising revenue. The forecast for 2026 is $243.46 billion in net worldwide ad revenues.
The global figures don’t show a tripling yet, but Nvidia is likely talking about U.S. ad revenue, or a narrower business segment. I wouldn’t attribute all this growth to AI. But either way, Meta is seeing an increase. Most of this ad money is going right back into AI through data center spending.
Is all this AI spending worth it for Meta?
Meta is catching a lot of flak for spending on its AI operations, but most of that spending isn’t really a waste when you consider the long-term returns.
These data centers are going to run the AI models that target users and drive up conversions and revenue. Over time, I would expect the increased advertising efficiency to pay off on this data center spending. I would still keep my expectations constrained, because Meta is spending hundreds of billions over many years. The payback period is going to be long.
Unlike Azure and AWS, Meta’s data center buildout is solely to support its own operations. Thus, if we ever see a time where compute demand declines for some reason, Meta is not going to collapse because it is not dependent on outside players.
Some of this advertising growth would’ve happened organically anyway. Plus, if there’s a downturn, fine-tuning ads for users will only go so far if they don’t have the money to spend in the first place.
All things considered, I do think the AI spending is worth it as long as it can keep advertising sales and profits terminally higher.
Should you buy META stock now?
No one knows what impact AI will have on ads during a downturn, and whether or not the market will focus more on the cash flow figures. META stock is trading at over 32 times free cash flow, which is fair value if you look at its historical valuations. The revenue growth isn’t translating into strong market performance anymore due to free cash flow staying low.
I don’t expect free cash flow to increase substantially unless Meta scales back on AI spending.
I’d buy the stock, but your main focus should be on the hardware side. Most of Meta’s free cash flow and advertising gains are flowing directly into AI hardware due to its data center buildout. You’re better off holding the companies making the GPUs and CPUs for Meta’s data centers right now.
You can then slowly rotate those gains back into Meta and other hyperscalers once the buildout starts to mature.