When Meta talks about AI, Wall Street seems to have a split personality. Investors cheer when the company credits AI for boosting revenue but get skittish over its astronomical spending on the same technology. Here’s why they love one and fear the other.
AI Revenue: Small Wins, Big Applause
When investors hear about Meta's 'AI revenue', they're mostly hearing about immediate, tangible improvements to its core advertising business. This isn't some far-off concept; it's AI working today to make the company's main cash cow even more profitable. AI models are enhancing ad targeting, optimizing campaign delivery, and increasing the price advertisers are willing to pay because the ads perform better. For example, new AI-driven ad models have improved conversions and click-through rates, leading to a double-digit rise in the average price per ad. This is the kind of AI that Wall Street understands and loves. It’s a low-risk, high-reward enhancement to an existing, wildly successful business that generates nearly all of the company's income. It’s a clear return on investment that shows up in the quarterly earnings report, and investors reward that predictability.
AI Spending: A Bet on a Different Universe
The spending side of the equation is a completely different story. Meta's capital expenditure plans are staggering, projected to be as high as $145 billion for 2026, nearly double the previous year's figure. This isn't just for improving ads. This is for building a future that doesn't exist yet. The spending goes toward massive data centers, custom chips, and the foundational infrastructure for both the metaverse (via Reality Labs) and new, ambitious products like agentic AI assistants. Unlike the ad improvements, the return on this investment is uncertain and years away. Investors see billions in spending that leads to negative free cash flow, as it did in the third quarter, and they get nervous. The Reality Labs division alone continues to post multi-billion dollar losses each quarter. This kind of spending is a long-term, speculative bet on creating entirely new computing platforms, and the market has a notoriously short attention span.
Not All AI Is Created Equal
The core of the issue is that there are two different 'AIs' at play. One is the practical, applied AI that makes Instagram Reels more engaging and Facebook ads more effective. This AI provides a direct, measurable lift to the current business model. The other is the speculative, foundational AI required to power the metaverse and create truly autonomous personal assistants like 'Muse'. While a product like Muse has already attracted millions of users, it is incredibly compute-intensive and its path to profitability is not yet clear. Investors are essentially evaluating two different companies: the hyperprofitable social media giant of today and the speculative R&D lab of tomorrow. The market cheers the former while anxiously questioning if the latter is a visionary leap or a money pit.
The Credibility Test
This tension isn't unique to Meta. Amazon's AWS was a massive cost center before it became the company's profit engine. Google's parent company, Alphabet, has its own 'Other Bets' division that consistently loses money on moonshot projects. The difference is that Wall Street is now demanding clearer proof that these massive AI investments will pay off. After seeing Meta pour billions into the metaverse with little to show for it so far, investors are more skeptical this time around. Unlike Microsoft or Amazon, Meta doesn't have a massive cloud computing business to help shoulder the infrastructure costs, meaning its ad business has to fund the entire expansion. Recent earnings reports from across Big Tech show that strong revenue growth is no longer enough; investors now want to see that the spending is translating into sustainable cash flow.
