Here is an honest question to sit with before the bell rings tonight: how does a company with a 41% operating margin, 33% revenue growth last quarter, and 3.4 billion people using its apps every single day trade nearly 10% below where it started the year?

The short answer is $145 billion. That is the top end of Meta’s 2026 capital expenditure guidance, raised again in April from a prior range of $115 to $135 billion. The market has been wrestling with that number ever since.

Tonight, Meta reports Q2 2026 results after the close. Wall Street is looking for revenue of roughly $60.2 billion, up about 26.6% year over year, and EPS of approximately $7.14 to $7.23. The company guided Q2 revenue of $58 to $61 billion last quarter, so consensus is sitting near the upper end of that range. Meta has now topped revenue estimates for 15 consecutive quarters and beaten EPS expectations in 13 straight. That is a real track record, not a lucky streak.

But a beat alone may not be enough. It was not enough last quarter.

What Actually Spooked the Market

In Q1 2026, Meta delivered $56.3 billion in revenue, up 33% year over year. The fastest growth rate since 2021. Operating income came in at $22.9 billion with a 41% margin. Ad impressions grew 19%. Average price per ad climbed 12%. By almost every measure, it was an exceptional quarter.

The stock fell 6 to 7% after hours anyway.

The reason was not the revenue. It was the capex raise. Meta bumped its 2026 infrastructure spending guidance to $125 to $145 billion, up from the prior $115 to $135 billion range, citing higher memory component pricing and data center costs. In after-hours trading that night, the reaction was swift. As one analyst noted, the investment community is growing frustrated at the pace of cash being spent without a clear, near-term return mapped back to earnings.

Slight tangent, but it matters: this is not exclusively a Meta problem. Alphabet raised its own 2026 capex guidance to a range of $195 to $205 billion when it reported last week, and shares dropped roughly 6% despite strong revenue growth. The entire hyperscaler class has made a collective bet that the cost of falling behind on AI infrastructure is higher than the cost of overspending on it. Whether that logic holds is genuinely one of the most consequential investing questions of this decade.

What the Fundamentals Actually Say

The core advertising machine is not just holding up. It is accelerating. Meta’s Family of Apps generated $55 billion in ad revenue last quarter, growing 33% year over year. Ad impressions and pricing both rose simultaneously, a dual-metric outcome that analysts flagged as the strongest combination in several quarters. AI-driven ad ranking improvements have been a direct driver of that outperformance, and management’s tone going into this quarter was notably confident.

This is the part of the AI spending debate that gets glossed over. The infrastructure spend is not purely speculative. A meaningful portion of it is reinforcing a business that already works. Revenue is not waiting for the AI investment to prove out. It is already happening inside the ad platform right now.

There is also a capex math point worth understanding before tonight’s call. In Q1, Meta spent $19.84 billion on capex, well below the roughly $27 billion analysts had modeled. Yet full-year guidance stayed at $125 to $145 billion. That means spending needs to accelerate sharply in Q2 through Q4 to hit the midpoint. Investors will be watching tonight’s actual Q2 capex figure closely. If it comes in light again relative to the pace needed, that could be a meaningful positive signal for near-term free cash flow.

The Valuation Picture Going In

Meta’s forward P/E sits at approximately 18.3x as of this week. The trailing P/E is roughly 21.3x. Both figures sit below the stock’s own three-year average of around 25.4x and its ten-year average of approximately 26.3x. The PEG ratio is 0.83. EV/EBITDA is around 14.2x.

For a business growing revenue near 27%, with return on equity above 32% and return on invested capital approaching 30%, those are not rich multiples. GuruFocus pegs fair value at roughly $825 per share, implying the stock at current levels near $593 trades about 27% below its estimated intrinsic worth. That gap does not automatically close, but it tells you where sentiment is relative to the underlying business.

Worth noting: EPS growth this quarter is expected to come in nearly flat year over year, around 1%, even as revenue grows close to 27%. That gap between the top line and bottom line is what capex pressure looks like in the income statement. It is real, and it is why some investors feel stuck. The business is growing. The free cash flow story is complicated.

The Bear Case Deserves Honest Treatment

Reality Labs has now accumulated over $90 billion in cumulative operating losses. Last quarter it posted $402 million in revenue against a $4 billion operating loss. That drag has not resolved despite years of investment, and no clear profitability timeline has been disclosed.

Regulatory headwinds in both the U.S. and EU remain active. Management flagged ongoing trials related to youth safety issues as potentially resulting in a material loss. And with capex spending needing to nearly double in pace across the remaining three quarters of 2026, free cash flow compression is likely before it improves.

The options market is pricing an approximately 8% swing in either direction around tonight’s results. The last two earnings reports both saw the stock sell off after hours, despite beating on the headline numbers. Direction has been driven by forward guidance signals, not by the quarterly performance itself.

What to Watch Tonight

The revenue number will matter, but it is almost certainly not the deciding factor. Wall Street will be listening for three things: the actual Q2 capex spend and whether it tracks below the pace implied by full-year guidance; any commentary on AI commercialization beyond advertising, particularly the reported cloud infrastructure talks with Anthropic; and the Q3 revenue outlook and whether management sounds confident or cautious about advertising demand heading into the second half of the year.

Family daily active people is expected to reach 3.61 billion, up from 3.48 billion a year ago. Ad revenue is expected near $59 billion. Those are the numbers that tell you whether the core machine is still running well.

Here is where I land on this. Meta is not a broken business. A business with 41% operating margins, 33% recent revenue growth, and 3.4 billion daily users does not have a fundamental problem. What it has is a genuine tension between exceptional operating performance and an unusually uncertain, unusually large capital deployment cycle. Markets hate uncertainty. That is what created the discount heading into tonight.

The patient investor’s question is simple: does the advertising engine keep compounding while the AI infrastructure pays off over time? The evidence from the last several quarters says yes on the first half. The second half is still a bet.

Tonight we get another data point. Whether the market chooses to believe it is a separate question entirely.

The Cheap Investor