Meta: An Institutional Trim Doesn't Make the Capex Selloff a Sell Signal

A fund trimmed Meta. Another fund did the same. A third fund is trimming Meta, and if you follow the weekly parade of 13F-driven headlines, it starts to feel like the smart money is exiting the building.

I've been puzzled by how much weight investors give to these position trims. Suncoast Equity Management, a boutique firm managing over $6 billion, reduced its Meta stake by 2,169 shares in a recent filing. That's the kind of rebalancing move that happens every quarter across the industry. Other large funds have also trimmed positions. These are large institutions adjusting weightings in portfolios measured in hundreds of billions of dollars - not thesis-based exits.

The real story isn't who's trimming what. The real story is whether the selloff behind these trims reflects a cracked business or a market that's confused a financing decision for a demand collapse. The numbers suggest the latter.

The Two Stories the Market Merged Into One

Meta's Q2 2026 earnings report on July 29 landed with a revenue beat and an EPS miss, and the stock sold off roughly 8-10% in after-hours trading. Almost every headline collapsed two unrelated stories into one. Let me separate them.

Story one: the ad engine accelerated. Revenue hit $60.8 billion, up 28% year over year, beating the roughly $60.2-60.3 billion consensus range. Advertising revenue specifically reached $59.4 billion, up 27%, with ad impressions up 14% and average price per ad up 12%. Family daily active people grew 3% to 3.60 billion. Reality Labs - the division that houses AI glasses and metaverse ventures - came in better than feared, with a $4.62 billion operating loss versus the $5.07 billion the Street modeled.

Fundamentals: Revenue growth of 28% YoY, EPS miss due to $3.58B in one-time charges (legal + severance). Ad revenue +27%, Family DAP +3% to 3.6B. Reality Labs loss $4.62B vs $5.07B expected. Free cash flow $784M, down 91% YoY, driven by $31.1B capex (98% of operating cash flow). No share repurchases; dividend exceeded FCF.

Story two: free cash flow cratered. Quarterly capital expenditure hit $31.1 billion, consuming roughly 98% of the $31.86 billion in operating cash flow. Free cash flow for the quarter was $784 million, down about 91% year over year from $8.55 billion. Meta repurchased no shares. The dividend alone exceeded the quarter's entire free cash flow.

One is a demand story. The other is a financing story. The market treated them as the same, and that's where the valuation gap opens.

The Valuation Disconnect

Meta's forward P/E sits at 21.4 times, nearly in line with the S&P 500. For a company growing revenue at 28% with an ROIC (return on invested capital - the efficiency with which a business generates returns from the capital it deploys) of 23.7% and a gross margin of 81.8%, this makes little sense. The stock trades at a lower trailing P/E than Microsoft at 27.4x, even though Meta's growth rate is more than double the software giant's. Even against Alphabet at 18.9x, Meta's multiple is only slightly richer while growing revenue faster.

The EPS miss itself deserves scrutiny. Diluted EPS came in at $6.18, versus roughly $7.22 in consensus - but the quarter absorbed a $2.4 billion legal-proceedings charge and $1.18 billion in severance from a May 2026 headcount reduction. By Meta's own disclosed math, operating income excluding those charges would have grown 9% year over year. The miss was accounting, not economics.

Full-year 2026 capex guidance has been narrowed to $130-145 billion from a prior range of $125-145 billion, raising the floor by $5 billion. That's a significant number, but TTM (trailing twelve months) capex already sits at roughly $91.8 billion, so the acceleration is front-loaded in the second half. The question investors should be asking isn't whether the spending is large - it is - but whether it's productive. Zuckerberg's clearest framing yet came on the Q2 call: selling intelligence, not just raw compute, will yield "significantly higher" margins than the cloud-rental model the hyperscalers are racing toward. That's the ROI argument, and it won't be proven in one quarter.

The Moat Holds Under Stress

Before calling a contrarian setup, I need to verify the competitive moat survives the pressure that crushed the price. Meta's moat isn't just scale - it's the combination of 3.60 billion daily active people, an ad-delivery machine running on both volume (14% impression growth) and pricing power (12% price-per-ad growth), and AI models that are accelerating targeting, creatives, and measurement.

The capex concerns echo the 2021-2022 metaverse spending cycle that wiped $307 billion from Meta's market value. But this is different. The 2026 AI infrastructure buildout is not a bet on a product that doesn't exist yet - it's an investment in compute and models that are already improving ad performance for the core business, which remains Meta's revenue engine. The ad engine grew 27% this quarter. Disruption of the core? I haven't seen the data yet.

AInvest's aggregate signal labels the stock Buy, and its fundamental rating of 7.01 out of 10 reflects the underlying strength. That doesn't override the need for an independent check, but it does confirm the institutional research consensus hasn't flipped on the business itself.

Price Action: Caution on the Entry

Here's where the picture gets more nuanced. The stock sits at roughly $588, down about 26% from its 52-week high of $796 and down roughly 11% year to date. It's trading below both its 50-day moving average ($601.50) and its 200-day moving average ($633.25). The RSI (a momentum oscillator that reads 30 as oversold and 70 as overbought) sits at 46.5 - not in oversold territory, but edging lower. The MACD (a trend-following momentum indicator) is negative at minus 8.96.

That tells me the technical trend is still down, and there's no confirmation of a bottom formation yet. The selloff following the Q2 report was sharp, and selling exhaustion - where volume dries up and price stabilizes - hasn't appeared. I don't think investors need to rush in right now.

So What's the Setup?

The fundamental case for Meta hasn't cracked. Revenue is growing at nearly 30%, margins (excluding one-time charges) are intact, the ad moat is widening, and the forward P/E has compressed to near-market multiples despite that growth rate. The market is punishing Meta for spending aggressively on AI infrastructure, which is a real risk - but one that the current multiple arguably over-penalizes.

What I'm arguing is that the Suncoast trim and the broader institutional trimming pattern - these are background noise. The valuation gap between a 21.4x forward P/E and 28% revenue growth is the load-bearing evidence. The capex surge is a financing concern, not a demand collapse.

The action here is Buy, but I don't think the best risk/reward is chasing today's price. The stock needs to show selling exhaustion or a bear-trap bounce above the 50-day moving average before the entry gets cleaner. Watch for price stabilization around the $560-580 zone - a level roughly aligned with the prior consolidation area before the Q2 selloff. If the stock holds there and volume dries up, that's the signal to double down.

I would reassess the thesis if two things happen: quarterly ad revenue growth drops below 15%, which would suggest the core engine is genuinely losing steam, or if full-year capex exceeds $150 billion without a corresponding acceleration in operating margins. Until then, the disconnect between this valuation and this growth rate is the opportunity, not the warning sign.

Investors who've been waiting for a better entry into a high-quality growth compounder at a reasonable price shouldn't let this setup go to waste - but they also don't need to catch a falling knife. Patience on the timing, conviction on the numbers.