Source: Mark Zuckerberg (@finkd) via X, July 6, 2023 | META close: $291.20
Source: Mark Zuckerberg (@finkd) via X, July 2026 | META trading near $610
The Bull Case for Meta's Next All-Time High
On July 6, 2023, Mark Zuckerberg posted the "we're the same" Spider-Man meme, needling Twitter days after Threads launched. META closed that day at $291.20. Everyone had a good laugh, Threads got written off as a clone with a countdown clock, and most of the internet moved on.
Three years later, almost to the week, Zuckerberg is posting something with considerably less punchline potential: a benchmark table showing Meta's Muse Spark model trading blows with Anthropic, Google, and OpenAI on nine separate evaluations. META is trading around $610 as we write this. That is a gain of roughly 110% since the meme, and the second chart is a lot less funny than the first one, which is usually a good sign for a stock.
We are not going to pretend the tweeting habits of a chief executive are an input into our models. But it is worth noticing what changed in the gap between these two posts. The 2023 version of Zuckerberg was on the back foot, mocking a competitor's app for staying alive. The 2026 version is publishing benchmark tables against Anthropic and OpenAI like a company that expects to be taken seriously in the conversation. That shift did not happen by accident, and it shows up in more places than his timeline.
Meta's market capitalization sits around $1.7 trillion. Before arguing about whether the AI business justifies a premium on top of that, it is worth doing the boring exercise of just listing what the company actually owns, because most of the market seems to have stopped doing this sometime around 2022 and never started again.
None of that is a growth story on its own. It is closer to a floor. Roughly $160 billion a year in advertising revenue is generated by pointing this attention at businesses, and that engine is what funds everything else in this memo, including the part where Meta spends more on AI infrastructure than most countries spend on their power grids and nobody has to ask where the cash is coming from.
If you pulled up Meta's most recent quarter and stopped at free cash flow, you would have found a genuinely bad-looking number: $784 million against an estimate of $848 million, and a trailing-twelve-month figure down nearly 15% year-over-year. Read in isolation, that looks like a company whose cash generation is cracking. Read next to the rest of the statement, it looks like the opposite.
Operating cash flow actually grew 26.8% on a trailing-twelve-month basis, to $130.3 billion. Revenue has compounded at 20-22% annually for three straight years, with the exception of the well-known rough patch in 2022. The business generating the cash is not the problem. The problem, if you want to call it that, is investing activities, where cash outflows more than doubled year-over-year to roughly $139.4 billion on a trailing basis, almost entirely the AI infrastructure buildout.
A company that generates more operating cash every single quarter and chooses to spend nearly all of the incremental gain on infrastructure is not in distress. It is making a bet. Whether that bet pays off is a fair question. Whether the falling free cash flow number by itself tells you the bet is failing is a different question, and the honest answer is no, it does not.
For a long time, the joke about Meta and artificial intelligence was that the company was mostly buying GPUs and hoping. That joke is getting harder to make. On the Artificial Analysis Intelligence Index, which aggregates nine separate evaluations, Meta's Muse Spark scores 62, tied with Claude Fable 5 and ahead of GPT-6 Astra, GPT-5.6 Sol, and Grok 4.6. It trails the top two entries on the board, and we are not going to pretend otherwise. Coming from a standing start against labs that exist purely to build frontier models, tied with one and ahead of two others is not a participation trophy.
What makes this a market-share story rather than just a leaderboard story is pricing. Meta has been giving away access to Muse Spark's contributor tier at ten cents per million input tokens, a fraction of what comparable frontier access costs elsewhere. One account tracking this called it a scorched-earth strategy, and looking at the pricing table, that is not an exaggeration so much as a description.
The result shows up directly in usage share. On a widely tracked coding-agent inference marketplace, Meta's share climbed from effectively nothing in mid-July to 43.2% by early September, ahead of DeepSeek at 28.5% and dwarfing OpenAI's 0.6%. You can debate whether subsidized pricing is a sustainable long-term strategy. You cannot debate that it is working right now, and working fast.
Buried in Meta's own second-quarter filing with the SEC is a line called Family of Apps other revenue, the kind of line most people skip on the way to the advertising number. In the second quarter of 2025, it was $583 million. In the second quarter of 2026, it was $1.007 billion. That is a 73% year-over-year increase, and it crossed a billion dollars in a single quarter for the first time, driven almost entirely by WhatsApp's paid business messaging and Meta Verified subscriptions.
WhatsApp's paid messaging business crossed a $2 billion annualized run rate by the end of 2025, the first time Meta has quantified the business at that scale, and Barclays estimates WhatsApp and Threads together could add up to $25 billion in incremental advertising revenue by 2027. For nearly a decade, the standard joke about WhatsApp was that Meta owned the world's largest messaging app and had somehow never figured out how to make real money from it. That joke has an expiration date, and the expiration date appears to have already passed sometime in the last two quarters without very many people noticing.
Here is a comparison that is currently making the rounds and is funny enough to repeat, with the caveat that we do not think it should be taken as rigorous math. Anthropic's private valuation sits near $1.9 trillion. OpenAI's sits near $1.2 trillion. Meta, a public company generating $60 billion of quarterly revenue and building models that trade blows with both of them, carries a market capitalization of $1.7 trillion for the entire business, advertising empire included. Taken literally, that would suggest the market is assigning Meta's AI effort a value of roughly zero.
Taken honestly, this is comparing two private, venture-funded labs whose entire valuation is a bet on future AI revenue against a diversified public company whose valuation is mostly still anchored to an advertising business the market understands extremely well. It is not an apples-to-apples comparison, and we are not going to pretend it is. What it is useful for is a gut check: a public market that is famously willing to pay up for AI optionality everywhere else has, so far, priced almost none of that optionality into Meta specifically. Whether that gap closes because Meta's AI business gets properly valued, or because Anthropic and OpenAI's private marks eventually meet gravity, is a genuinely open question. We think the more likely direction is the first one.
Put the pieces together and the setup looks like this. A four-billion-person distribution network that was never really the question. An advertising engine still compounding at 20 percent-plus that funds the entire AI buildout without touching the balance sheet. A model that has gone from irrelevant to competitive with the two most expensive private AI companies on the planet in roughly eighteen months. A WhatsApp business that quietly crossed a billion dollars a quarter while everyone was busy arguing about capex. And a stock still sitting more than twenty percent below its own all-time high.
We do not think this requires a heroic assumption about where artificial intelligence goes from here. It requires believing that a company executing this well, on this many fronts simultaneously, eventually gets priced like it is executing well, rather than like it is still apologizing for 2022. Over the next six to twelve months, we expect that repricing to happen, and we expect it to take the stock through its previous high rather than stall in front of it.
Disclosure: Double Eleven Capital holds a position in Meta Platforms, Inc. (META) as of September 5, 2026.
Zuck's cooking. The rest of the industry is watching the pan. We would rather be at the table than reading about it later.
