Totally Addicted to FCF
I remember the Facebook earnings call in October 2022 where their operating margins dropped to 20% and they almost zero'ed their FCF on the back of crazy hiring and the metaverse pivot (lol). Stock was already down tonnes then tanked post-earnings.
Markets drew a red line, all this spending with no reward, to which Zuck responded by controlling costs. FCF rose dramatically and the market absolutely loved it. Headcount control became the name of the game. Cost discipline was suddenly sexy.
And higher earnings x higher multiples = higher share prices.
META FCF (LHS) vs Share Price (RHS)
Ooft, good stuff right there.
All of a sudden, those negative cash flow stories or burn for growth narratives all turned it on full blast.
UBER FCF (LHS) vs Share Price (RHS)
Uber is probably the ultimate poster boy for this playbook. They used the negative rate environment to build a business moat that competitors simply cannot approach, then flipped the switch to profitability when rates turned. Check out the share price rally once the last negative quarter was cycled out (31 Dec 2022).
Companies replied to a high cost of capital when markets thought they never could. The dopamine hit was formed.
Totally addicted to cashflow
Markets rewarded financial discipline, those expanding multiples combined with growth equals stocks mooning.
But you'll note above that META FCF is falling again which is completely on the back of massive Capex spend around AI. And many ask, rightly so what about the returns on this capex as we've already been burnt once. So rather than excitement, there's more fear than anything else. People inevitably are calling this a bubble. Where is the killer app? When does this capex turn into profit?
Return on Investment
AI folk will have you know that jobs will be replaced. And everyone keeps questioning exactly who will be replaced? Doctors? Engineers? Accountants? CFOs? CEOs?
But tech c-suite got that dopamine hit in 2022, and they will do what they know best.
They want free cash flow.
They're totally and completely addicted to it. If you want a clear and simple answer to what all this AI capex spend is going to target: tech margins.
After capex, headcount represents the biggest expenditure line item for these companies. If you can replace the average worker who comes with the pressures of management, rising wages and all the other associated human complexities, by empowering a smaller number of workers to deliver the same level of output, the prize at the end is enormous. These companies already pay some of the highest average salaries in the world. If AI can maintain productivity levels with meaningfully reduced headcount, the margin expansion potential is staggering.
The AI Capex spend is effectively a buyback mechanism for remaining employees at the Mag7, higher profits shared among a smaller pool of workers.
(ChatGPT GPT5-Codex which is their agentic coding tool just dropped and its utterly friggin wild)
An Uncomfortable Conversation
My thesis for all this capex spend is purely higher Mag-7 earnings. This is a view I'm incredibly uncomfortable writing, particularly given how far markets have already run.
But the above thesis is not about controlling downside risk but rather thinking about answering that "AI killer app" and ROI question while also aligning ideas about getting convex exposure to AI in public markets. What if earnings keep rising? Who are the winners here?