ZM Keeps Printing About $1.9 Billion a Year in Cash. The Market Just Called It Peaked

On Tuesday Zoom reported that revenue from its enterprise customers grew at its fastest rate in three years and raised its full-year sales and profit targets. The stock fell about 8% the next day. That is the setup worth understanding, because the market that sold it was not confused. It anchored on the one number that threatens the whole "Zoom is back" story: guidance for free cash flow this fiscal year came in below what the company actually generated last year. The crowd's read was blunt — the cash machine has peaked.

The cash-flow path says something else. Start with what Zoom actually produces. In fiscal 2026, which ended in January, the company generated $1.92 billion of free cash flow — cash from operations less capital spending — up 6.4% from the prior year, on $4.87 billion of revenue. That is almost 40 cents of cash for every dollar of sales. This year it guides to $1.78 billion to $1.82 billion, and management has lifted that range since May, when it was pointed at $1.70 billion to $1.74 billion.

The reason Zoom converts so much revenue into cash is structural, not heroic. Subscribers pay ahead for a service that is pure software: gross margin runs about 77%, and capital spending is roughly $75 million a year, about 1.5% of sales, because the company rents computing instead of building it. There are no factories, no fleets, no inventory. So a roughly 40% non-GAAP operating margin is mostly cash by the time it hits the bank. That is what "prints cash" means for a software business — revenue collected, a small slice spent on infrastructure, nearly everything else left over.

Where the cash goes matters as much as the fact that it arrives. Zoom holds about $7.2 billion of cash and marketable securities against roughly $2.3 billion of convertible notes — call it about $5 billion of net cash. On top of that sits nearly $3.8 billion of strategic investments, the centerpiece an early stake in the AI lab Anthropic. And Zoom returns the cash: it completed a $2.7 billion buyback program, repurchased about 3.7 million shares last quarter, and has roughly $1.3 billion still authorized. The share count keeps shrinking.

One caution before anyone quotes a headline valuation. The Anthropic stake produced a $1.6 billion paper gain in the quarter, and GAAP "earnings" of $5.15 a share — a single-digit price-to-earnings multiple that is an illusion, a mark-up on the balance sheet rather than operating profit. The honest earnings lens is the $6.10 or so of non-GAAP EPS guided for the year, which puts the stock near 16 times. The cleaner lens still is the cash itself.

Now the part that should hold a buyer honest. Free cash flow this year is guided flat at best against last year's outturn — not compounding — and the third-quarter revenue guide arrived below what the Street expected. Management points to higher AI-compute costs as a near-term drag on margins. The online consumer business — still more than a third of revenue, at 38% — grew just 0.6%. And the company's net dollar expansion rate is 99%, meaning existing enterprise customers on average spend slightly less with Zoom each year; growth now comes from new logos, not from the installed base growing. None of that is a broken thesis, but it is the honest reason the multiple has stayed humble, and it is why the stock, after trading as low as $70.70 in the past year and spiking above $114 in June — some of that a one-day 11% pop when Anthropic filed to go public — now sits near $98.

Here is what has to happen for the "cash machine has peaked" read to be the stale one. Fiscal 2026's growth was only 4.4% after two consecutive years stuck at 3.1%, so the acceleration is recent and specific. Enterprise revenue — now 62% of the total — grew 7.8%, its strongest rate in three years, and management pointed to longer, multi-product contracts. Backlog tells the same story: total remaining performance obligations rose 14% to about $4.5 billion, and the slice expected more than a year out grew 25%. The AI monetization is showing up in the plumbing: paid AI appeared in nine of the top ten customer-experience deals, licensed AI users grew 125% year over year, and the count of customers paying more than $100,000 reached 4,625, up 8.2%. None of that shows up in a depressed cash multiple yet, because it has not been earned yet. The market is still pricing the old risk profile while the operating setup is already getting cleaner.

The arithmetic at about $98 is simple. A market cap near $28.7 billion against the guided $1.8 billion of free cash flow is roughly 16 times — or about 12 times enterprise value to free cash flow once the pile of net cash is subtracted. For a business whose enterprise engine just hit a three-year high and whose cash yield is above 6%, that is a reasonable price for the setup, not a stretched one. AInvest's aggregate signal labels the stock a Buy, and analyst targets moved only modestly after the report — Bernstein nudged its target to $108 from $103 and held it at Market Perform. Consensus is not the point. The point is that the multiple embeds almost no improvement at all.

Here is the falsifiable version. The case works if free cash flow holds at or above $1.8 billion through the coming quarters while enterprise growth stays above roughly 7% and net dollar expansion finally crosses 100% — evidence the installed base is growing again, not just the invoice list. It is wrong if the cash guidance slips again, if enterprise growth rolls over, or if the AI expansion stays decorative instead of showing up in the top deals. I can be wrong again, but this is a setup where the market's doubt is testable with a number. The cash is already real. The question the next two quarters answer is whether it is a peaked metric or the base for a return to growth — the difference between a 16 times multiple that stays a value trap and one that quietly re-rates as the enterprise engine keeps feeding it.