Empty Wallet: Big Tech’s AI Spending Pushes Cash Flow to the Brink

Cash is flowing out of Big Tech faster than ever as AI investment accelerates. Backlogs and promises must now carry the weight cash once did.

KEY TAKEAWAYS

• Alphabet produced negative $5.9 billion in quarterly free cash flow–its first negative quarter since going public. The decline came despite strong revenue and operating income because capital spending nearly doubled.

• Meta’s quarterly free cash flow fell to $784 million, while Amazon’s trailing figure dropped dramatically. AI infrastructure spending is the common thread connecting all three companies.

• These businesses are investing heavily in data centers, chips, and power contracts that could produce returns for years. The unanswered question is how long shareholders must wait for those returns to arrive.

• Big Tech is increasingly using debt, equity, joint ventures, and potential infrastructure leases to support its spending. That introduces new financial obligations and reduces the independence created by self-funding operations.

• Income statements can look spectacular while cash flow deteriorates underneath. Backlogs and contracted revenue may support the investment thesis, but they remain promises rather than cash already collected.

MY HOT TAKES

• The AI boom should be evaluated through cash flow, not merely revenue growth, earnings, or headline backlog figures. The cash flow statement is where the true cost of the infrastructure race becomes visible.

• Negative free cash flow does not mean Alphabet is in financial distress. It does mean the company’s historically dependable financial model has changed in a way investors should not dismiss.

• Wall Street is giving Big Tech considerable latitude because these companies have pristine balance sheets and decades of credibility. That trust is valuable, but it is not unlimited.

• The word “eventually” is carrying the AI investment case. If demand weakens, prices fall, or technological change shortens the useful life of today’s infrastructure, expected returns could look very different.

• Big Tech is temporarily living on tomorrow’s expected revenue rather than today’s available cash. That makes even the strongest technology companies subject to the oldest financial rule: cash is king.

• You can quote me: “Right now, for a handful of the most trusted names on Wall Street, the cash is flowing out faster than it is flowing in.”

Empty wallet. Back in the early 1980s a trend emerged in suburban New York shore towns. Surf shops began selling nylon wallets that literally velcroed shut. From a practicality standpoint, they offered significant improvements over the cowboy-themed leather billfolds most young men carried at the time. The most important feature--for me at least--was the velcro, which ensured that I wouldn't lose any of the wallet's important contents. I had a paper library card and...wait for it...wait for it...MY VALUABLE CASH. The cash would increase if I earned more and would only decrease if I spent it. The goal was simple: keep the cash flowing IN faster than it was flowing out. This was my first lesson that cash is king.

That lesson would stay with me as I came to Wall Street and eventually went on for graduate studies. While in graduate school, one of my most memorable classes was taught by a professor who had left academia to start one of the first electronic trading firms (recognized as the precursor to today’s high-frequency trading firms) and returned with an atypical, oversized, gold watch, and renewed interest in teaching. I credit that professor with inspiring me to go on for my doctorate (he pitched it to me in a vestibule outside a lecture hall one late night), maybe even more importantly, solidified that on Wall Street: cash is–indeed–king. He introduced me to Marty Fridson who wrote what appeared to be a simple book entitled Financial Statement Analysis. The book was unique in that it was not written by an academic but by an accomplished Wall Street bond trader who worked on high yield bond trading desks from legendary Solomon Brothers to Merrill Lynch, to Paine Weber. Because he was a bond analyst, he took a unique approach which could be easily contrasted to the bible of financial analysis from Graham and Dodd. Fridson focused heavily on what was–at the time–FASB (SFAS) 95 cash flow statements. Sure, he covered income statements and balance sheets, but his emphasis was on the cash flow statement which was relatively new at the time. WHY? Because you can't pay your bills without cash! You are probably wondering why I burned through two paragraphs to introduce this topic. And, I am sure that you know that cash is important for your own finances. It has always been important on Wall Street, but right now, I would argue to you that it has not been this important since the late 1980s. Story time is over, let's dig in.

Free cash flow. It sounds like jargon that only bond guys care about, but it is really as simple as my old velcro wallet. Take the cash a company generates from running its actual business–operating cash flow, in the parlance–and subtract what it spends to keep the machine running and growing (capital expenditures, or CAPEX–new servers, new buildings, new chips). Whatever survives that subtraction is free cash flow. It is the money left over to buy back stock, pay dividends, pay down debt, or simply sit in the vault as a rainy day fund. Positive, and a company is generating more cash than it needs to sustain and grow itself. Negative, and the company spent more building its future than its current operations brought in the door. It is, in the simplest terms I can offer, an empty wallet–even if the velcro is still holding it shut.

For Alphabet, that number went negative last month for the first time since the company went public back in 2004. Twenty-two years of unbroken positive free cash flow, and this quarter it produced NEGATIVE $5.9 billion. Not because the underlying business faltered–quite the opposite. Revenue grew 24% to nearly $120 billion. Google Cloud grew 82%. Operating income came in at $40.8 billion. But capital expenditures nearly doubled year over year to $44.9 billion, chewing through more cash than the $39.1 billion operations generated. Alphabet also raised its full year capex guidance yet again, to a range of $195 to $205 billion, and warned that 2027 will run higher still. For the second straight quarter, the company bought back zero shares of its own stock–something it had done routinely, and heavily, for years.

Meta did not technically cross into negative territory, but it came close enough to smell it. Free cash flow fell to $784 million for the quarter–down from roughly $12 billion the company had averaged over the prior two years, and its lowest level since 2022. Call it a rounding error away from negative. Amazon may have gotten there first: earlier this year it reported trailing twelve month free cash flow of just $1.2 billion, down from $25.9 billion a year earlier–practically nothing for a company that size, and some trackers using slightly different capex assumptions already put the trailing figure into outright negative territory. The common thread across all three: capital expenditure tied almost entirely to AI infrastructure–the data centers, the chips, the power contracts required to keep pace in a spending race none of them seem willing to lose.

This is a remarkable admission from companies long treated as the gold standard of corporate financial health–the ones that historically generated so much cash they didn't know what to do with it all, buying back their own stock by the tens of billions and sitting on cash hoards that made central banks jealous. Analysts have a defense ready, and it is not unreasonable: data centers carry useful lives measured in decades, and the chips inside them, while shorter lived, are being deployed against contracts that are already signed. This is not, they argue, cash being lit on fire. It is cash being converted into an asset that will, eventually, throw off cash of its own. Maybe so. But the word "eventually" is doing an enormous amount of work in that sentence, and Wall Street has a long memory for companies that asked shareholders to trust the "eventually."

So how does a company operate without cash flowing in fast enough to cover what is flowing out? The same way any of us does when the velcro wallet runs empty: you borrow, or you sell something. Alphabet raised $49.6 billion issuing stock and preferred shares during the quarter, and another $20.3 billion through senior notes–corporate bonds, in plain English. Its long term debt has more than doubled since December, to $98.2 billion. That is Wall Street's version of a payday loan, except the collateral is a pristine credit rating and thirty years of trust with bondholders instead of a car title. Meta, for its part, is exploring ways to monetize the very infrastructure it is racing to build–Zuckerberg told analysts the company is fielding offers to lease out excess compute capacity at a premium to what it paid, and it is pursuing cloud ambitions of its own, including a $14 billion data center joint venture with BlackRock. In other words: sell what you built before you have even finished paying for it.

None of this means these companies are in financial distress, not yet, and maybe not ever. Alphabet still sits on roughly $240 billion in cash and marketable securities, and its trailing twelve month free cash flow, smoothing out this one lumpy quarter, remains a healthy $53 billion. But a company that must issue debt and equity to fund its own operations has, by definition, ceded a measure of control. It now answers to bondholders and equity investors, not just to its own cash register. It must convince Wall Street, quarter after quarter, that the spending will eventually convert into revenue large enough to justify it. Alphabet points to an eye popping $514 billion cloud backlog as its evidence. Amazon points to over $225 billion in committed AI infrastructure revenue. Promises, in other words, are being asked to carry weight that cash used to carry alone. And promises, unlike cash, can be revised downward the moment demand softens or a competitor undercuts on price. That is the trade these companies have made, whether they say it out loud or not: certainty today for the possibility of something much larger tomorrow.

This is precisely the sort of thing Marty Fridson would have flagged decades before anyone thought to ask the question out loud. A rosy income statement, dressed up with accounting gains (and we have been seeing lots of those lately 😉), can make a quarter look glorious even while the cash flow statement tells a very different story underneath--Alphabet's headline profit this quarter, remember, was inflated by roughly $99 billion in unrealized gains on its stakes in other companies. That was Fridson's entire point in writing a book about cash flow when almost nobody else was bothering to. Earnings can be adjusted, defined, and redefined in a dozen creative ways. Cash cannot. It is either sitting in the account, or it isn't.

Which brings me back to my velcro wallet, and to that paperback copy of Financial Statement Analysis still sitting on my shelf, spine cracked from a semester of late night reading more than three decades ago. I have not opened it since. I have not needed to–because the lesson stuck the way only a teenage lesson about your own cash running dry can stick. Whatever these companies say about backlogs, and contracted revenue, and an AI future large enough to justify it all–and it may well be, in time–the wallet does not lie in the meantime. Right now, for a handful of the most trusted names on Wall Street, the cash is flowing out faster than it is flowing in. That does not make them reckless. It makes them, for a stretch, exactly like the rest of us: living off of what is coming rather than what has already arrived. Cash, even for Alphabet, Meta, and Amazon, is still king. It always has been. Keep your eyes open.

YESTERDAY’S MARKETS

Yesterday, the S&P 500 slipped 0.18%, the Dow Jones Industrial Average fell 464 points (0.85%) to 53,885.10, and the Nasdaq Composite eased 0.06%, snapping the Dow's recent record-setting win streak. The 10-year Treasury yield rose roughly 5 basis points to about 4.67% as stronger jobless claims and productivity data trimmed hopes for near-term Fed rate cuts. West Texas Intermediate crude jumped 3.37% to $77.75 a barrel on renewed uncertainty over the Strait of Hormuz reopening timeline.

NEXT UP

• Nonfarm Payrolls (July) is expected to come in at 80k after a June print of 57k.

• Unemployment Rate (July) may remain unchanged at 4.2%.

• Next week: lots more important earnings announcements in addition to Consumer Price Index / CPI, Producer Price Index / PPI, Retail Sales, and University of Michigan Sentiment. You better check back in next week–if you want to remain the life of the party.