Google has been a public company since 2004. In all that time, through financial crises, a pandemic, and more failed product launches than anyone wants to count, it had never once burned through more cash than it brought in. Until now. AI did what two decades of catastrophe could not.
The Numbers Are Genuinely Staggering
According to BBC News, Alphabet reported negative free cash flow of $5.9 billion in the second quarter of 2026. That's the money left over after you pay for everything you need to keep the lights on and grow the business. When it goes negative, you are, in the plainest possible terms, spending more than you're keeping. For Google, a company that has essentially printed money since the Bush administration, this is a first.
Tesla joined the party too, posting negative free cash flow of $1.1 billion for the same quarter, its first negative showing in two years. Elon Musk's car company that stopped being primarily about cars a long time ago is now apparently also in the business of lighting billions of dollars on fire in the general direction of AI infrastructure.
Alphabet's stock dropped more than 7% on Thursday. Tesla's fell 13.5%. That's not a bad day. That's a rout. Combined, these two companies shed an eye-watering amount of shareholder value in a single session because investors looked at the earnings reports and apparently did not like what they saw.
What They're Actually Spending On
Google's chief financial officer Anat Ashkanazi told analysts on a call, as BBC News reports, that the company spent $45 billion in the second quarter alone. Sixty percent of that went to servers. The remaining forty percent went to data centers. If you're keeping score at home, that is $27 billion on servers and $18 billion on data centers in three months.
Alphabet now expects to spend up to $205 billion this year, a $15 billion upward revision from the estimate it gave just three months ago. Three months ago. They looked at their own projections from April and said, nope, we're going to need more. Ashkanazi's position on this is essentially that demand for AI capability is outpacing even these historic levels of investment, so the spending will continue.
Tesla, for its part, says it will spend up to $25 billion this year on what its filings describe as "unspecified projects." That phrase is doing a lot of heavy lifting. Twenty-five billion dollars of unspecified. Tesla's CFO Vaibhav Taneja told analysts the company is in "a big investment cycle" and that spending would probably increase further over the next three years. Cool. Great. Very reassuring.
The Question Nobody Has a Good Answer To
Here's the thing that is making investors sweat through their shirts: nobody has convincingly explained when any of this turns into actual profit. Alphabet's revenue is still growing, up 23% year-over-year to $119.8 billion for the quarter. The business is not dying. But the gap between what these companies are taking in and what they are spending is widening at a pace that is making serious financial people visibly uncomfortable.
Russ Mould, investment director at AJ Bell, told BBC News there is "a healthy degree of scepticism about the ability of these investments to generate a commensurate level of return." That is analyst-speak for: nobody is sure this works out. Rachel Winter, a partner at wealth management firm Killik & Co, put it more bluntly, saying the size of Google's spending figures produced genuine surprise and that the stock drop reflected real concern.
Sundar Pichai, Google's CEO, went on the call and described the AI moment as "early innings" and said the company's approach to generating returns was "disciplined." When a CEO uses the word disciplined to describe spending $45 billion in a single quarter, you are allowed to raise an eyebrow.
The Broader Race Nobody Knows How to Win
This is not a Google problem or a Tesla problem in isolation. The entire tech industry has looked at AI and collectively decided that the cost of losing the race is higher than the cost of running it, so everyone is sprinting with their wallets open and their eyes half-closed.
What we are watching, in real time, is one of the largest capital allocation experiments in corporate history. The bet is that whoever builds the biggest, fastest, most capable AI infrastructure right now will own the economic returns for a decade or more. That might be true. It also might be a story we tell ourselves while watching hundreds of billions of dollars disappear into server farms in the Nevada desert.
The uncomfortable historical parallel nobody wants to say out loud is the late 1990s, when companies spent like maniacs on fiber optic cable and internet infrastructure because everyone agreed the future was online and nobody wanted to be left behind. They were right about the future. A lot of them still went bankrupt.
The Dingo Take
Let's be clear about what is happening here. Two of the most valuable companies on earth just told their investors, in official financial documents, that they spent more money than they made last quarter. Google did this for the first time in its entire history as a public company. And the stated reason is that AI demand is so enormous, so insatiable, that $45 billion in a single quarter is not enough to keep up with it. If you find that either inspiring or terrifying, you are not wrong on either count.
The part that should genuinely concern people is the vagueness baked into all of this. Tesla's $25 billion in "unspecified projects" is a phrase that should not be allowed to exist in a public earnings report without a journalist following up until someone goes home. Google's CEO is talking about "frontier capabilities" and "extraordinary returns" in language so abstract it could mean anything. These are not small companies spending rounding errors. These are world-historical sums of money being committed to a technology that, by the CEO's own admission, has not yet been fully translated into actual products that actual users actually benefit from.
Maybe this all works out. Maybe in five years we look back at 2026 as the moment the foundation was poured for the next era of computing and everyone who worried was just scared of change. Or maybe we look back at it the way we look back at every other moment in history when enormous amounts of money chased a guaranteed future that turned out to be more complicated than advertised. Either way, the bill is already being written.
