Alphabet's P/E ratio fell from 28.6x in December to 21.7x in April. Revenue grew 22%. Earnings per share grew 82%. And the market is pricing the company lower relative to its own profits.
That is not how a company in trouble trades. That is how a company trades when investors are worried about something that hasn't happened yet - but might.
The more interesting question isn't whether Alphabet's growth story is intact. It is. The question is whether the market is right to be nervous about $185 billion.
Here is the setup. Google Cloud hit $20 billion in revenue last quarter - up 63% year over year. Cloud alone is now running above a $70 billion annual pace. Backlog is $240 billion. Gemini, the AI product, has more than 750 million users. Search revenue, the old cash register, grew 17% to $63 billion in the fourth quarter. This is a company where every major line item is accelerating.
And Alphabet just told investors it plans to spend $175 billion to $185 billion on capital expenditures this year - nearly doubling from $91.4 billion last year. The money goes into servers, data centers, chips. Infrastructure to feed the AI hunger.
$185 billion of capex is a number that makes your brain hurt until you see what other companies in the same space are doing. Microsoft and Amazon are spending comparably, and the total AI infrastructure buildout across all three will probably exceed $500 billion this year. Alphabet isn't the outlier. They're trying not to fall behind.
But here is where the puzzle is. Capex at this scale eats free cash flow. And the market already seems to be pricing that in, even though the spending hasn't fully hit the cash flow statement yet. That is what compressed the P/E.
The way to think about this is not as a cost problem but as a capacity problem. Cloud backlog grew 55% quarter over quarter in Q4. $240 billion in commitments is the kind of pipeline that tells you demand is outstripping supply. If Alphabet doesn't build, someone else will - and the window to be competitive in enterprise AI is not decades wide. It is months.
This is where most of the analysis ends. "They have to spend, so the spending is justified." But that is not the same thing as "the spending will pay off." A company can build exactly the right capacity and still not earn enough margin to justify the investment.
I haven't found clear evidence that cloud margins will sustain at current levels as AI inference demand scales. AI inference - running the model for each user query, not training it - is notoriously expensive. If Google Cloud's AI customers are consuming more compute than they pay for, the backlog becomes a margin problem rather than a growth story. This is a real risk, and the quarterly reports don't yet separate inference economics from training economics in a way that lets you test it.
Still, let me come back to the valuation. At 21.7x earnings, Alphabet is cheaper than it was three months ago despite reporting the kind of results most companies would kill for. The analyst consensus price target sits around $429, roughly 22% above the current ~$350 price. That kind of upside is unusual for a company this size unless either the market is pricing in a breakage that hasn't happened, or the consensus is wrong.
I suspect the market is doing something that often happens with capital-intensive transitions: it is pricing the spending as if it's a drag, not an investment. The market treats the capex as a cost that shrinks near-term cash flow. It's less willing to treat it as the infrastructure build that creates superlinear returns once capacity meets demand. Both readings are possible. Only one is right.
The test is simple. Watch free cash flow over the next two quarters. If the capex is absorbing almost all operating cash flow, the market's skepticism has legs. If free cash flow holds above $20 billion per quarter despite the spending surge, the P/E compression is a buying opportunity - the market is pricing a cash flow problem that the numbers don't support.
There is a second test that matters even more. Watch whether AI Overview - Google's AI-generated search results - starts eating search ad revenue. Publishers are already reporting major drops in organic click-through rates. If Google's own AI answers start cannibalizing the ad inventory that funds the capex machine, you have a company where the growth engine is undermining the cash engine.
That would be a real structural shift, not a quarterly noise event. For now, search ad revenue is still growing. But the mechanism that could break the model is live, and nobody knows yet whether Google can monetize AI-generated answers at the same margin as traditional search.
So here is where I land. Alphabet at 21.7x earnings, with 22% revenue growth and 82% EPS growth, is not a company the market has decided to love. It is a company the market is waiting to prove it can carry $185 billion in spending without breaking. The uncertainty is real. The capex is enormous. The inference economics are unproven.
But the market has already priced the pessimism. The question for an investor is whether they believe a company that prints $110 billion per quarter in revenue and dominates search can also win the AI infrastructure race - or whether that's two things that require different kinds of execution.
I lean toward yes. But I'm not betting on the growth. I'm betting on the gap between what the market fears and what the numbers actually show. If free cash flow holds and search ads don't collapse, the compression is the story - not the capex.