Alphabet earnings: Google’s $185 B Spend is the Real Story

Alphabet, Google’s parent company’s earnings this week did two things at once.

One - they reassured investors that Google is still the most effective money-printing machine on the internet. And two - they warned, quietly but unmistakably - that the next phase of that machine is going to be far more expensive to run.

Alphabet reported $113.8 billion in Q4 revenue, up 18% from a year earlier. Net income was $34.5 billion, up 30%. Earnings per share came in at $2.82. For the full year, revenue crossed $400 billion.

If you wanted a single datapoint to push back against the popular “chatbots will kill Google Search” storyline, it’s sitting right there in those figures: advertisers have not left, and users have not stopped searching. 

The behaviour change that dominates tech conference conversations still hasn’t shown up on Google’s main ledger.

What is interesting is that the quarter wasn’t only about proving Google’s resilience, the more consequential part of the release was what Alphabet said about spending.

The company told the market to expect $175–$185 billion in capital expenditure in 2026. That is not a “we’re investing in growth” number. It is a redefinition of what Alphabet is willing to become: a more industrial business, with heavier fixed costs, because it believes the next platform shift is built on compute capacity.

This is the crux of the story: Google is not collapsing. It is preparing for a world in which staying on top requires building and owning more of the physical layer of AI.

Search is still the centre of gravity

For years, Alphabet has had the luxury of being able to fund long bets—self-driving cars, health science, moonshots—because its core product sits in a privileged position. Search catches people at the moment they declare intent. It is hard to beat that moment for monetisation.

This quarter, that structure still held, as search revenue rose strongly again, and the broader Google Services business continued to throw off cash. Even YouTube, now a mature product rather than a growth miracle, remained a solid contributor. None of these read like a business being rapidly displaced, but they read like a company that still controls the default pathways of online life.

That matters because “AI disruption” is often narrated as if consumer behaviour changes in clean jumps: search today, chat tomorrow. In practice, people adopt new tools without abandoning old ones. The new ones are added and blended, and they keep using the that is fastest when they’re tired, rushed, or simply habitual. 

Google’s earnings suggest that reality is still on its side.

Cloud is where AI becomes real money

If Search is the cash engine, Cloud is the bridge into the next phase.

Alphabet’s Cloud unit grew fast again - around 48% year-on-year to roughly $17.7 billion in Q4, and profitability improved. 

Cloud has been the company’s long-running insecurity next to Amazon and Microsoft. These results, at least for now, show something closer to momentum: meaningful revenue expansion, with profits that suggest the business is no longer just subsidised ambition.

Cloud is also where the AI boom stops being a consumer trend and becomes corporate procurement. Enterprises don’t buy “AI vibes.” They buy capacity, contracts, and integration. That is why Cloud’s growth matters more than almost any product demo. It is a sign that demand is showing up in the most conservative place: budgets.

And that demand is why Alphabet is about to spend so much money.

The capex guidance is the tell

Alphabet’s 2026 capex guidance - $175–$185 billion, is the kind of number that changes your mental model of the company.

For most of the last decade, Google looked like a software firm with extraordinary margins that happened to run enormous infrastructure. In the AI era, the infrastructure is no longer background, it is the competitive arena and edge.

Large-scale AI is a physics problem disguised as a software story. You need chips, networking, data centres, power, supply chains that don’t buckle. Apart from this, you also need the ability to deliver capacity reliably to customers who will simply leave if all this isn’t delivered.

Alphabet is effectively saying: we see the bottleneck, and we are going to build all this and more.

There is a rational defensive logic here. If compute capacity is scarce, the worst outcome is being a major platform with more demand than you can serve. That is how customers drift to rivals. That is how a cloud business loses credibility.

But the risk is also straightforward: once you start spending like an infrastructure company, you expose yourself to infrastructure economics. This means higher fixed costs, more depreciation, more sensitivity to cycles, and more pressure to keep utilisation high. It doesn’t mean Alphabet becomes unprofitable.

It means the company’s famous elegance - its ability to grow without adding proportionate physical weight, simply becomes harder to preserve.

A more honest read of “AI threatens Google”

The simplistic version of this story says AI will replace Search. The earnings argue that’s not happening in any immediate, measurable way.

The more honest version is subtler: AI may not kill Search, but it may change what Search costs.

If users come to expect direct answers, summaries, and actions, those interactions can be more compute-intensive than traditional query pages. If advertisers want new formats, Google has to experiment and if regulators scrutinise defaults and bundling, distribution becomes more contested. 

None of that requires Search to collapse. It just requires Alphabet’s margin profile to compress over time as the company spends more to defend the same behavioural default it already owns.

That is the “AI tax” Alphabet is signalling with its capex.

This is a global infrastructure story, not a Silicon Valley feature story

Spending at this scale doesn’t stay inside a quarterly narrative.

It runs through advanced chip supply chains that are still geographically concentrated. And through energy systems and grid politics. 

This kind of spending also runs through construction capacity, permitting, and local resistance to data centre expansion.

 It runs through government thinking about “strategic compute” in the same way governments think about ports, pipelines, and semiconductor fabs.

When a handful of firms decide that AI competitiveness equals physical buildout, the industry starts to look less like software and more like utilities. That raises barriers to entry. It concentrates power. It makes the winners harder to dislodge, and it makes the losers less relevant.

What this quarter really says

What it’s important is that Alphabet’s earnings do not describe a company losing its grip on the present - the numbers are too strong for that.

Nevertheless, they do describe a company that thinks the next era will be won by whoever can afford to build it—literally.

That’s the tension investors should sit with after this quarter: Alphabet can fund the transition because Search still works. 

But the transition will change the company’s shape. It will make Alphabet heavier, more capital intensive, and more exposed to the constraints of the real world.

In the old Google story, the internet was the infrastructure and Google was the layer that monetised it. In the next story, Google is spending like it wants to be the infrastructure too.