---
title: "Google is releasing its earnings report tonight, I'm only watching one number"
type: "Topics"
locale: "en"
url: "https://longbridge.com/en/topics/42852809.md"
description: "Last night (Thursday early morning Beijing time), Google released its Q2 earnings, kicking off the earnings season for the Magnificent Seven. Everyone will look at profits. But I only focus on one number — capital expenditure (capex): this quarter, how much money did Google actually dump into 'heavyweights' like data centers and AI chips? Why only this one number? Because it determines not just Google's fate, but whether the next phase of funding in the entire AI supply chain will continue to flow. And this wasn't a last-minute decision. At the end of my previous article on HBM, I said we would only watch three signals going forward..."
datetime: "2026-07-22T11:27:34.000Z"
locales:
  - [en](https://longbridge.com/en/topics/42852809.md)
  - [zh-CN](https://longbridge.com/zh-CN/topics/42852809.md)
  - [zh-HK](https://longbridge.com/zh-HK/topics/42852809.md)
author: "[熊猫校长Ming](https://longbridge.com/en/profiles/16386316.md)"
---

# Google is releasing its earnings report tonight, I'm only watching one number

Tonight (early Thursday morning Beijing time), Google released its Q2 earnings report, kicking off the earnings season for the Magnificent Seven.

Everyone will look at profits. **But I only focus on one number — capital expenditures (capex)**: this quarter, how much money did Google actually dump into 'heavyweights' like data centers and AI chips?

Why only this one number? Because **it determines not just Google's fate, but whether the money in the next phase of the entire AI industry chain continues to flow downstream.**

And this wasn't a last-minute idea. At the end of my previous article on HBM, I said we would only watch three signals next, with the first being 'cloud providers' capex wording'. **Tonight, the first signal is revealed.**

![图片](https://pub.pbkrs.com/uploads/2026/d29843da5e1986faf86937e16c5c94c0?x-oss-process=style/lg)

## **I. First, let's clear a mine: The 'profit' this time might be fake**

Let's point out a pit first, so you don't get fooled by the numbers tomorrow.

Google's earnings per share (EPS) this time are likely to look particularly good, even ridiculously so — Bank of America has given a profit forecast of $8.38, nearly three times the mainstream market expectation (around $2.9).

Why such a big difference? Because Google holds over 10% stake in the star AI company Anthropic, and Anthropic's valuation has surged recently. According to accounting rules, this paper unrealized gain must be counted into Google's profit for this quarter.

To put it bluntly: **this part of the profit is 'paper wealth', not earned through real cash from the core business.** Cheering for one-time paper gains is like celebrating a salary raise because your house price went up — two different things.

So the more it's times like this, the more you should take your eyes off 'profit' and look at that number that won't lie: how much real money they are actually dumping into the future.

## **II. Why I only look at capital expenditures**

To explain this, we have to go back to that 'map of money' I've been talking about.

In the game of AI, money flows station by station downstream: at the very top are a few giants who pay (Google, Microsoft, Amazon, Meta), crazily dumping money to build AI; money flows down to chip makers like NVIDIA and TSMC; further down, it flows into HBM memory, advanced packaging, optical modules, power...

**Google stands at the very upstream of this river — it is the payer, the faucet of the entire AI industry chain. Its capital expenditure is the switch of the faucet:**

Turn the switch wide open (capex increased): The river water is still sufficient, and downstream TSMC, memory factories, and power companies all have food to eat;

Turn the switch back (capex loosens): No matter how lively the downstream is, once the water source shrinks, the whole river will eventually follow suit.

Advertising is Google's own business, while capex is its bet on the entire future of AI — and also the rice bowl for thousands of downstream companies. As for whether the people receiving the water downstream can keep the profit, that's a matter of the 'three locks', which we discussed in previous articles, so we won't expand on it today.

## **III. To what extent is this faucet currently opened?**

Four numbers, feel them:

**1\. Full-year guidance of $180-190 billion.** Last year was still $91.5 billion, doubling in one year. This is already one of the most money-burning tech infrastructures in history.

**2\. In Q1 alone, they dumped $35.7 billion, more than doubling year-on-year (+107%).**

**3\. Free cash flow almost halved year-on-year (-47%).** It's not that they aren't making money, but too much of the earned money is being dumped into construction sites. The scene I described when talking about TSMC — giants' 'paper profits' are turning into 'rebar and cement on construction sites' — Google has also reached this stage. Those who make money and those who can keep cash in hand are becoming two different groups.

**4\. The harshest is this: Not enough money to burn, so they started financing.** In June, Google raised $84.75 billion by issuing additional shares, specifically for AI infrastructure; this year it also issued about $40 billion in bonds. **A company that earns tens of billions in net profit annually is now selling stocks and borrowing money to increase stakes — this isn't ordinary 'optimism', it's betting on its balance sheet.**

Did it win the bet? At least for now, the bet is paying off: Cloud business grew 63% last quarter (Microsoft Azure 40%, Amazon AWS 28%), operating profit margin jumped from 18% to 33% in a year, backlog orders nearly doubled, reaching $460 billion. Google CEO Sundar Pichai also had a quote on the earnings call: **'If we could meet demand, cloud revenue would have been even higher.'** Hear that — demand is 不缺， capacity is lacking. That's why capex cannot stop.

## **IV. But this earnings report is released in a new sentiment**

The background has changed.

For the past two years, the market talked about 'compute shortage' — whoever dumps more money is right. But in recent months, the story has become 'will there be overcapacity?' — Google's stock price has already pulled back 15% from its May high, and the market is starting to scrutinize the return on every dollar spent with a picky eye.

At the same time, two new variables have appeared:

**First, architecture is starting to 'save'.** This week, it was reported that Google is developing a new chip codenamed Frozen v2 — directly 'freezing' parts of the Gemini model structure into silicon, claiming it can spit out 6-10 times the tokens per unit of power. Remember that race I talked about regarding HBM — architecture optimization vs. usage growth, who runs faster? This is 'architecture optimization' accelerating.

**Second, TPU is starting to make money on its own.** In April this year, for the first time in history, Google sold its self-developed TPU chips as independent products externally, and partnered with Blackstone to start an AI cloud company. Self-developed chips have transformed from 'internal cost-saving tools' into a 'second revenue curve' — this is a brand new chapter in the story of cloud provider vertical integration.

So tonight's capex number is not being released in an environment of 'dump money and rise', but in an environment of 'prove it to me'.

![图片](https://pub.pbkrs.com/uploads/2026/d83760f7e117fbf0198c617d97035e39?x-oss-process=style/lg)

## **V. Tomorrow morning, match yourself to one of three scenarios**

Ordinary people don't need to guess ups and downs or stay up late watching the market. When you wake up tomorrow, just match yourself to one of three scenarios:

**Scenario 1: Capex increases again, or maintains $180-190 billion with strong wording** ('Demand is too fierce, continue to add')→ **Green Light**. The main gate continues to open, and the story of the entire downstream chain — chips, HBM, packaging, power — continues.

**Scenario 2: The number remains unchanged, but the rhetoric changes** (words like 'optimize efficiency' and 'prudent investment' emerge)→ **Yellow Light**. The shift in rhetoric often leaks the wind direction earlier than the shift in numbers.

**Scenario 3: Guidance is lowered, or the tone for 'significant increase' in 2027 becomes vague** → **Red Light**. The upstream water source tightens, and no matter how lively the downstream is, the valuation of the entire AI hardware chain must be re-evaluated.

By the way, check two corroborating factors: Can cloud growth maintain above 60% (market consensus expectations have already opened to 67%, and the hidden threshold is close to 70%); Is free cash flow continuing to halve? The former proves that spending money yields returns, and the latter measures the cost of this gamble.

## **VI. Ultimately**

Tomorrow, the question shouldn't be 'how much did Google earn', but: **Is this AI money bag willing to continue pouring out?**

Don't ask if the earnings report looks good, ask which way the faucet is turned. This answer determines not just Google, but the entire chain downstream.

**What I research is never AI itself, but the money in the AI era — where it comes from, where it flows. Because capital never disappears, it just constantly migrates, and the starting point of all this migration is this faucet tonight.**

When the numbers come out tomorrow, I'll be back 第一时间 to watch with you which way it turns.

Before that, I want to hear your judgment first: **Do you think this Google earnings report tonight will be** **Green Light,** **Yellow Light,** **or Red Light?** Will capex continue to go up, or start to loosen? Leave your answer in the comments, and we'll check together tomorrow morning.

See you at the earnings report.

Real trading observation, does not constitute investment advice.

— Panda Principal Ming | AI Infrastructure Research

Related research targets:  
$Alphabet - C(GOOG.US) $Alphabet(GOOGL.US) $Micron Tech(MU.US) $Sandisk(SNDK.US) $SK Hynix(SKHY.US) $NVIDIA(NVDA.US) $Broadcom(AVGO.US)

### Related Stocks

- [GOOG.US](https://longbridge.com/en/quote/GOOG.US.md)
- [GOOGL.US](https://longbridge.com/en/quote/GOOGL.US.md)
- [MSFT.US](https://longbridge.com/en/quote/MSFT.US.md)
- [META.US](https://longbridge.com/en/quote/META.US.md)
- [TSM.US](https://longbridge.com/en/quote/TSM.US.md)
- [GGLS.US](https://longbridge.com/en/quote/GGLS.US.md)
- [GGLL.US](https://longbridge.com/en/quote/GGLL.US.md)
- [MSFL.US](https://longbridge.com/en/quote/MSFL.US.md)
- [MSFO.US](https://longbridge.com/en/quote/MSFO.US.md)
- [MSFD.US](https://longbridge.com/en/quote/MSFD.US.md)
- [MSFU.US](https://longbridge.com/en/quote/MSFU.US.md)
- [MSFX.US](https://longbridge.com/en/quote/MSFX.US.md)
- [MSFY.US](https://longbridge.com/en/quote/MSFY.US.md)
- [SKHY.US](https://longbridge.com/en/quote/SKHY.US.md)
- [NVDA.US](https://longbridge.com/en/quote/NVDA.US.md)
- [AVGO.US](https://longbridge.com/en/quote/AVGO.US.md)
- [GOOGN.US](https://longbridge.com/en/quote/GOOGN.US.md)
- [MU.US](https://longbridge.com/en/quote/MU.US.md)
- [SNDK.US](https://longbridge.com/en/quote/SNDK.US.md)