This is the biggest week of the year for the AI money story. Microsoft and Alphabet report on Wednesday, October 28, after the US market closes. Amazon follows on Thursday, October 29.
Together with Meta, which reported last week, these companies are spending close to $800 billion a year on AI: data centres, chips, power and the land to put them on. To give that number a human scale, it is more than $2,000 for every person in the United States, every year.
Most previews will tell you what analysts expect for each share. This one asks a different question: who is paying for all this, and does any of it reach your wallet? The short answer is yes, in three ways most people do not notice: through their retirement savings, their subscriptions and, increasingly, their electricity bill.
The week at a glance
| Company | When | What analysts expect | The number that matters |
|---|---|---|---|
| Microsoft | Wed, Oct 28, after the close | revenue about $90.6-91 billion, up about 17% | Azure cloud growth against guidance of 44-45%; spending above $50 billion this quarter |
| Alphabet (Google) | Wed, Oct 28, after the close | revenue about $127-128 billion, up about 25% | Google Cloud growth, expected around 90% after 82% last quarter; cash flow |
| Amazon | Thu, Oct 29, after the close | revenue about $202 billion | AWS growth, the swing factor for the whole week |
Apple's date had not been confirmed at the time of writing, and Nvidia, which sells the chips all of them buy, is expected to report around November 18.
Why this week is a test
For two years the market rewarded every increase in AI spending. That patience is thinning, and the reason is cash.
| Signal | What happened |
| Meta | reported revenue up 26% last week, but spent $34.9 billion in one quarter and its free cash flow turned negative |
| Alphabet | free cash flow was reported negative at about minus $5.9 billion in the second quarter |
| Microsoft | free cash flow of $19.6 billion last quarter, still positive but down 23% |
| IMF | warned this month that weaker-than-expected AI profits could trigger a market correction |
In plain terms: these are the most profitable companies in history, and they are now spending so fast that some of them take in less cash than they pay out. That is fine if the spending turns into revenue soon. It is a problem if it does not.
The evidence for revenue is real. Microsoft's contracted future business, its remaining performance obligations, was about $678 billion last quarter. Google Cloud's backlog was about $514 billion. Those are signed contracts, not hopes. What investors want to see this week is that cloud growth is fast enough to make the spending look like investment rather than a habit.
I covered the same question for Meta in Meta's October results, and the bigger picture in is AI a bubble, and what to do with your money.
Who pays, part 1: your retirement savings
If you have a 401(k), an IRA or almost any broad US index fund, you own Microsoft, Alphabet, Amazon, Meta and Nvidia already, and together they make up a very large share of the US stock market. That means a big part of your retirement balance is riding on one idea: that this $800 billion will pay off.
This is not a reason to panic or to trade around Wednesday. It is a reason to know your exposure. Two practical checks:
- Look up the top ten holdings of your main fund. Most fund pages list them. If the same five or six AI spenders dominate, you are more concentrated than the word "index" suggests.
- Decide whether that suits your age and goals. Someone 30 years from retirement can ride out a correction. Someone five years away may want a different mix. That is a question for a licensed adviser, not a news article.
How AI-focused funds are built, and what they actually hold, is in the AI ETF guide.
Who pays, part 2: your subscriptions
AI is expensive to run, and the companies are passing part of the cost to customers.
| Where you may notice it | What happened |
| Microsoft 365 | Microsoft raised US consumer prices by $3 a month in January 2025 when it added Copilot to Personal and Family plans |
| Chatbot subscriptions | the main plans now range from about $8 to $20 a month for everyday use, and $100-200 for heavy users |
| Ads you see, and pay for if you advertise | Meta's average price per ad rose 14% last quarter |
| Cloud tools small businesses use | software built on AWS, Azure or Google Cloud carries the cost of the computing underneath |
None of these is dramatic on its own. Together they add up, and the trend is clear: AI is being built into products you already pay for, and the price is rising with it. If you are deciding what is worth paying for, ChatGPT Go vs Plus and what people actually pay for AI are a good starting point.
Who pays, part 3: your electricity bill
This is the newest and most argued-about part of the bill, and the honest answer depends on where you live.
| Study | What it found |
| Rutgers, New Jersey State Policy Lab (24 states, 2014-2024) | homes near data centres paid about $0.90 to $2.09 a month more, against an average bill of about $164; the authors note this covers the period before the latest AI build-out |
| IEEFA, on the PJM grid (13 eastern states) | extra data centre demand added about $9.3 billion in capacity costs for 2025 and 2026, roughly $16 a month more on an average home bill in Ohio and $18 in western Maryland |
| Federal Reserve Bank of Dallas | data centres have raised wholesale power prices by 2% to 6% nationally, more where they cluster, and AI could raise generation costs by 20% to 30% by 2028 |
So for most households the effect so far is small, around a dollar a month. In data centre hotspots it can already be $16 to $18 a month, close to $200 a year, and every major forecast says the pressure will grow as the facilities now being built switch on. That is why towns such as Trumbull, Connecticut have started passing moratoriums on new data centres.
If your utility is in a fast-growing data centre region, it is worth checking your bill's capacity or transmission charges, and following your state utility commission's rate cases. That is where the decision about who pays, households or the data centre owners, is actually made.
What to watch on Wednesday and Thursday
Microsoft: Azure growth versus spending. If Azure grows at or above 45% while spending passes $50 billion in a quarter, the market will likely see the spending as justified. A slowdown would raise questions fast.
Alphabet: cash. Google Cloud is growing faster than anything else in big tech. The question is whether Alphabet's cash flow recovers or stays negative, and whether it restarts share buybacks.
Amazon: AWS. It is the largest cloud business and the one most investors are unsure about. Its growth rate on Thursday will set the tone for the whole sector.
Everyone: 2027 plans. Meta said it would give its 2027 spending plans in January. If any of these three gives a number for next year this week, that will be the headline.
The honest take
What strikes me about this week is how far the AI story has moved from the chatbot on your phone. The interesting numbers are no longer users or benchmarks. They are cash flow, electricity and backlog. AI has become an infrastructure business, like railways or power grids, and infrastructure is always paid for by somebody.
Some of that somebody is the shareholders, who are accepting lower cash today for growth tomorrow. Some of it is customers, through subscriptions and prices. And some of it, in a growing number of places, is ordinary households through the power grid. None of that makes the investment wrong. It does mean the AI boom is no longer something that happens only in Silicon Valley.
So before Wednesday: do you know how much of your retirement fund sits in these five companies?
This article will be updated with the actual results after October 28 and 29.
Frequently asked questions
When do Microsoft, Alphabet and Amazon report earnings?
Microsoft and Alphabet report after the US market close on Wednesday, October 28, 2026. Amazon reports after the close on Thursday, October 29.
How much is big tech spending on AI?
Microsoft, Alphabet, Amazon and Meta together are spending close to $800 billion a year on AI infrastructure such as data centres and chips.
Do AI data centres raise electricity bills?
For most US households the measured effect so far is small, around a dollar a month. In data centre hotspots on the PJM grid, one analysis puts the increase at about $16 to $18 a month, and forecasts expect the pressure to grow.
How does big tech AI spending affect my 401(k)?
Broad US index funds hold large positions in Microsoft, Alphabet, Amazon, Meta and Nvidia, so their results and the success of their AI spending can move your balance. Whether that concentration suits you is a question for a licensed adviser.
What should I watch in Microsoft's earnings?
Azure cloud growth against guidance of 44-45%, and capital spending, which Microsoft guided to above $50 billion for the quarter.
What should I watch in Amazon's earnings?
AWS growth, which analysts see as the swing factor for the whole sector this week.
Sources: Mitrade: Big Tech's $800 billion AI bill faces its first real test; SentiSense: big tech earnings this week; Capital Insider: AI capex is the number under the number; Rutgers NJ State Policy Lab: are data centers raising your electric bill?; Yahoo Finance: the AI data center boom is raising power costs; Fox News on the Dallas Fed data center research; The Neuron: AI news, October 26.
