Content About Stocktwits
No Image Available

'That’s a Wrap'- Big Tech's Quarterly Bow. ARD #131

Author
Michael Parekh AI: Reset to Zero
6 hours ago
19 MIN READ

Today’s theme, ‘That’s a wrap’, at least for the five big tech companies in the quarter, with a little more to go.

Amazon and Apple brought down the curtain, after Google last week and Meta and Microsoft on Wednesday. Two more to come. SpaceX on August 4, with earnings and a lockup release of over 900 million shares. Then the big one, Nvidia, on August 26.

A lot of implications, a lot of market volatility, both here and abroad. This is the AI Tech Wave in its fourth year after ChatGPT, and the ‘pendulum swings’ this week were the size of national economies.

Three events, each with my Take, then my Overall Take. Plus a Gadget AI on Zuck’s uphill climb with consumer AI agents, and two questions. Let’s get started.

(1) Amazon Pulls a Microsoft

MP TAKE: Amazon, the number one cloud provider with AWS, pulled a Microsoft, the number two with Azure. Investors like what they see at both, and both are investing super aggressively in AI infrastructure, hundreds of billions of dollars over the next few quarters.

And do not forget Google, the number three cloud provider, whose cloud business was up over 80 percent last week. All three cloud businesses are off to the races.

Microsoft had the biggest move of any stock in US history, about $450 billion added on Wednesday. Then Amazon came in and said hold my beer, and finished up over 15 percent on Friday. CEO Andy Jassy should be smiling

Let’s zoom into some details.

AWS revenue: $42.2 billion, up 37 percent. Jassy put the precise figure at 36.7 percent and called it their fastest growth in eighteen quarters. That is four and a half years. AWS is accelerating, not maturing.

AWS operating income: $16.6 billion, a 39.4 percent operating margin. So the biggest cloud business in the world is growing faster than it has since 2021, at nearly forty points of margin.

Total net sales: $200.6 billion, up 20 percent. First time over two hundred billion in a quarter. Total operating income: $27.5 billion, up 43 percent. Revenue up 20, operating income up 43.

And then the capex number. Amazon raised its 2026 capital budget to $220 billion, up from $200 billion. The reason they gave is the one I have been writing about here for weeks. Higher memory costs.

Jassy’s line is worth seeing in full.

“Even at that amount, we will still not have enough capacity to meet all of the demand we have in 2026.” And then: “this dynamic will also be true in 2027 too.”

Two hundred and twenty billion dollars is not enough. That is a demand statement in investment terms, and it gave investors a direct way to link the two.

The thing worth highlighting at the highest level is that investors are differentiating between the companies investing in this wave who also have huge enterprise cloud businesses, where you can see the metrics on revenue growth in the tens if not hundreds of billions. You can see that with these three, Amazon, Microsoft and Google, the global top three cloud businesses. Now showing AI momentum from businesses large and small.

You cannot see it as clearly with Meta or SpaceX, who are spending equivalent numbers, 100 to 150 billion dollars a year each, but do not have the enterprise business. They are hinting at it. I have called Zuckerberg’s version ZWS, Zuck Web Services, and Elon’s EWS, Elon Web Services. Neither is an official name. Placeholders for aspirations and possibilities.

There have been rumblings that Meta may rent excess data center capacity to Anthropic for ten billion or more. Investors have already seen the Elon playbook, two major deals ahead of the SpaceX IPO, Anthropic for $30 to $40 billion over several years and Google for $30 billion, cancellable by either side on three months notice. That detail matters.

Renting out capacity gets you a revenue bump. It does not get you millions of business customers tying themselves into your infrastructure for years at predictable rates on which you can hang long term multiples and valuations.

One more item. Net income was $62.6 billion. But $53.4 billion of that was a non-operating gain, mostly a mark-up on Amazon’s stake in Anthropic. The operating number, the $27.5 billion, is the real story, and it is a very good one on its own.

It took a few years, but Amazon has earned a spot at the front of the AI Tech Wave line. I called AWS the AI underdog back in 2023, and poised to offer AI enterprise services at scale in RTZ #925, before that was consensus.

And it showed something else. Having proprietary frontier AI models is not the only path to success. Being a global platform purveyor of a whole host of these models and applications, both open and closed, is also a path to possibly win. Once the AI business goes into global distribution and diffusion mode.

(2) Apple Skates Above the AI Spending Fray, but misses a tad on Services

MP TAKE: Four years after ChatGPT, in the world of AI agents, investors are saying that if we are going to invest in these companies aggressively, we want to see the results.

We are seeing the results on the cloud side, and those are being rewarded to the tune of hundreds of billions of dollars of volatility in a single day. And it goes both ways. We saw a vivid example with Apple.

Apple was skating above the fray on AI spending, but they have issues too, potentially, in the short term. Investors were spooked by services revenue, their second biggest business after iPhone, coming in a tad below expectations.

The numbers underneath were a record. Revenue $109.4 billion, up 16 percent. Tim Cook called it their strongest June quarter ever, with double-digit growth in every geographic segment. EPS $2.02, up 29 percent. iPhone $54.3 billion, up about 22 percent. Mac $10.4 billion, up 30. Greater China $18.8 billion, up 22 percent, which is a real reversal.

Services came in at $30.7 billion, up 12 percent. The street was looking for closer to fifteen. And that one line is what moved it.

The stock is down about nine percent. On a company near five trillion dollars of market cap, that is almost four hundred billion dollars of moves in a session. Swings equivalent to the GDPs of most countries. Opposite of Microsoft’s historic move the other way.

When we talk about AI bubbles and volatility, recognize that we are seeing some of the biggest single-day moves in our biggest companies, the ones that make up big chunks of our indices, and which we would normally never see in established mature companies. The reason is we are not in mature markets. We are in growing AI markets. Early innings. These companies are acting like startups, investing ahead. At unprecedented scale.

Now the apples-and-oranges number that makes all the difference between Apple and its big tech peers. Apple’s capital expenditure was $6.8 billion across nine months. Amazon’s was $173 billion in twelve. Apple is spending roughly two percent of what Amazon is spending, and it earned $29.8 billion in the quarter. That is what skating above the fray actually looks like.

Apple is doing the same investing-ahead that the others are, which is why I think this short-term drop, and this is not a stock recommendation, is an opportunity for Apple to build ahead of the fall introduction of their next generation of products and services.

Under new management, of course. Who have been at Apple for decades and are steeped in the ‘Apple Way.’ CEO John Ternus, at the company over twenty-five years, and Johny Srouji, now Chief Hardware Officer, who built Apple Silicon.

Customized bottoms-up with billions of users’ own data. Protected with unique safety, privacy and trust capabilities that only Apple can wield at scale.

Apple remains one of my two favorite consumer AI companies over the next two or three years. Google is the other.

It may be a wrap on Tim Cook’s spectacular fifteen-year run. He took over in August 2011. But it is the beginning of something special at Apple, in the AI Tech Wave to come.

(3) Mr. Market Swings Back, and Google Takes a Bow

MP TAKE: Mr. Market, as wise investors like Buffett call it, is swinging back and forth.

Google came back for a bow with a hinted fifteen billion dollar infrastructure deal with Anthropic, potentially providing much needed AI compute. Google was up five, six percent.

So huge volatility on this stuff. And not just in the US.

South Korea. I have written a lot about this. The two biggest companies in the memory world, SK hynix and Samsung, comprise over 50 percent of Korea’s index, and both have been up several hundred percent. Almost 15 million Korean citizens out of a 50 million population have been aggressively meme-stock day trading, with those two companies in particular.

Both reported great numbers. Both were down eight or nine percent after having been up a lot. And millions of Korean investors have faced margin calls.

Which are similar to the margin calls that made the news this week around Leopold Aschenbrenner. He ran $45 billion of public and private positions and had to fire-sale some of them to Ken Griffin’s Citadel.

What I am describing here is just the volatility in these markets, at the individual company level and at the country level.

So the quarter for now is ending, and everyone is taking a bow. And it is a wrap for the big companies so far this week.

MP OVERALL TAKE

The good news is we have seen Meta, Microsoft, Amazon and Apple this week, and we will see SpaceX and Nvidia soon.

These are all companies with relatively open-ended opportunities in the AI Tech Wave. Each has different pros and cons to its strategy, very different. They are all spending extraordinary amounts of money, with the exception of Apple, and with the exception of Nvidia, who are selling the picks and shovels to everybody else.

But this business is a long-term structural bull market, where the demand for AI compute runs as far as probably the end of the decade. The demand for memory is the same. The demand for infrastructure and compute is the same.

The models are evolving. There are a lot of tussles going on between open source and closed source, US versus China, and I have written a lot about that. Nvidia has taken big positions on it with the consortium they put together, which I wrote about a few days ago.

So the ‘AI Trade’ is very much a multi-trillion dollar trade that is very much intact. Goldman Sachs, my former firm, had a report with over five trillion dollars of infrastructure spend expected by the end of the decade.

But in the meantime, there is going to be a lot of volatility. And this week’s theatrical ups and downs are the case in point.

So buckle up. Next quarter is already underway. The curtain will rise again for that in a few months, and we will be back.

(Not Stock Advice)

Gadget AI. Zuck’s Uphill Climb on Consumer AI Agents

MP TAKE: I have written a lot about this before. At the end of his call a couple of days ago, Mark Zuckerberg literally said he is going to be focusing hard on consumer AI agents, for the three and a half billion users of his core properties. Instagram, Facebook, Messenger, WhatsApp.

Remember he tried a three billion dollar acquisition of the Chinese company Manus, which was disapproved by the Chinese government, so he had to unwind that. But they are leaning very hard into creating all kinds of agentic services for consumers, which I think is interesting.

We have seen a big market for AI agents in AI coding and enterprise, which is why Anthropic and OpenAI are going up against each other and competing hard for a market worth tens of billions of dollars in revenues for those companies.

It is not yet clear what the revenues look like on the consumer side. It is good to see Meta having the ambitions, but it is a long hill that will take many quarters.

He also hopes to see traction with AI smart glasses. He has a multi-billion dollar bet on those. He has already got about seven million of them out there, and hopes they will grow to 70 million, 700 million. I think that too is a very uphill climb, but we will see how it rolls out.

Question 1: Have I found useful activities for AI agents yet?

Answer: Yes. I think Claude Cowork is terrific for a lot of business and personal users. I use it every day.

But they are utterly disconnected from social media. And that is why I wanted to highlight the Meta and Zuck thing, because he is obviously hoping to leverage social networks. There is not a single thing I have done with AI agents that needed the social media part of things.

Maybe there will be applications further down the road. But I have found AI agents to be very useful more on the business side, in the context of daily professional work.

Question 2: What do I think of consumer AI agent applications like shopping and delivery services?

Answer: This is an area all the companies are focusing on very hard. Almost every demo you see of consumer AI agents is that it can shop for you, it can buy stuff on DoorDash for you and get deliveries.

I think those are great demos of products, and they are not real world cases for the most part. It is still far easier and more convenient to just open your phone and order whatever you want than to talk to an agent. They are more wrong than they are right.

This is early. We are the beta users, testing these things and improving them. I am excited about what is possible. It is not quite there. Right now it is a demo, and the companies involved are throwing a lot of spaghetti against the wall to see what sticks.

WRAP

Today’s AI-RTZ #1164, on why open source AI ‘rockets are hard.’ I did a lot of work on this one, more than usual, with a lot of new charts and tables I spent time creating. It looks at Nvidia’s focus on leading the open source AI side against China, and at the US open source profile of our companies versus theirs, and who is doing what. Given the issue is coming up in the papers around US versus China, especially in DC, you may find it of interest.

AI Ramblings Daily on AI-RTZ is here to think through AI and reset. Together.

Tomorrow, the Saturday Weekly Roundup. And Sunday, The Bigger Picture.

I want to wish you all a great weekend. Thanks for joining us today, AI Curious Folk. Stay tuned.

— MP

Full Source Reading

For the broader context, see the canonical sources for ARD 131, in today’s narrative order:

Event 1. Amazon Pulls a Microsoft

Event 2. Apple Skates Above the AI Spending Fray

Event 3. Mr. Market Swings Back, and Google Takes a Bow

Gadget AI. Zuck’s Uphill Climb on Consumer AI Agents

Also referenced. ARD #130, ‘Time to Crunch the Numbers’ · ARD #110, ‘Blind Ambitions Abound’

Standing reference. The AI Tech Wave and the AI Tech Stack.

Clips from today

Clip 1. Why AI Shopping Agents Are Still Just a Demo

Almost every demo you see of consumer AI agents is that it can shop for you, it can buy stuff on DoorDash for you and get deliveries.

MP Take: I think those are great demos of products, and they are not real world cases for the most part. It is still far easier and more convenient to just open your phone and order whatever you want, rather than talk to an agent. They are more wrong than they are right. This is early. We are the beta users of this, testing these things and improving them. So it is early. I am excited about what is possible. It is not quite there. Right now it is a demo, and the companies involved are throwing a lot of spaghetti against the wall to see what sticks.

Clip 2. The Best AI Agents Have Nothing to Do With Your Feed

I think Claude Cowork is terrific for a lot of business and personal users. I use it every day.

MP Take: But they are utterly disconnected from social media. And that is why I wanted to highlight the Meta and Zuck thing, because he is obviously hoping to leverage social networks. There is not a single thing that I have done with AI agents that needed the social media part of things the way I use it. Maybe there will be applications further down the road. But I have found AI agents to be very useful more on the business side, in the context of daily professional work.

Clip 3. Why the AI Trade Runs to the End of the Decade

This business is a long term structural bull market where the demand for AI compute runs as far as probably the end of the decade.

MP Take: The demand for memory is the same. The demand for infrastructure and compute is the same. The models are evolving. Yes, there are a lot of tussles going on between open source and closed source, US versus China. Nvidia has made big positions on this with the big consortium they put together. So the AI trade is very much a multi-trillion dollar trade that is very much intact. Goldman Sachs, my former firm, had a report with over five trillion dollars of infrastructure spend expected by the end of the decade. But in the meantime, there is going to be a lot of volatility, and this week’s theatrical ups and downs are case in point. So buckle up.

Clip 4. Zuck’s Uphill Climb on Consumer AI Agents

At the end of his call a couple of days ago, Zuckerberg literally said that he is going to be focusing hard on consumer AI agents for his three and a half billion users.

MP Take: They are leaning very hard into creating all kinds of agentic services for consumers, which I think is very interesting. We have seen a big market for AI agents in AI coding and enterprise, which is why Anthropic and OpenAI are going up against each other and competing hard for a market that is worth tens of billions of dollars in revenues for those companies. It is not yet clear what the revenues look like on the consumer side of the equation. It is good to see Meta having the ambitions, but it is a long hill that will take many quarters.

About AI Ramblings Daily (ARD), and AI-RTZ

Both are daily. Both are free. Both are about AI. But they’re different mediums carrying different messages.

AI-RTZ is the morning text, a deeper written take on one idea, published by at least 5 AM EST. Today: post #1164.

AI Ramblings Daily is the afternoon video + podcast, my ad hoc takes and perspective on the day’s AI issues and news flow, around 20 minutes, with short 1-2 minute clips for quick topic views. Today: episode #131.

Subscribe to either or both on michaelparekh.substack.com. They run as separate Sections you can opt into or out of.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here.)





Want the latest?

Sign up for Michael Parekh's Newsletter below:


Subscribe Here

More like this

BROWSE ALL ARTICLES