Greetings from Kansas City and Cedar Rapids. Pictured are a couple of chili cheese sausages from one of my favorite Cedar Rapids lunch spots – The CR Midnight Smoker. We don’t do a lot of food pics but really like their BBQ (that’s from a KC BBQ lover) and have seen them grow from a food truck to a full blown restaurant.
This week, after a shorter than usual market commentary, we are diving into capital spending for several of the hyperscalers: Microsoft, Amazon, Meta and Alphabet/Google. Implications for the wireless industry are also discussed.
The Fortnight That Was
Under all the headlines, the Fab Five quietly topped $2 trillion in market cap increase for 2026, gaining $373 billion last week. This was driven by Apple (+$72 billion), Microsoft (+$166 billion) and Meta (+$219 billion). Apple (+25%) and Meta (+14%) currently lead the YTD percentage growth range, and Meta is now on the verge of a $2 trillion market cap milestone.
Meanwhile, the Telco Top Five slowly continued their attrition this week (-$12 billion). As we mentioned in the previous Brief, there is a mere $3 billion separating the value of T-Mobile and AT&T (and $19-22 billion separating these two from Verizon). What a difference a year makes.
Meta Connect 2026 was last week, and the keynote did not disappoint. Fresh on the heels of their Muse (personal AI agent) launch, the company announced several updates to their wearables lineup and teased the Muse Charm as part of a larger screenless device push. The VR glasses pictured nearby were the big hit from the show and demonstrate how much can be packed into a mere 100 grams. Now comes the hard part: generating high and sustained interest from the developer community. The company announced IMAX support (great for The Odyssey fans) and a deal with Disney+.
The concept of personalization was prevalent throughout the keynote. VR glasses are the beginning of that journey, but they also dovetail into the entire Muse strategy. If Meta were to announce a VR relationship with YouTube (including their creator community), that would be big news. As it stands, however, VR will move the needle only slightly while Muse will steal the show.
Muse is going to be a very big deal, and we believe that many monetization opportunities exist beyond collecting a percentage of each transaction facilitated through the app. The app has been out for a week, and Mashable already has an article on Top 10 prompts (here). We think there will be a community (Facebook group) focused on “best prompts to save money” and “best prompts for fitness/fashion/food finds.” With a large base to quickly build personal algorithms, Muse gets larger much faster.
Amazon is taking notice. To no one’s surprise, the online commerce giant has blocked Muse from their site (CNBC article here). But we were quick to notice that Meta is building close relationships with Walmart, Best Buy and Sephora (article from The Information here). We are reminded of the early days of smartphone app stores when iOS refused to allow access to Google Voice (and removed Google Maps as the default mapping app while they launched Apple Maps). Like that short-lived incident, we think that the Amazon/Meta dustup will be resolved soon. If Muse inquiries were not rapidly multiplying, there would be nothing for Amazon to worry about. Muse is working, and Amazon is now faced with a difficult decision.
There wasn’t a lot of new economic data released this week, but what was released pointed to a rapidly growing economy. Chief Economist Chris Williamson’s comments in the latest S&P Global US Flash PMI release say it all:
“To put the growth surge in context, barring the spike in demand following the opening up of the economy after the COVID-19 lockdowns, the latest improvement in business activity is the greatest recorded since early 2015. Business is clearly booming now in both manufacturing and services.
However, this growth is being accompanied by some of the most severe supply chain bottlenecks seen in the near-two-decade survey history if the pandemic is excluded, with companies also reporting increasing problems finding suitable staff. Backlogs of work are consequently rising sharply. While this accumulation of uncompleted orders bodes well for the further expansion of output and capacity in the coming months, it also indicates that companies are developing more pricing power and hence is a worry for the inflation outlook.
Firms’ input costs have meanwhile jumped in September at the steepest rate for four years, with fuel and transport costs spiking higher thanks to the rise in oil prices seen during the month, which will add further to the upward pressure on selling prices and inflation in the coming months.”
The third quarter tends to be the fastest growing quarter of the year, and 2026 is no exception. The 5.0% real GDP growth rate currently estimated by the Atlanta Fed (here) would imply an 8.1-8.3% nominal growth rate. Something to consider when your favorite large cap telecom company boasts of 3.0% annual growth.
All this debt-driven growth fuels the fire of higher interest rates (great for retirees who are looking for additional Certificate of Deposit interest income, but bad for their grandkids looking for their first mortgage). Nearby is a chart from Macrotrends showing a 60+ year trend for the US 10-year T-Bond. Bad news for companies who have moderating cash flows looking to rollover debt, but great news for bondholders who are looking for higher yields.
Where is AI Capital Going?
There are many headlines for AI infrastructure spending, and we thought it would be good to provide a concise view of where that capital is being spent as we head into 3Q earnings and 2027 outlooks. Specifically, it’s important to distinguish between capital commitments (e.g., $3 billion announcement to build a data center) and specific quarterly outlays (which are detailed in 10-Q and 10-K reports and other public releases). Here are summaries for Amazon, Microsoft, Alphabet/Google and Meta based on publicly available information:
Amazon. Here is their statement on capital expenditures from their latest 10-Q:
“Cash capital expenditures were $31.4 billion and $53.1 billion during Q2 2025 and Q2 2026, and $55.6 billion and $96.3 billion for the six months ended June 30, 2025 and 2026, which primarily reflect investments in technology infrastructure (the majority of which is to support AWS business growth) and in additional capacity to support our fulfillment network, both of which we expect to increase in 2026.”
The leader in cloud infrastructure is continuing to spend money on third-party chipsets (as well as the development of their Tranium chips). They are also growing their core and edge data center footprints and making power commitments (the 2025 Talon Energy commitment as well as their acceleration of investment in Small Modular Reactor (SMR) technology through their investment in X-Energy). As a reminder, AWS has a major relationship and an ownership interest in Anthropic (Claude) and has committed two Gigawatts of Tranium training capacity to OpenAI (announcement here).
AWS is also the largest cloud infrastructure provider in the world with an estimated 28-30% market share. That base is driving increased capital expenditure as their demand grows. We think that many of these enterprises will also quickly grow their use of Claude, Gemini, and OpenAI applications.
Amazon also announced on the most recent earnings call that they anticipate spending $220 billion for capital expenditures in 2026. Second half spending should increase 10-15% over the first six months. We expect their cap ex trends to parallel revenue growth for the foreseeable future.
Alphabet/Google
Alphabet spent $80.6 billion in capital expenditures in the first half of 2026. Second half investments are expected to be about 50% higher. In the second quarter, Alphabet spent 60% of its capital on chipsets with the remainder on data center infrastructure (including IRUs on fiber that connects each data center). CFO Anat Ashkenazi indicated that 2027 cap ex levels would be “significantly higher” than 2026.
Alphabet also disclosed the following in their most recent 10-Q:
“As of June 30, 2026, we had material purchase commitments and other contractual obligations totaling $811.0 billion, of which $200.7 billion was short-term. These purchase commitments primarily relate to costs for technical infrastructure and inventory through long-term supply agreements and open purchase orders. Additional contractual obligations include commitments for content licenses and energy take-or-pay contracts.”
With $811 billion in commitment backlog, the level of capital spending will not slow down anytime soon. Even though Alphabet designs their own chipsets (called Tensor Processing Units or TPUs – manufactured by TSMC), they also remain one of NVIDIA’s largest customers. While Google holds a $2+ billion investment in Anthropic, it does not have a business relationship with or ownership stake in OpenAI.
Microsoft
Like Alphabet, Microsoft has a large capital commitment backlog. Here is the latest information from their annual report (fiscal year ending June 30, 2026):
Even with the debt commitments removed (which are financing the AI-related investment), Microsoft still has $660-690 billion in commitments remaining. Assuming most of these commitments are related to chipsets, there are a lot of reasons to smile if you own NVIDIA, AMD or Intel. Microsoft’s largest commitments, however, are for operating leases related to data centers. Some of those power Azure, while others power other Microsoft platforms (including internal use).
Because of Microsoft’s investment stake and partnership with OpenAI (over $13 billion in either direct investment or Azure credits), their data centers, power, and customer-related capital spending are quickly used as soon as they are received. 85-90% of all 2026 capital spending is going towards existing Azure customers or upcoming AI infrastructure needs.
Meta
Finally, Meta is driving the highest capex as a percentage of current revenues (51% of 2Q revenues were spent on capital – nearly double the percentage (26%) of Amazon). This is not sustainable but highlights how important recent investments are to the company.
As mentioned above, Meta’s entire strategy evolves around personalization and privacy. Individual preferences require storage and are constantly being revised by applications such as Muse. Add in the complexities of hardware information gathered in real-time (from glasses and pendants), and it’s clear that Meta will need a very large internal network to create the next social media platform.
The company spent $31 billion in 2Q 2026 and projects that they will spend more than $130-145 billion this year (so an additional $80-95 billion in the second half). Like their peers, the company spends a lot of capital on internal and proprietary chipset design and manufacturing. Unlike their peers, however, Meta does not currently derive revenues from leasing capacity to commercial customers (they remain open to doing that in the future if/ as needed, however, as this article details). That’s what makes the 2026 estimated spending amazing. All the 2026 and 2027 ramp is for internal revenue growth (from applications like Muse discussed above).
Bottom line: Chipsets, land acquisition, data center construction, materials, labor and fiber. That’s a pretty good list of where the hundreds of billions of capital expenditure are going. Despite constraints emerging with suppliers, spending should increase in 2027. This is why we are encouraged by the outlooks for value creation for the Fab Five (including Apple) but wonder how wireless keeps up. Currently, it looks like significant wireless spectrum is going to be needed at lower latencies to mirror the fiber-based in-office/ in-home experience. The heavier the task, the greater the network constraint. One sporting event with strong AI dependencies could bring wireless networks to their knees. Our biggest worry (outside of demand) is the creation and sustainability of experience across the high-count fiber and spectrum-constrained wireless networks.
That’s it for this week. We will have our second iteration of Apple iPhone 18 Pro and Pro max charts out this afternoon after the Chiefs game. The next two Briefs will be focused on 3Q wireless earnings (spoiler alert: T-Mobile will still lead the pack) In the meantime, if you have friends who are interested in being notified each time we publish a Brief, please have them sign up at www.sundaybrief.com.
Go Chiefs and Sporting KC!
Important disclosure: The opinions expressed in The Sunday Brief are those of Jim Patterson and Patterson Advisory Group, LLC, and do not reflect those of CellSite Solutions, LLC, or Fort Point Capital.
