META top 10 posiotn --FH26 Results and commentary - Core Strong- Doubling Down on Growth

 META FH26 Note

My thoughts

There is little doubt that Meta is a complex story, and markets do not like complex stories. In addition to this, markets usually do not price in a derisked scenario for a stock if capex is still growing strongly; it will wait for signs of an inflexion in capex growth or clearly strong returns on that capex, and we are not seeing that yet here.

I like to think of Meta as three stories in one, the first being the traditional business infused with AI, which is growing phenomenally well, perhaps the best example of what AI can achieve in a traditional business. The second, which has some overlap with the first, is Meta Superlabs' goals and developments. MSL will be the driver of new products and expansions for Meta. That business is heavy on spend and generating little revenue so far, but promises a large payoff if successful. The last business is Reality Labs, which continues to generate losses; Meta signalled a slowdown in losses, but they remain large (TTM $19B), even for a group the size of Meta; that capital is sorely missed as the prize has moved to MSL, not RL.

The table below shows the extent of the new initiatives' costs; we have revenue growth driven by the traditional business up 38% yoy, a phenomenal result on a large revenue base and showing the positive impact of AI improvements, plus other initiatives. The operating margin, however, adjusted for one-offs, has fallen from 45% to 39%. IMO this is not negative operating leverage but the cost of new growth (hopefully) yet to appear. Splitting the expenses between the two would be very useful but difficult due to the overlap, but we have commentary that much of the growth in expenses is yet to produce significant income. The question then becomes: if, when, and how much. That is the uncertainty overhanging the stock. The huge increase we see in Capex in the graph further down for 2026 is likely to take at least a year or more, likely two, to generate significant returns. That will test the patience of the market.

Revenues
FH26	117112	38%
FH25	84830
Opy 26	41647	10%
Opy 25	37996
excl one offs		19%
Margin 26 adj	38.6%
margin 25	44.8%

 

Perhaps the best indication of Meta’s progress in the traditional business is that before the start of AI, Meta’s advertising business was about three-quarters the size of Google’s. By next year, at current run rates, it will overtake Google, and Google has been growing strongly over that period.

As can be seen in the graph below, ad prices and volumes were in a trade-off until about the start of 2024, when both prices and volumes began to grow in the double digits and have continued to do so. Meta is using AI to analyse the vast data available to it, including a DAU of half the world's population, to improve the product across the board. In brief, this includes analysing user intent and serving up individually interesting and new content (better matching), increasing time on site, improving advertisers' ability to create quality ads, and better matching those ads with users. In fact, it goes further than that, with Meta tools being used by advertisers to run their whole ad campaign. With MSL moving into the area, that's where the overlap occurs. The outcome of these efforts is improving conversion rates and increasing supply and pricing of ads as they become more profitable to the client. The increase in supply comes from better engagement leading to more time on site, which opens the way for splicing ads into viewing times. A virtuous cycle, as Meta’s bread-and-butter clientele are millions of small businesses that are hooked into the ecosystem. However, with possible positives but upfront costs, Meta's ambitions do not end there.

AI is not only improving products and client experiences; it is improving Meta operations themselves, with 8k employees (10-11% of total) being moved on over the next period. The mix of tasks within Meta is changing with AI. Productivity and efficiency measures continue to climb for the group.

The second business (to my thinking) is MSL’s plethora of new products being developed and introduced. Some of these are quite different to traditional ad selling. The products include Meta One subscriptions for more tools, and AI features across the apps; Forum, a new standalone sellers marketplace; Agents with a personal agent being a big initiative as well as business agents.

Meta’s vision of a personal agent is based on the data they have on social media, and the aim is to make it easy to use, so that billions of customers can use it without having developer skills. Meta’s focus should be on agents that are easy to use and available for mass distribution. This makes sense.

WhatsApp, after a lengthy period of growth but little revenue, is now changing, with AI infused into it, and is targeted as a large revenue earner as the business messaging app moves into more business tools, with catalogues, CRM and inventory management mentioned. Monetisation is occurring.

Broader business agents are also being talked about, with Meta’s enormous clientele of small businesses being the target. WhatsApp, Messenger and Instagram are all being used for the development of business agents.

“And over time, we'd like to build this into a business-in-a-box service that can help you start and run a whole business using Meta's platforms.”

Monetisation was described as occurring through subscriptions, volume-based pricing and success-based pricing. There will likely be an evolution in monetisation as the business evolves.

To build these products takes an awfully large amount of compute and capex. Meta did mention that they are open to selling any excess compute, due to the huge prices currently being bid for compute. That appears to be a short-term, revenue-generating timing issue, as the requirement for core business compute will grow over time.

The third business, Reality Labs, is focusing on glasses, and not so much on the metaverse. Whether AI glasses become the iPhone killer remains a distant dream at this stage. Of course, Zuckerberg would love his own platform and not be reliant on Apple, so the losses are likely to continue as this is potentially more than just a new product.

After a phenomenally strong Q1, Q2 was a letdown, Q3 revenue guidance was lowish, and one-off charges, regulatory $2.1B, and severance charges $1.18B dented margins.

As for Meta having the leading frontier models, the following is relevant.

Longer term, it's harder to predict the exact usage scaling curves, but we believe that our distribution advantages will give us the opportunity to serve AI products that are valuable for everyone, both our 3.6 billion users and millions of businesses. This should be true regardless of whether our models are on the frontier, but we believe that being on the frontier will unlock new markets and opportunities for which we may need additional compute.

My reading of this statement is that it is too soon to give up on the frontier. The product possibilities are still evolving, and Meta should be competitive enough and have sovereignty over their models to direct them to where they are of most use and importance for Meta. Over time, once product and distribution are settled, the cost of being at the frontier becomes easier to define in cost/reward terms. Paying perhaps a few tens of billions as insurance protection on over a trillion market cap is the bear view; having control of the vast opportunity is the bull view.

At $556 SP, Meta trades at just over 10X TTM OCF. That OCF carries some of the cost of running losses in RL and also the cash cost of the new growth initiatives. There is also a substantial capital spend. The thesis for Meta is that, with 3.6B DAU worldwide (as well as millions of embedded SME customers) and with the ability to increase time on site by improving engagement, etc, the opportunity is vast for AI-focused products to make inroads. In fact, AI looks particularly well suited to the huge data load that Meta owns. Effective execution is required, and Meta has some form of delivering product into the +billion-number base. The products are in the early stages and encompass both consumer and small business. The cost to play, being high capex, is currently pricing the stock very lowly.

Despite having an extremely strong base biz, there is reasonable optionality apparent, including RL, WhatsApp business plans, API services, Productivity services, new Apps, broader business agents, Personal agents, possibly building a neo-cloud biz (I think this is a second prize) and progressing MSL, which could unlock something much bigger as AI evolves. Some of these are moderate in size, some not, some easy to envisage, some harder. At the current pricing, which is wearing some of the cost, not much success is priced in. IMO.

A negative that is hard to price is the regulatory risk that floats along with the stock. Meta stated that European growth was slower than other geographies due to restrictions on personalised ads. Meta, along with other US big tech, has attracted huge fines for various infringements. Possibly this is BAU for big tech, a European tax by stealth. Section 230 immunity in the US is a strong pillar of defence in that jurisdiction. The rule protects social media companies from being held legally responsible for what users post on their sites. Therefore, a mixed regulatory bag for the companies, GOOG and Meta.

Although Meta has close to no net debt, debt is increasing but unlikely to threaten financial ratios anytime soon. There is significant off-balance-sheet debt that looks spread out over the next decade at about $68Bpa. That is a large number and highlights the need for capex to produce results. However, the spreading of this spend over the long term makes a financial crisis for Meta unlikely.

There is a portfolio construction issue here. AMZN, MSFT, GOOG and Meta are all spending big on capex. We are seeing some positive signs of ROI, but it is early and spend continues to escalate. The overall exposure to this group in the portfolio should be monitored. Meta arguably has the most complex story and the most unclear path to a profitable end than the others. The portfolio aggregate exposure to the group should be capped, IMO.  

Summary

The thesis is that Meta will find a way to monetise the huge capex and opex spend into its enormous global user base and SMB’s. The ultimate Roi on the spend may take some time to become apparent.  Tracking the progress will not be that easy, as new products will be lumpy and take time to scale. Successive poor releases will be a thesis breach. We can take some comfort in that the market is not pricing Meta for any type of attractive growth, which is a positive. Overall, I like the blend of strong base biz, some significant optionality and low pricing at this stage, although it could take a couple of years to prove the case. Another way to look at Meta is a value story with a lot going on under the hood that could significantly change the narrative.

 

Graphs below, solid and solid.