Meta Platforms: The iPhone Moment?

«The biggest risk is not taking any risk. In a world that is changing really quickly, the only strategy that is guaranteed to fail is not taking risks.»

– Mark Zuckerberg, Founder and CEO of Meta Platforms

arvy's teaser: 19 years ago a man in a black turtleneck sweater walked on stage with a glass rectangle and changed how billions of people live. Now, another man walked on stage with a keychain. This week: Meta’s bet that agents replace apps, the 3.6-billion-user pipe behind it, and the bill that comes with it.

9 January 2007.

Steve Jobs walks on stage and holds up a glass rectangle.

No keys. No stylus. A screen and a finger. Within a decade it becomes the most profitable product in history and turns Apple into the most valuable company on earth. That became the «iPhone moment»: shorthand for the point where a technology breaks into the mainstream, changes daily behaviour and resets the standard for a whole industry.

Last week, somebody tried to do it again.

On 23 September, Mark Zuckerberg walked on stage at Meta Connect and did not present a better phone. He presented the thing that is meant to replace it.

Muse Charm (chart 1).

A keychain-sized device, roughly an AirPods case, with a two-inch screen, a fluffy animated avatar and a fingerprint sensor. Its own 5G, so it works without your phone. Designed by Alan Dye, until recently Apple’s head of user-interface design.

Meta is targeting the holidays to put it in people’s hands.

Beside it: VR glasses at around 100 grams that make Apple’s Vision Pro look like a brick. Ray-Ban glasses with no camera at all. And behind all of it, one promise: personal superintelligence, carried on your body rather than sitting in a data centre far away.

Artificial intelligence is coming off the desk and onto the face.

Now, Zuckerberg has looked for his Big Thing before.

The metaverse? An expensive flop. Headsets? A niche. But the smart glasses, the ones we wrote about whenMeta teamed up with EssilorLuxottica, quietly turned into the first AI hardware people actually wear. So the question writes itself.

Is this Meta’s iPhone moment?

Let’s dig in.

Chart 1: Muse Charm — keychain-sized, a two-inch screen, a fluffy avatar and its own 5G

Meta’s Muse, or the Difference Between Asking and Getting

You ask ChatGPT how to book a flight.

Muse books the flight.

That is the whole difference, and it is far bigger than it sounds.

Muse launched on 8 September as Meta’s personal AI agent, not a chatbot. It has its own email address, runs in its own isolated virtual machine, and plugs into Walmart, Instacart, Expedia, Notion and GitHub. It answers mail, blocks your calendar, compares prices, fills in forms, books the restaurant and pays with a one-time virtual card. For anything sensitive it asks permission. Everything else runs in the background.

Sit with that for a second.

Not a chat window.

An assistant that works while you sleep.

«Hey Patrick, you have paid for Netflix for three months and never opened it. Shall I cancel?»

Or

«Hey Patrick, you are paying far too much for your phone and use 20% of your contract. Shall I renegotiate, and go elsewhere if they refuse?»

Yes, that is the future. A bit scary, I know. We will have to adapt.

And the small example that will land with everyone: you save a recipe on Instagram, and Muse turns it into a grocery order that shows up on Thursday.

Although, as a Swiss household, we will probably still end up at Coop or Migros… And will not use Coop at Home…

And the uncomfortable half of the same sentence: an agent that can spend your money is an agent that can be wrong with your money.

Muse looks like a Tamagotchi. Small screen, fluffy creature, hanging on a keychain and demanding attention. Yes, I know what you are thinking of: The nineties are back. Except this one books your holiday.

And it is selling. Roughly three million downloads in the United States in the first two weeks, and the number one spot on the App Store on 18 September. Look at the slope in Chart 2. At the same point after launch, ChatGPT sat at half that. Meta’s own AI app lower still. Google’s Gemini barely off the floor.

But the downloads are not the point. The pipe is the point.

And that is Meta’s biggest edge.

3.60 billion people use a Meta app every single day. Facebook, Instagram, WhatsApp, Messenger, Threads, roughly 45% of humanity. Meta does not need to win a download war. It can switch Muse on inside apps you already have on your phone.

That is the competitive advantage nobody else owns. OpenAI has to acquire every single user. Meta has to send a notification.

And that is what makes the iPhone analogy more than a headline.

An iPhone moment needs more than a clever product. It needs distribution. Meta already sits in the pockets of 3.6 billion people. If Muse works, Meta does not have to teach the world a new habit one user at a time. It can put that habit inside apps people already open every morning.

That does not make an iPhone moment inevitable.

It makes one possible.

And the money?

Potential monetisation: hardware, subscriptions at $20 and $100 a month, potentially a cut of whatever the agent buys, plus more time inside the apps, which means more advertising.

Which raises a question that Cupertino, Apple’s headquarters, will not enjoy. If the agent does the clicking, who still needs an app store?

2007: touch instead of keys.

2026: agents instead of apps.

And it is not only Cupertino. If the agent compares and books, the comparison sites, the travel portals and the search box all quietly lose a customer they never knew they were renting.

The market is already pricing that uncertainty: Booking, Expedia and their peers have fallen some 20%.

A beautiful theory. And one already finding its way into share prices.

Now let it meet the income statement.

Chart 2: Muse’s US downloads since launch — outpacing ChatGPT, Meta AI and Gemini at the same

Meta’s Numbers: The Engine Pays for the Dream

The engine is extraordinary.

Second-quarter revenue of $60.8 billion, up 28%. Of that, $60.4 billion came from the Family of Apps. Revenue is expected to compound north of 20% over the next three years.

Almost all of it is advertising, and advertising is cyclical. That is the first honest risk in this «Good Story».

Here is the second.

Reality Labs, the glasses, the headsets and now Charm, produced $431 million of revenue last quarter and lost $4.6 billion doing it. Capital expenditure guidance for 2026: $130 to $145 billion. Trailing free cash flow: $41 billion.

Read those last two numbers again.

Meta is spending roughly three times its free cash flow to build a future that does not exist yet. Net income fell to $15.8 billion from $18.3 billion even as revenue grew 28%. That is what the dream costs, and it is charged to today’s shareholder.

Part of the bill does not even appear on the balance sheet. Roughly $27 billion of data-centre financing sits in a joint venture with an outside investor rather than in Meta’s own debt. Entirely legal, entirely common, and precisely the kind of thing that belongs in the «know what you own» file.

So, what do you pay for all this?

On next year’s earnings, 23.3 times, which happens to be almost exactly its own ten-year median of 23.4 (chart 3). You are paying the average price for a company that is anything but average.

But look at the lower panel. The free-cash-flow yield is 2.14%, against a ten-year median of 3.30%. On earnings you pay the median. On cash you pay up. That gap, drawn as a picture, is the capex bill.

A wonderful business, a «Good Story» then, with some caveats. But you know our rule: a wonderful business is not automatically a wonderful investment.

So, what does Mr. Market say?

Time for the «Good Chart».

Chart 3: Meta’s next-twelve-month P/E against its own ten-year median — and the free-cash-flow yield underneath it

Meta’s «Good Chart»: A Year of Going Nowhere

Here is the strangest number in this whole piece.

In the latest quarter, revenue grew 28%. Over the past twelve months, the share price went nowhere. Flat. A touch negative, in fact.

A business compounding at nearly thirty percent, and a stock that has stood still for a year.

That is not weakness. That is a base.

Every dollar of earnings the company added went into working off the multiple rather than lifting the price. The stock did the boring, necessary work: it consolidated while the fundamentals caught up with the valuation.

Bases are where new leaders are quietly assembled. You know our line by now.

The longer the base, the higher the space.

And now the tape is speaking.

Two years of coiling inside a narrowing range. Then, last week, the price broke clean out of it, at least through the first resistance. Up strongly in five sessions, on the heaviest buying volume in years (Chart 4). That looks much more like institutional buying than retail drifting in.

And here is what often follows. A move out of a long base on volume that dwarfs the average can take more than a single week to play out. The textbook sequence is breakout, a pause to digest it, then continuation. Nothing on a chart is certain. History gives us a pattern, not a promise.

What makes this breakout more interesting than most is the catalyst standing behind it. Charts move on stories, and this story is a product pointed at hundreds of millions of lives at once. Meta says it will ship in time for the holidays.

Which brings us back to the man on the stage.

Zuckerberg has been wrong in public before, expensively and repeatedly. The metaverse burned tens of billions and delivered avatars without legs. He has been mocked for every one of these bets.

He is also one of the few large-cap chief executives who keeps making them. And he told us why, in the line at the top of this page.

And that is precisely what Cupertino has not managed since 2007. Apple has spent a decade hunting for another iPhone: something that big, that profitable, that universal. It has not found it. Meta has just put its candidate on a stage.

Is Charm the iPhone or the Newton?

We do not know. Nobody does, barely a week in.

It may become another expensive Meta experiment. A strange little keychain people stop talking about by Christmas. Shareholders may spend tens of billions finding that out.

But there is another possibility.

Meta has 3.6 billion people already using its products every day. It has an advertising machine capable of paying for the experiment. And it has a founder willing to spend before the answer is obvious.

In January 2007, nobody sitting in that audience knew they had just watched an iPhone moment.

Nineteen years later, Zuckerberg walked on stage holding something much smaller.

Maybe it's just a keychain and will be dismissed as a cool gadget.

Or maybe that's exactly how an “iPhone moment” begins.

Chart 4: Meta weekly — two years of coiling inside a narrowing range, then a breakout on high-volume buying