Meta Rises On The Back Of Connect

Meta’s AI Moment: Can Muse Justify the Bill?

There are moments in the market when investors stop asking whether a company has spent too much and start asking whether it has got there first. Meta appears to be in one of those moments now. Shares have climbed more than 30 per cent in September and are on course for their strongest month in more than 13 years, driven largely by enthusiasm around Muse, the company’s new personal AI agent, and the broader message delivered at Connect 2026.

That matters because only a few weeks ago the conversation around Meta was far less forgiving. When the company reported second-quarter results in July, revenue rose 28 per cent to $60.8bn, yet free cash flow collapsed to just $784mn from $8.55bn a year earlier as spending on AI infrastructure intensified. Meta also narrowed its 2026 capital expenditure guidance to between $130bn and $145bn, an extraordinary figure even for a business of its size.

The gap between those two narratives is the real story. On one side is a market that has decided Meta may finally have found the consumer AI product investors have been waiting for. On the other is the awkward arithmetic of whether a potentially successful AI assistant can ever generate returns commensurate with the bill being run up in the background.

From metaverse to Muse

What Connect 2026 seemed to confirm was that Meta has become much clearer about the story it wants investors to hear. For several years, the company’s technology ambition was framed around the metaverse, a concept that was grand in scale but often vague in commercial timing. Muse changes the emphasis. It is more immediate, more consumer-friendly and far easier for investors to understand.

At Connect, Mark Zuckerberg positioned Muse at the centre of Meta’s next phase. The company said Muse is coming to its AI glasses, while also expanding the retail and payments network around the assistant. Walmart, Best Buy, Gap, Sephora, Ulta, Wayfair and others are being added, while PayPal and Shop Pay are part of the payments infrastructure. Expedia and Instacart are also being added, broadening the practical use cases from shopping to travel and everyday errands.

That shift is important because it turns Meta’s AI effort into something closer to a platform strategy. An assistant that can search, recommend and transact across a widening ecosystem is more valuable than one that simply answers questions. Investors have understood that point quickly, which helps explain why the stock rallied even before the keynote itself.

Hardware as distribution

The other reason Connect landed well is that Meta did not present Muse as a standalone app. It presented it as the intelligence layer across a family of devices. The new line-up included Ray-Ban Meta Gen 3 glasses starting at $449, camera-free audio glasses priced from $349, a new VR headset expected in spring 2027, and Muse Charm, a small keychain-style device designed to give users access to Muse without reaching for a phone.

That matters because hardware gives Meta a route to distribution that many AI rivals simply do not have. OpenAI may dominate public discussion, but it does not yet have a mass-market consumer hardware channel. Apple has the hardware but has looked less decisive in consumer AI. Meta is trying to do something more integrated by tying software, devices and commerce together in one ecosystem.

There is also a more subtle point here. Smart glasses and wearable AI devices create repeated, habitual use. If Muse becomes part of how people shop, ask questions, navigate and communicate, then Meta has a better chance of building something stickier than a novelty chatbot. The market is effectively betting that Muse can become a daily utility rather than a short-lived launch sensation.

The cost problem has not gone away

None of that removes the basic financial issue. Meta’s spending plans remain vast. The company’s own guidance puts 2026 capital expenditure at between $130bn and $145bn, while second-quarter operating cash flow was $31.86bn and free cash flow slumped to $784mn. That collapse in free cash flow is not a side note. It is the clearest measure of what the AI build-out is costing shareholders right now.

This is where the bull case begins to look more complicated. According to market reporting, Morgan Stanley estimates Muse could generate around $1.3bn in annual revenue by 2028. That is clearly not trivial. Any new product capable of producing revenue in the low billions would be meaningful for most technology companies. But Meta is not most technology companies, and $1.3bn looks modest set against capital expenditure running well above $100bn a year.

The chart below makes that imbalance visible. It does not prove the investment is wrong, but it does show why investors will eventually need more than excitement around a keynote or strong app-store momentum.

Why investors are still buying it

The reason the market has been willing to overlook the bill, at least for now, is that Meta has a rare combination of advantages. Its advertising business remains immensely powerful, giving it the cash generation to fund experiments that would break smaller rivals. It has global consumer reach across Facebook, Instagram, WhatsApp and Threads. And unlike many AI challengers, it already owns the social graphs, engagement data and product surfaces where an assistant can be integrated at scale.

In other words, investors are not valuing Muse as a single product. They are valuing the possibility that Muse becomes the connective tissue across Meta’s entire consumer ecosystem. If that happens, the economics could improve far beyond direct app revenue, particularly if AI improves shopping conversion, advertising relevance, retention and hardware sales.

That is the optimistic case, and it is not irrational. But it does rely on several assumptions holding true at once. Muse has to maintain consumer momentum. Retail partnerships have to translate into actual transaction volume. Wearable devices have to move beyond niche adoption. And the company must show that all this infrastructure spending produces operating leverage rather than simply a more expensive version of growth.

The next test

For now, Meta has done the hardest short-term thing, which is to persuade investors that the spending has a story attached to it. Connect 2026 gave the market a product narrative, a hardware roadmap and a consumer use case that felt far more tangible than the old metaverse pitch. That is why the shares have responded so strongly.

The more difficult phase comes next. Earnings will have to demonstrate whether Muse is becoming commercially useful rather than merely culturally fashionable. Investors have bought the narrative. What they will soon want is evidence that the economics are starting to move in the same direction.

That is the real question hanging over Meta after its best month in 13 years. Muse may well be the product that finally gives the company’s AI spending a convincing public face. The issue is whether it can also become the product that justifies the bill.