Apple folds the iPhone, Meta gives artificial intelligence a job

Key takeaways

• Meta’s rally reflects excitement about AI that acts, not proof that Muse will become a profitable product.

• Apple still follows a different model: better hardware, premium pricing and deeper ecosystem spending.

• Investors increasingly want AI spending to produce measurable behaviour, revenue or customer lock-in.

Apple unveiled a phone that folds. Meta unveiled an artificial intelligence (AI) assistant that can send emails, book travel and make purchases.

Investors reacted very differently.

On 9 September 2026, Meta closed at 653.69 USD, up 6.55%. Apple finished at 315.34 USD, down 0.28%. That is not a verdict that Meta is winning and Apple is losing. It is a sign that the technology cycle is changing.

Investors have spent years funding enormous AI investments. Increasingly, they want evidence that new technology changes what customers do, what they pay for, or how difficult they are to lose.

Three roads to the customer’s wallet

Apple’s model remains straightforward. Build desirable devices, control the software around them and charge premium prices.

Its new iPhone Duo pushes that formula further. The first foldable iPhone starts at 1,999 USD. Apple is testing whether a new form factor can encourage upgrades, protect pricing power and pull customers deeper into its ecosystem.

That ecosystem matters. Apple generated 54.25 billion USD from the iPhone and 30.74 billion USD from Services last quarter. Hardware remains the front door. Services increase the value of keeping customers inside.

Meta approaches AI differently.

Muse is designed not just to answer questions but to perform tasks across applications. If users repeatedly rely on it, Meta could monetise that activity through subscriptions, commerce and stronger engagement across its existing platforms.

That last point matters because advertising still generates almost all Meta’s revenue. Even small improvements in engagement can become valuable when applied across billions of users.

Amazon and Alphabet offer a third model: connect AI directly to commercial intent.

Amazon can use AI to help customers compare products and complete purchases. Google can insert AI into search, shopping and advertising. The distance between an AI query and actual revenue can therefore be much shorter.

From AI capex to AI cash flow

This matters because AI spending is becoming enormous.

Meta expects capital expenditure of 130 billion USD to 145 billion USD in 2026. In the second quarter alone, capital expenditure reached 31.08 billion USD, while free cash flow fell to just 784 million USD.

Muse therefore matters beyond the product itself. It offers one possible answer to the question investors increasingly ask: what does all this infrastructure spending eventually produce?

The 6.55% share-price jump does not prove the economics work. It shows how valuable a credible monetisation path can become when spending expectations are already high.

Risks

Apple still needs to prove consumers want an expensive foldable iPhone in meaningful numbers.

Meta faces a different challenge. The more useful an AI agent becomes, the more responsibility users give it. Reliability, privacy and payment security therefore become more important.

There is also a simpler risk across Big Tech: people may use AI heavily without paying much for it. Usage is not the same as attractive economics.

Investor playbook

• Follow behaviour, not demonstrations. Watch repeat usage, transactions and conversion rather than headline AI capabilities.
• Compare revenue with spending. AI growth matters less if infrastructure costs rise just as quickly.
• Watch Apple’s upgrade cycle. Strong foldable demand could show hardware innovation still supports pricing power.
• Track the route to money. Advertising, subscriptions and transactions provide clearer signals than vague promises of future AI value.

From spectacle to habit

Yesterday looked like a contest between a folding screen and an AI agent. It is better understood as a contest between monetisation models.

Apple makes money when better hardware encourages customers to upgrade and remain inside its ecosystem. Meta needs AI to deepen engagement and eventually create new revenue streams. Amazon and Google sit closer to the transaction itself.

None guarantees attractive returns.

For investors, the question is becoming less about which demonstration looks cleverest and more about which technology changes behaviour repeatedly, produces cash and does so without costs rising just as fast. 

The next technology cycle may still reward spectacle. But usefulness is becoming the receipt investors increasingly want to see.