The next phase of the AI boom will belong to companies that can turn innovation into profits.
Listen to the audio version of this article (generated by AI).
Hello, Reader.
Wall Street loves a big number, and Microsoft Corp. (MSFT) delivered several this week.
First, on Wednesday, the company reported its fourth-quarter earnings results. Earnings reached $4.81 per share, up from $3.65 per share a year earlier. Revenue came in at $90 billion, beating expectations of $87.6 billion, while Azure cloud revenue surged 43%.
Microsoft also plans to spend $175 billion on AI infrastructure this year.
But one number stood out above the rest: Microsoft’s Remaining Performance Obligations (RPO) backlog rose 84% to $678 billion.
RPO is the official accounting term for a specific type of backlog – the contracts Microsoft has signed but hasn’t fulfilled yet.
But does a bigger backlog actually prove AI is profitable?
The market certainly thought so. Microsoft shares jumped 15% on Thursday following its blowout report. Semiconductor stocks also rebounded strongly on the numbers. The Philadelphia Semiconductor Index – which tracks 30 leading semiconductor firms – jumped about 6.7%, ending a five-day losing streak.
But there’s an important distinction being overlooked: Future revenue is not the same thing as future profits.
I think investors are getting too excited about Microsoft’s future AI revenue – especially before knowing whether it will actually turn into profitable cash.
So, in today’s Smart Money, let’s take a look at why Wall Street may be celebrating AI profits before they actually arrive.
Then, I’ll share one company that is already turning AI investments into real business results.
A Bigger AI Backlog Doesn’t Mean Bigger Profits
Microsoft’s RPO is significant, but it isn’t cash in the bank.
RPO is like a restaurant with booked with dozen’s of $40-per head banquets. That’s great because demand is strong. But food still has to be bought, staff has to be paid, and meals have to be served before those reservations turn into profits.
Revenues are fixed, but costs aren’t.
When Microsoft signs a long-term contract with a customer to provide cloud or AI services over a period of time, the amount of revenue it will receive is largely fixed. But the cost of delivering those AI services is not.
If expenses such as AI chips, electricity, data centers, or labor continue to rise, Microsoft could end up making less profit than investors expect. In other words, a growing RPO shows that Microsoft has secured future business, but it doesn’t reveal how profitable that business will be.
A signed contract is only valuable if the customer can ultimately pay. A company like OpenAI, which accounts for about one-third of Microsoft’s RPO, might not be “good for it” down the road.
Investors are acting as if this future revenue is already earned, but it isn’t.
And Microsoft’s financials suggest the cost of delivering AI is already increasing.
- Cloud gross margin fell from around 67% to 65%. That means Microsoft is making slightly less profit on every dollar of cloud revenue.
- Overall company gross margin also declined. So, profitability is slipping across the business, despite strong revenue growth.
- Capital spending (CapEx) keeps rising. Microsoft spent $41 billion this quarter on infrastructure and expects to spend even more next year. This suggests that the cost of building and operating Microsoft’s AI business is rising quickly.
Microsoft’s surge this week shows that the market is treating bigger RPO like bigger profits.
But that isn’t necessarily true.
The company clearly has a large backlog of AI business. And a growing backlog tells us that customers want AI. But what it doesn’t tell us is if those AI contracts will generate strong profits.
I think investors are celebrating a victory before it has been achieved… just like they’ve been doing with lots of other high-profile AI names.
The bigger takeaway here is that the winners of the next phase of the AI boom won’t just be the companies spending the most, but the ones that can prove their AI investments are generating returns.
With that in mind, here’s a better “match” instead…
Wall Street Swiped Left. AI May Swipe Right.
Match Group Inc. (MTCH) is in the business of making online love connections, but the company hasn’t been able to make a love connection with investors for several years. Most of them have been “swiping left” – i.e., saying “no thanks” – on the shares of this online dating leader.
But the company’s new management is engineering an AI-powered overhaul that could produce a new era of robust earnings growth.
On the consumer side, Match is deploying AI to reshape its most important brands, Tinder and Hinge. At Hinge, for example, a new AI-powered recommendation algorithm launched in March 2025 has increased matches and contact exchanges by 15%. This meaningful improvement translates into more real-world dates and higher rates of payer conversion.
On the Tinder app, which is facing sharp user declines, Match is introducing an even broader suite of AI tools. This AI-enabled upgrade is prioritizing deeper compatibility and better user outcomes, rather than optimizing for superficial swiping or other short-term engagement metrics.
By embedding AI at multiple touchpoints – from onboarding and match recommendations to internal product development – Match is building a unified ecosystem where data, technology, and human creativity reinforce each other.
This dual focus on consumer experience and operational efficiency is not simply an experiment; it is the strategic foundation for the company’s revitalization plan.
While Microsoft spends billions to build the infrastructure needed to power AI, Match is using the technology to make its existing apps more effective – and helping turn more users into paying customers.
Match is set to release its second-quarter 2026 earnings next week. And it currently offers a unique and compelling opportunity: It is successfully using AI to reinvent its business and drive future growth.
To learn about more companies turning AI investments into real business growth, click here.
Regards,
Eric Fry
