Why Apple Has the Edge in Big Tech’s Race to Deliver AI Profits
Rivals have invested billions with little to show for it
When Apple Inc. (AAPL) unveiled its artificial intelligence plans at its Worldwide Developer’s Conference, (WWDC) in June, it went from AI laggard to AI leader overnight.
Apple has leapfrogged everyone else by concentrating less on the generative AI bandwagon in favor of a less splashy approach. Apple demonstrated an array of AI features – which it calls Apple Intelligence – that are integrated into the operating system and built into apps people use every day.
Less ambitious and not nearly as sexy as what we’ve seen from the likes of ChatGPT-maker OpenAI, but more practical and a better fit for the vertically integrated Apple ecosystem.
It’s also a far better business strategy.
As it turns out, building out generative AI capabilities like ChatGPT is monstrously expensive. Companies like Microsoft Corp. (MSFT), Alphabet Inc. (GOOGL), and Meta Platforms, Inc. (META) have spent tens of billions of dollars on the research to build it and the hardware to run it – particularly specialized chips from Nvidia Corp. (NVDA).
This has been wildly profitable for Nvidia, (perhaps the greatest pick-and-shovel play of all time) but the Big Tech companies doing the buying have yet to see much of a payoff themselves.
Until recently, Wall Street was willing to wait for the profits it knew were coming. From Nov. 1, 2022 to July 1 of this year, both Alphabet and Microsoft stock rose 92%, while Meta rocketed a whopping 444%.
But now investors are showing signs of impatience, reflected in headlines like these:
Has the AI Bubble Burst? Wall Street Wonders if Artificial Intelligence Will Ever Make Money
Investors Are Suddenly Getting Very Concerned That AI Isn't Making Any Serious Money
The End of Investors' Generative AI Honeymoon
But Apple’s plan to put integration first avoids this pitfall. Apple’s approach is structured to generate significant AI-based profits in less time and with far less up-front spending.
Today I’m going to dig into how the other Big Tech companies got to this point – and the details that make Apple’s strategy a masterstroke.
Trying to Make the AI Math Add Up
There’s no question AI will generate a mountain of revenue, which is why Big Tech has been willing to pour so many billions into it.
According to a June report from Grand View Research, the global AI market is expected to grow from $196.63 billion last year to $1.81 trillion by 2030. That represents a jaw-dropping CAGR (compound annual growth rate) of 36.6%.
But so far at least, the heavy spending is canceling out most of the returns.
Take Microsoft, which reported earnings July 30.
The company said its capital expenditures rose a startling 77.6% to $19 billion, with the lion’s share going to cloud and AI-related spending. For its fiscal 2024 (the June quarter is Microsoft’s Q4), capital spending was $55.7 billion.
As a percentage of revenue, capex rose from 19.04% to 29.35%.
Meanwhile, revenue growth in Microsoft’s cloud business, where it expects to realize the benefits of its AI investing, was slightly lower than expected (29% vs 30.6%). Worse still, Microsoft’s guidance projected further slowing in the current quarter.
That triggered a 7% drop in MSFT stock.
The root of the problem is that generative AI has multiple costs. The Nvidia chips are expensive. Training the AI takes time and is resource-heavy (while creating no revenue). And keeping AI services up and running requires a vast amount of electricity.
The primary generative AI app, ChatGPT, uses about 500,000 kilowatt-hours of electricity per day to process about 200 million requests. That’s the equivalent to the daily electricity use of about 17,240 U.S. households. Even that crazy figure doesn’t include the power the Big Techs are using to train the AI models.
The monthly cost just of fulfilling ChatGPT’s request is more than $4 million per month. And as AI use grows, so will the bill.
Wells Fargo estimates that AI’s total annual power requirements will explode by 550% to 52 terawatt-hours in just two years – enough juice to power more than 3.5 million U.S. homes for a year. That’s about $6 billion worth of electricity, and it’s an expense that will keep going up.
Revenues can’t begin to cover these costs.
Just look at OpenAI, a pure play AI company. According to The Information, the dominant provider of generative AI could actually run out cash within the next 12 months.
Here’s the math: OpenAI’s sales for 2024 will be between $3.5 billion and $4.5 billion. But training the models and processing the requests will cost OpenAI $7 billion, with staffing adding another $1.5 billion. That will leave a shortfall of between $4 billion and $5 billion.
This is why Big Tech is struggling to show a profit from AI. It may take years before the revenue outpaces the expenses.
That brings us back to Apple…
How Apple Jumped to the Front in the AI Sweepstakes
Apple Intelligence, Apple’s marketing term for its version of AI, is far less focused on costly cloud-based generative AI. Instead, the center of gravity for Apple Intelligence is local – it happens right on the company’s devices.
Apple has been preparing for this moment for years by building a “neural engine” into Apple Silicon – the chips that power most of the company’s hardware. The neural engine is designed specifically for machine learning and AI tasks – without the need to access expensive resources in the cloud.
Just look at Apple’s capex spending, which was steady at just under $11 billion per year in both FY 2022 and FY 2023. For the current fiscal year, it’s on track to decline to well under $10 billion — a stark contrast to the soaring capex its rivals are reporting.
While Apple has been spending more on R&D — the main focus of Apple’s AI investment — it’s still spending less than other Big Tech companies.
Apple’s R&D spending grew by 36.55% between 2021 and 2023. But R&D spending at both Meta (up 47.97%) and Alphabet (up 41.25%) grew faster. Only Microsoft’s increase was lower at 31.27%, but it doesn’t include Redmond’s multi-year $13 billion investment in OpenAI.
But spending less is only part of the equation. Apple also has better ways to monetize AI more quickly.
In the short term, Apple will use its new AI features as a marketing tool to encourage customers to buy new iPhones, iPads, and Macs. That will deliver an immediate boost to profits over the coming quarters.
That’s why in the six weeks following the WWDC announcements, Wall Street analysts jacked up their AAPL price targets by an average of about $45 a share.
"Apple Intelligence is a clear catalyst for a multiyear product upgrade cycle," Morgan Stanley analyst Erik Woodring said in a client note following the Apple event.
Woodring estimated Apple Intelligence will help boost iPhone sales by 6% over the next two years while pushing the iPhone’s average selling price (ASP) up 5%. He raised his price target on Apple to $273. The highest current target is Loop Capital’s $300.
Though the impact will be smaller, Mac and iPad sales will also benefit.
Other PC makers have been hoping that the arrival of AI PCs will trigger an upgrade cycle and rejuvenate the PC market.
But since all Macs are now equipped with an AI-capable M-series chip, Apple has a huge head start. Research firm Canalys says Apple accounted for 60% of the AI-capable PCs shipped worldwide in the second quarter, quite a feat for company that typically has a market share in the neighborhood of 10%.
In the longer run, Apple has other ways to profit from AI.
Two Magic Words: Recurring Revenue
While working on the Apple Intelligence features, Apple realized it still needed to offer generative AI in some form.
So it made a deal with OpenAI to offer ChatGPT as an option on Apple devices. The deal cost Apple nothing; OpenAI was happy to accept access to Apple’s 2.2 billion active devices as compensation. Apple hinted that more such deals with generative AI vendors are in the pipeline.
It means Apple can offer easy access to generative AI while passing on the costs to others.
But there’s also money to be made here.
ChatGPT offers a limited free version as well as a more robust paid version. The details of the Apple-OpenAI deal were not disclosed, but it’s a good bet that any Apple users who convert to paid ChatGPT customers will earn Apple a commission of at least 15%. That’s pure gravy for Apple.
Eventually Apple is expected to develop a “pro” version of Apple Intelligence that will join its other subscription services, such as Apple Music, Apple TV+, and Apple Arcade.
For an idea of the potential here, OpenAI’s paid version is $20 per month. That’s probably what Apple would charge for a standalone AI service. Apple Intelligence also would likely join the company’s top-tier services bundle – along with a bump up in price from the current $38 per month.
Apple has focused on boosting services revenue for years, and for good reason. Gross margins on services are about 73% – far higher than the 35% margins Apple gets from hardware sales. Apple Intelligence should ensure double-digit growth in services revenue for years.
Here again Apple has a built-in advantage over the other Big Techs. Apple already has a base of about three times as many subscribers as Microsoft and about 10 times as many as Alphabet.
Does This Mean AAPL Is a Buy?
There’s no doubt that Apple Intelligence will boost AAPL stock. But those gains will be spread out over several years, and some of it is already priced in.
Since the June 10 announcement of Apple Intelligence, shares are up about 15%. And the P/E ratio is a lofty 34, well above the five-year average of 21.
If AAPL reaches Woodring’s $273 target, which I believe likely, that’s a modest 21% gain. I think $300 is a reasonable target 12-18 months out, which is a solid 33% gain.
But this will be a multi-year catalyst for Apple stock. And it doesn’t factor in any other Apple innovations or products we may see over the next few years, such as a cheaper version of the Vision Pro or an Apple Watch that can measure blood glucose levels without penetrating the skin.
Never underestimate Apple.