Hyperscale Firms See "Clear" AI Revenue Gains

Recent third-quarter financial reports by Meta, Alphabet, Microsoft and Amazon should quell some of the concern about AI contributions to revenue, at least for the hyperscalers making the biggest investments.

Google CEO Sundar Pichai said its investment in AI is paying off in two ways: fueling search engagement and spurring cloud computing revenues. That’s a good example of one way AI will be monetized in many instances: indirectly, as existing products are improved.

Google services revenue--which includes search--grew 13 percent in the third quarter of 2024. Google Cloud, whose revenues grew 35 percent in the third quarter of 2024. 

Separately, Microsoft likewise reported robust AI-linked revenue gains. “Our AI business is on track to surpass an annual revenue run rate of $10 billion next quarter, which will make it the fastest business in our history to reach this milestone,” says Satya Nadella, Microsoft CEO.

Andy Jassy, Amazon CEO, said “our AI business is a multi-billion dollar business that's growing triple-digit percentages year-over-year and is growing three times faster at its stage of evolution than AWS did itself.” 

Nadella points to Azure cloud computing as a service revenue as one component of that growth. But silicon (accelerators, for example); Azure AI; developer tools (GitHub); CoPilot and LinkedIn as AI-linked products with revenue contributions, if perhaps indirect. 

One thing Microsoft does not appear to be doing is renting graphics processor unit compute cycles, as some other cloud computing firms are doing. 

“We're not actually selling raw GPUs for other people to train,” says Nadella. “In fact, that's sort of a business we turn away because we have so much demand on inference.”

“We kind of really are not even participating in most of that (renting GPU compute cycles) because we are literally going to the real demand, which is in the enterprise space or our own products like GitHub Copilot or M365 Copilot,” Nadella says. 

In fact, Microsoft seems to be going the other way, leasing compute cycles and GPU access  from firms such as CoreWeave. 

Google CEO Sundar Pichai said its investment in AI is paying off in two ways: fueling search engagement and spurring cloud computing revenues. That’s a good example of one way AI will be monetized in many instances: indirectly, as existing products are improved.

Also, as expected, the cost of inference has declined dramatically. “Since we first began testing AI Overviews, we have lowered machine cost per query significantly,” said Pichai. “ In 18 months, we reduced cost by more than 90 percent for these queries.”

And though an argument can be made that AI might cannibalize some significant amount of search, Google has found, since AI Overview was introduced, that “strong engagement” leads to “increasing overall search usage and user satisfaction,” Pichai noted. “People are asking longer and more complex questions and exploring a wider range of websites.”

That, in turn, fuels the advertising revenue potential. 

Google Cloud usage to support AI operations also has skyrocketed. “Gemini API calls have grown nearly 40 times  in a six-month period,” Pichai said. 

Likewise, Google Cloud has seen 80 percent growth in BigQuery ML (machine language)  operations over a six-month period, he noted.

Google AI Monetization: Ads and Cloud Computing as a Service

Google CEO Sundar Pichai said its investment in AI is paying off in two ways: fueling search engagement and spurring cloud computing revenues. That’s a good example of one way AI will be monetized in many instances: indirectly, as existing products are improved.

Google services revenue--which includes search--grew 13 percent in the third quarter of 2024. 

On the other hand, for at least some infrastructure providers, AI will drive usage of cloud computing resources as well, in this case Google Cloud, whose revenues grew 35 percent in the third quarter of 2024.  

Also, as expected, the cost of inference has declined dramatically. “Since we first began testing AI Overviews, we have lowered machine cost per query significantly,” said Pichai. “ In 18 months, we reduced cost by more than 90 percent for these queries.”

And though an argument can be made that AI might cannibalize some significant amount of search, Google has found, since AI Overview was introduced, that “strong engagement” leads to “increasing overall search usage and user satisfaction,” Pichai noted. “People are asking longer and more complex questions and exploring a wider range of websites.”

That, in turn, fuels the advertising revenue potential. 

Google Cloud usage to support AI operations also has skyrocketed. “Gemini API calls have grown nearly 40 times  in a six-month period,” Pichai said. 

Likewise, Google Cloud has seen 80 percent growth in BigQuery ML (machine language)  operations over a six-month period, he noted. 

AI capital investment levels will remain an issue for some time, given the huge leap in capex for AI infrastructure, models and inference that happened in 2023 and has continued into 2024. Google itself projects an increase in AI capex spending for 2025, as well. 

Some idea of the ramp up of investment can be seen in venture capital investments alone, and excluding investments by leading firms such as Google, Microsoft, Meta, Apple and Amazon, to support generative and other forms of AI. 

Year	Estimated VC Investment (Billions USD)
2020	0.2
2021	1.2
2022	2.7
2023	22.4
2024 (projected)	30+

Those figures do not include any sums spent by enterprises, software or hardware firms to create AI features, apps or platforms. Nor doe those amounts include investment by hyperscale app providers or  device firms to add AI features to their existing products. 

source: Our World in Data 

The big takeaway from Alphabet’s most-recent earnings call is that significant revenue attributable at least in part to AI investments has been seen.

Consumer Feedback on Smartphone AI Isn't That Helpful

It is a truism that consumers cannot envision what they never have seen, so perhaps it is not too surprising that artificial intelligence smartphone features people say they want are augmentations or enhancements to features they already use. 

source: YouGov

Consumer surveys might provide more-relevant feedback for existing  products that might be enhanced in some way using AI or other digital technologies.

Product	Description	Reason for Lack of Consumer Insight
Smart Sleep Mask	A sleep mask that uses sound and light therapy	Consumers may not fully understand the benefits of sleep tech.
AR Navigation Glasses	Glasses providing augmented reality navigation	Difficulty in envisioning practical daily use and advantages.
Wearable Stress Monitor	A device that tracks stress levels and suggests coping methods	Lack of awareness about stress management tools.
Personalized Nutrition App	An app that tailors meal plans to individual genetics	Consumers might not be aware of how genetics affect nutrition.
Modular Smart Furniture	Furniture that can be reconfigured for various needs	Novelty in furniture concepts may confuse potential users.
Biodegradable Phone Cases	Eco-friendly cases that decompose after use	Limited consumer knowledge about environmental impacts of plastics.
Virtual Reality Fitness Games	Immersive games that promote physical activity	Skepticism about fitness benefits of gaming; unfamiliarity with VR.
Smart Pet Feeder	Automatic feeder that dispenses food on schedule	Pet owners may not see the value of automation in pet care.
Language Translation Earbuds	Real-time translation via wireless earbuds	Consumers might not grasp the practical applications in daily life.
Holographic Displays	3D displays for personal and professional use	People may not fully understand how holography can enhance experiences.

But there arguably is more difficulty ascertaining consumer demand for products that did not exist in the past, even if, in some cases, digital versions of previous analog consumer goods were introduced. The iPod might have been “revolutionary” in the music player category, but people already were familiar with portable music-playing devices. 

Smartphones already existed before the Apple iPhone, but the ease of use was a major advance. E-book readers were a new appliance concept, but the advantages of immediate content downloading, similar to the advantage of music player content downloading, might have eased consumer acceptance, even if nobody really had experience using such a device in the past. 

Product	Company	Category	Description
iPod	Apple	Digital Music Player	Revolutionized how people listen to music with a portable device and iTunes.
iPhone	Apple	Smartphone	Combined a phone, iPod, and internet communicator, transforming mobile tech.
iPad	Apple	Tablet Computer	Created a new segment between smartphones and laptops for media consumption.
Apple Watch	Apple	Smartwatch	Integrated fitness tracking, notifications, and health monitoring in a wearable.
AirPods	Apple	True Wireless Earbuds	Popularized wireless earbuds, emphasizing convenience and sound quality.
Kindle	Amazon	E-Reader	Pioneered the e-reader market, changing how people consume books digitally.
Nest Learning Thermostat	Nest (now Google)	Smart Thermostat	Introduced smart home technology that learns user preferences for energy saving.
Fitbit	Fitbit	Fitness Tracker	Created a category for wearable fitness technology focused on health tracking.
Oculus Rift	Oculus (now Meta)	Virtual Reality Headset	Brought immersive VR experiences to consumers, paving the way for gaming and more.
Microsoft Surface	Microsoft	2-in-1 Laptop	Blended laptop and tablet functionality, creating a new form factor for devices.

The point is that most AI innovations are extrapolations from what we already know. In most cases that is likely useful, but also unlikely to create a platform for whole new industries with new value propositions and products. 

As a corollary, it also is virtually impossible to quantify demand for such innovations, as consumers would not be able to envision their use of the new products, even if innovators can try to describe them. 

New Product Category	Year Introduced	Difficulty of Evaluating Consumer Demand
Smartphones	2007	High
E-cigarettes	2003	High
Smart speakers	2014	Medium
Ride-sharing services	2009	High
Virtual reality headsets	2016	High
Meal kit delivery services	2012	Medium
Cryptocurrency	2009	Very High
Electric scooters (sharing)	2017	Medium
Streaming media services	2007	Medium
Wearable fitness trackers	2009	Medium

At least so far, AI features for lots of products are appearing. It remains unclear whether any of those adaptations will decisively change product demand. For example, observers speculate about the ability of Apple Intelligence to spur sales of Apple iPhones. But each new generation of iPhones creates sales volume for other reasons. 

Each new generation of iPhone typically features a mix of technological advancements, design updates, and features.

Camera Upgrades; better displays; longer battery life; updated operating systems; faster processors and handset designs (colors, bezels, materials) tend to be the focus. Apple Intelligence is such an enhancement, but so far seems focused on enhancing existing iPhone functions. 

We still do not see a breakthrough use case that redefines the category or creates a new category. And even if such a use case were to emerge, consumers would not be able to provide much useful feedback about demand.