• by Laurie Sullivan , Staff Writer @lauriesullivan, March 28, 2025

Alphabet, Amazon, and Microsoft are projected to spend more than $250 billion on capital expenditures needed to support generative AI (GAI) services in 2025 on public cloud servers.

Morningstar analysts believe the trend is similar to what the industry witnessed with the birth of the internet.

So many people began testing OpenAI's newly released 4o Image Generation feature in ChatGPT that the company temporarily set limits how often they could use it. OpenAI, which is backed by Microsoft, said the demand has been so high that it is straining its computer resources.

OpenAI CEO Sam Altman posted on X that it is "super fun” to see people enjoy the feature, but joked that “GPUs are melting.”

The need for CPU processing power and cloud space has left some investors wondering whether these company’s investments in GAI will produce strong economic returns for these tech companies.

Most of us are watching the next phases of advertising take shape, while the lucky ones are making it happen. 

Dan Romanoff, senior equity analyst at Morningstar, believes even if GAI demand doesn’t materialize as expected the data centers these firms are building have a value to create other uses as all types of workloads continue to shift to the cloud. In the case of advertising, that could be ad servers and other media networks and platforms.

Morningstar, which provides independent investment research, released a report on Thursday that estimates generative AI (GAI) will produce high economic returns similar to what the tech industry experienced in public cloud access.

The GAI market is expected to have a compound annual growth rate of 75% during the next five years.

The report points to several factors for long-term success that will rely on investments to train large language models (LLMs) and investments in infrastructure that can run AI inference workloads.

Investments in GAI stand to catalyze an increase in cloud spending, with early signs of increasing demands in GAI workloads. While capex could accelerate in 2025, Morningstar analysts see a slowing in 2026 and beyond for all three public cloud players.

Morningstar asserts that GAI demand is “not a fad.” The report also asserts that “demand for this technology is real and stands to transform the modern enterprise.”

“We have seen prices for generative AI tools dramatically decrease over the past two years,” according to Morningstar. “As prices continue to go down, demand for generative AI tools has skyrocketed.”

Amazon, Google and Microsoft have a “distribution advantage, allowing them not only to directly monetize generative AI via their cloud businesses, but also indirectly monetize the technology by improving their core businesses such as advertising, productivity tools, and retail.”

These companies also are building out their own custom AI chips, which in time can provide them with a meaningful cost advantage over firms running their training/inference workloads on off-the-shelf GPUs.