AI Arms Race Reaches 'Point Of Inflection' As Tech Giants Pour Billions Into Data Centers

An intensifying competition to be the leader in artificial intelligence has the world’s largest tech companies ramping up data center development.

Since the emergence of ChatGPT late last year, tech giants Microsoft, Google and Amazon have all made hard pivots toward artificial intelligence, pouring resources into incorporating generative AI technology into products and services across their different business lines. But earnings reports released last week provided the first real insight into the scale of the investments these cloud providers are making in data centers and other infrastructure needed to support these technologies.

It may be the early days of the AI economy, but these companies — by far the largest users of data centers already — are ramping up spending on the facilities needed to support both their own AI products and AI cloud infrastructure for customers. They are doing so even as they make significant spending cuts elsewhere in the face of economic headwinds, positioning themselves to compete for a share of an anticipated AI demand wave they believe will fundamentally transform their businesses — and the data center landscape in the process. 

“I’ve compared it to the successful transition we made from desktop to mobile computing over a decade ago,” said Google CEO Sundar Pichai, speaking on the company's first-quarter earnings call Tuesday. “This has been an important moment, as pretty much every organization is thinking about how to use AI to drive transformation ... across the board, from startups to large companies, they are engaging with us. I view it as a point of inflection.”

This may indeed be an inflection point for AI, but it was rival Microsoft’s success over the past quarter that presented the strongest evidence that growing customer demand for these products is more than hype. 

Microsoft — whose partnership with ChatGPT-maker OpenAI has created at least the perception that it is the clubhouse leader in the early days of the AI boom — has seen immediate returns on its AI investments. The company's Azure cloud service increased its market share by a full percentage point, a gain that Microsoft's leadership largely attributes to the incorporation of OpenAI’s products. Meanwhile, the company’s Bing search engine — little more than a punchline since its launch in 2009 — saw user numbers skyrocket and app downloads grow 400% after adding ChatGPT functionality.

Microsoft’s competitors have rushed to respond over the past quarter, with Google launching a competing AI chatbot called Bard and adding AI features to both its search and cloud services, and with Amazon touting AI integrations across its various business lines.

But while chatbot search engines and other consumer-oriented AI products may signal burgeoning AI demand, it is the cloud where these companies see the biggest potential for growth through providing the infrastructure to support the burgeoning AI economy.

Developing and operating generative AI products requires enormous computing power, with more powerful chips and fundamentally different supporting infrastructure than what traditional cloud offering support. Amazon Web Services, Microsoft and Google are in a race to provide these specialized cloud offerings.

This means building new data centers capable of supporting these high-performance computing needs, and all three cloud giants say they are ramping up spending on data center development in preparation for growing demand. 

“We will continue to invest in our cloud infrastructure, particularly AI-related spend as we scale with the growing demand, driven by customer transformation, and we expect the resulting revenue to grow over time,” said Microsoft Chief Financial Officer Amy Hood, speaking on its Tuesday earnings call.

“We have continually focused on pivoting our resources aggressively to the future as we execute at a high level in the moment to deliver value to our customers," Hood added. "We are committed to leading the AI platform wave and making the investments to support it.”

Microsoft’s capital expenditures rose significantly last quarter to $7.8B, up more than $1B year-over-year in an increase Hood says is driven by the build-out of AI-focused data centers.

Although rival Amazon’s overall capex is trending lower, CEO Andy Jassy has indicated spending on AWS infrastructure has increased to meet demand for generative AI, with capital cutbacks in logistics and other areas of the business being redirected to AWS. 

Google is also scaling up its data center portfolio to grow it’s AI computing capacity, according to CFO Ruth Porat, with the pace of data center spending accelerating over the coming year.

“We’re expecting a step-up in the second quarter, and that will continue to increase throughout the year … AI is a key component — it underlies everything that we do,” Porat said Tuesday. “We’re continuing to invest in support of AI … the increase in capex for the full year 2023 reflects the sizable increase in technical infrastructure investment.”

This accelerated spending on data centers is happening even as the tech giants slash jobs and budgets across other segments of their business, and as the cloud industry’s growth has slowed. Revenue growth for the cloud industry as a whole — for which AWS, Microsoft and Google comprise 63% of market share — was 19% last quarter, down significantly from 26% a year ago, according to Synergy Research Group.

This continues a trend that began last quarter, as economic headwinds tempered the growth of corporate IT budgets. After three years of scaling up spending on cloud services to meet pandemic-driven digital transformations needs, companies are now focusing on getting more bang for their cloud services buck and optimizing their IT spending. 

The slowdown is expected to continue in the months ahead, with Amazon leadership indicating on its Thursday earnings call that AWS’ revenue growth this month is down around 5% from first-quarter performance.

“In AWS, what we’re seeing is enterprises continuing to be cautious in their spending in this uncertain time — customers are looking for ways to save money however they can right now,” Jassy said Thursday. “One of the great attributes of the cloud is that you can scale seamlessly up or down as demand dictates, which is not the case with on-premise infrastructure.”

Yet sluggish revenue growth isn't deterring cloud providers from investing in AI-focused capital investments, and the resulting surge in hyperscale development could reshape the data center landscape. AI doesn’t just mean more data centers; it means greater power requirements and fundamental changes to how projects are designed to support the more powerful chips and network equipment this technology requires. 

The coming months may mark the start of a significant division within the data center sector between new assets capable of supporting AI workloads and older assets that cannot, industry insiders say.

Experts also expect the major cloud providers, along with social media giant Meta, to become increasingly aggressive in gobbling up large swaths of land with access to inexpensive power, resources that will become increasingly critical as hyperscalers try to hold down the cost of energy-intensive, high-performance computing for AI as it becomes central to their business model. 

“The move to generative AI is going to completely unhinge the cost basis for most of these businesses,” said George Slessman, CEO of data center firm DCX, speaking in March at Bisnow’s DICE Southwest event Phoenix. “What's coming in the underlying infrastructure is as important to this generation of infrastructure as the internet was to data centers … this is not a moment to be taken lightly.”