Capital spending by Big Tech companies in their capital projects has doubled through this year, crossing the $100 billion mark, in a mad rush to lay down the necessary infrastructure for the fueling of artificial intelligence. That is despite Wall Street’s creeping skepticism regarding payoffs from such big spending.

Each quarterly earnings report from Microsoft, Alphabet, Amazon, and Meta has indicated that spending was at $106 billion, increasing at a blistering pace, as released in its first half of 2024 results. Throwing caution to the wind, the tech titans, meanwhile, promised to continue increasing investments over the next 18 months.

“I would rather take on that capacity too early,” Meta CEO Mark Zuckerberg said, anticipating that Meta’s capital expenditures could jump to $40 billion this year.

They predict that the tech giants are expected to spend over twice as much on AI by the end of the year. Dell’Oro Group analytics is that there could be as much as $1 trillion in patchwork investment over five years, although doubt has been expressed, given customers’ reticence to spend anywhere near that on products and services related to AI.

“Spending decisions are being pressed by tech management teams,” said Jim Tierney, head of U.S. growth at AllianceBernstein. “The business models and returns are still unclear to investors as they gird themselves in a ‘trust us’ environment in light of all that spending.”

These results came against a wider Wall Street downturn, in which the Nasdaq fell into the correction territory on Friday after softer-than-expected U.S. jobs data. While these gains are being made across the board, semiconductor stocks have been particularly dicey of late, as investors digest Big Tech spending plans.

Intel, which hasn’t really taken a lot of AI infrastructure spending because of product uncompetitiveness, saw its shares drop by over 25% on Friday after a large batch of job cuts it announced.

The heads of Big Tech were undaunted, however, by the immediate share sell-offs in Google, Microsoft, and Amazon on the back of their earnings reports.

In an interview, Zuckerberg said that his company’s next big language model consumed “almost 10 times” as much computer power as its predecessor and that some of the features Meta AI has developed are more likely to make real money “over years” rather than in the near future.

“In tech, during transitions like this, the risk of under-investing in AI is much greater than over-investing,” said Google CEO Sundar Pichai.

Alphabet, Google’s parent company, said that capital spending increased to $25 billion in the first two quarters of 2024, representing a 90% increment over the same period last year. Microsoft had its own reproduction of staggering numbers by citing a rise in capital spending by 78% during the quarter to $33 billion. Amazon reported a 27% increase in property and equipment investments to $32.5 billion, which includes its extensive e-commerce and logistics network.

Alphabet, Google’s parent company, said that capital spending increased to $25 billion in the first two quarters of 2024, representing a 90% increment over the same period last year. Microsoft had its own reproduction of staggering numbers by citing a rise in capital spending by 78% during the quarter to $33 billion. Amazon reported a 27% increase in property and equipment investments to $32.5 billion, which includes its extensive e-commerce and logistics network.

In a follow-on filing, Amazon said total capital spending is expected to “meaningfully increase” in 2024, with a big chunk of that going to new cloud infrastructure. That generative AI, as a result, has become a “multibillion-dollar business” for Amazon. But as its shares ended on Friday 8.8 percent lower, with the entire market index being weighed down, the first weak spots in its consumer business appeared.

Google executives pointed to robust advertising revenues -11% higher at $64.6 billion only in the second quarter, as proof that the big spending is underpinned by the company’s core business.

Staving off an Alphabet stock drop, it seemed to appease Microsoft investors somewhat hearing CFO Amy Hood describe the company’s data centers as long-term assets to be “monetized over 15 years and beyond”.

In the worst-case scenario, she added that Microsoft could pace the installation of expensive AI hardware into data centers if demand for the AI products and services turned out to be underwhelming. Again, Microsoft said recent growth in the cloud was actually capacity-constrained, not driven by customer demand.

A large part of the investments from Big Tech are in the purchase of lands to build new data centers required for cloud computing. And billions are being spent on hardware, particularly special chip clusters that are in the main made by Nvidia, required for the training and operation of large language models, which power chatbots.

Demand for cloud services has surged as companies experiment with generative AI to automate processes and boost productivity – even though many of the trials aren’t yet fully operational. In this race, start-ups like OpenAI, Anthropic, Elon Musk’s xAI, and France’s Mistral are trying to get as many computing resources as possible to train increasingly sophisticated LLMs.

These latest stock market gyrations come in the wake of a record-setting run-up. New York’s technology-dominated Nasdaq 100 has jumped around 70 percent since early 2023, when excitement about AI began to really take hold and propelled Apple, Microsoft, Nvidia, Alphabet, and Amazon to become the five most valuable public companies in the world.