(Reuters) - Big tech including Microsoft ( MSFT ) and Meta are stepping up spending to build out AI data centers in a rush to meet vast demand, but Wall Street is hungry for a quicker payday on the billions invested.

Microsoft ( MSFT ) and Meta both said on Wednesday their capital expenses were growing due to their AI investments. Alphabet, too, reported on Tuesday that these expenditures would remain elevated.

Amazon ( AMZN ), which is set to report results on Thursday, is likely to echo these forecasts.

The extensive capital spending could threaten fat margins at these companies, and pressure on this metric is likely to spook investors.

Big tech shares fell in after-hours trading on Wednesday, highlighting the challenges the companies face as they seek to balance ambitious AI pursuits with the need to reassure investors they are focused on short-term results.

Meta's stock fell 2.9% in after-hours trading, and Microsoft's ( MSFT ) stock price fell 3.6%, despite each topping profit and revenue expectations for the July-September period. Amazon ( AMZN ) stock also dipped.

"It's costly to run AI technology. Getting capacity is expensive," said GlobalData analyst Beatriz Valle.

"It has become a competitive race among the big tech companies to build out capacity. It's going to take time to see the returns, to see widespread adoption of the technology."

Microsoft's ( MSFT ) capital spending for a single quarter now is more than its annual expenditure used to be until fiscal 2020, according to Visible Alpha. For Meta, a quarter's worth of spending is in line with what they spent in a year until 2017.

Microsoft ( MSFT ) said capital spending rose 5.3% to $20 billion in its first fiscal quarter, and predicted increased spending on AI in the second.

But growth at its key cloud business Azure is likely to slow, it warned, blaming capacity constraints at its data centers.

"I think what investors are missing is that for every year Microsoft ( MSFT ) overinvests - like they have this year - they're creating a whole percentage point of drag on margins for the next six years," said Gil Luria, head of technology research at D.A. Davidson.

Meta, meanwhile, warned of "significant acceleration" in artificial intelligence-related infrastructure expenses next year.

BOTTLENECKS IMPEDE GROWTH

Capacity constraints are rippling through the tech industry.

Chipmakers including powerhouse Nvidia ( NVDA ) are struggling to keep up, in turn making it harder for cloud companies to build out capacity.

Advanced Micro Devices ( AMD ), which reported results earlier this week, said demand for AI chips was rising much faster than supply, limiting its ability to tap the order surge. It warned that supply of AI chips would be tight going into next year.

Despite the concerns, Meta and Microsoft ( MSFT ) said it was still very early in the AI cycle and emphasized the long-term potential of AI.

The investments are reminiscent of when Big Tech was developing cloud businesses and waiting for customers to embrace the technology.

"Building out the infrastructure is maybe not what investors want to hear in the near term, but I think the opportunities here are really big," said Meta CEO Mark Zuckerberg during Wednesday's earning call. "We're going to continue investing significantly in this."