The promise of a generative-AI gold rush has whipped the tech sector into a frenzy of anticipation, and with the exception of Microsoft, the hype has yet to translate into revenue growth for cloud providers.
Rates of cloud-computing sales growth slowed to historical lows this year as customers slashed IT budgets. Amazon Web Services, the largest cloud provider in the world, reported a 12% year-over-year sales increase for each of the past two quarters, less than half its growth rate for the same time last year. Microsoft and Google, the next two biggest American cloud providers, also saw rates of cloud-revenue growth slow this year.
Microsoft’s Azure and other cloud services saw revenue growth of 29%, down from 35% this time last year. Google this week reported a 22% year-over-year increase in cloud revenue the past quarter, down from 28% the previous quarter. The company attributed this to customer “optimization,” or looking for ways to save money on cloud services.
Wall Street has been hopeful that generative artificial intelligence will provide a much-needed revenue boost for the three big cloud providers. Still, analysts say it’s too soon to expect much of a lift. The one exception is Microsoft, which has spent billions investing in the generative-AI startup OpenAI, the maker of the chatbot ChatGPT.
“On the AI front, the near-unanimous view was that enterprise spend remains very early stage,” Karl Keirstead, UBS’s managing director, wrote in a research note to investors last week.
For its last quarter, Microsoft’s cloud-services unit reported that 3% of its sales growth came from generative AI, one percentage point more than expected. Sales are coming from services such as Microsoft 365 Copilot, a virtual AI assistant made with OpenAI’s technology.
Microsoft’s OpenAI investment gave it a head start on selling generative-AI services such as Copilot, which can easily be purchased as an upgrade to existing plans. Still, cloud customers surveyed by Keirstead’s team broadly said that they’re still figuring out what to do with the nascent technology.
“Enterprises are just trying to understand AI, such as that we’re being pulled in to do AI strategy work,” one cloud partner told UBS. “Actual spending is not material yet and even the proof-of-concepts you hear about are really just companies playing around.”
Another partner suggested that while the corporate world eagerly jumped on the generative-AI bandwagon, company leaders were either realizing they’re not prepared for it or they just didn’t understand it, making it difficult for cloud providers to close deals.
“People seem stuck between FOMO (fear of missing out) and FOJI (fear of jumping in),” the partner told UBS. “It’s caused a bit of paralysis in some cases where decisions are taking longer.”
A third partner said they’d seen some generative-AI deals close but not many. “It’s still fewer and far between,” the partner told UBS.
Wall Street analysts blame the spending delay on the economic environment, which remains persistently uncertain because of rising inflation, the war on Ukraine, and now the fighting between Israel and Hamas.
“We believe there will be a long ramp to monetization and generative AI may not be material for most companies until 2025,” Rishi Jaluria, RBC Capital Markets’ managing director, wrote in a research note this week. He described investors as “jittery” going into software earnings season.
AWS this week touted the launch of general availability for its large language model Bedrock, which gives customers access to a wide variety of AI models from third-party providers such as Anthropic and Stability AI.
Amazon CEO Andy Jassy said on the company’s earnings call Thursday that generative AI was still in its “early days.”
“Companies are still learning which models they want to use, which models they use for what purposes, and which model sizes they should use to get the latency and cost characteristics they desire,” Jassy said on the call. “In our opinion, the only certainty is that there will continue to be a high rate of change.”
Got a tip for this reporter? Contact Ellen Thomas at [email protected] or on the encrypted-messaging app Signal at +1 (646-847-9416).
Read the full article here