A generalised euphoria over generative artificial intelligence has gripped Wall Street. But the technology had little to do with the strong business performance reported by most of the big US tech companies in recent days.
Understanding where the technology is starting to yield real business results — and where it isn’t — will be key to distinguishing the AI winners from the AI losers in the coming months and years.
Consider, for example, the business rebound that the biggest cloud computing platforms have experienced this year. Last week, reaccelerating growth in the cloud computing divisions at Microsoft and Google fed hopes that AI was starting to make a noticeable impact. This week, Amazon Web Services, the cloud market leader, has added to the upbeat mood.
The recent results provide little insight, though, into how much this growth rebound reflects a pick-up in spending on generative AI, how sustainable any such spending will prove to be, and how expensive it will be to deliver the new AI services.
The last two years brought a drastic fall-off in cloud growth. Many customers who had seen their cloud bills soar during the pandemic put a brake on new spending as they tried to work out how to get more bang for the buck.
This pause, referred to euphemistically by the tech companies as a period of “optimisation”, laid waste to one of the industry’s biggest drivers of expansion. Revenue growth at AWS tumbled from 40 per cent at the end of 2021 to a relative trough of 12 per cent 18 months later.
That it has now rebounded to 17 per cent in the latest quarter is a sign that the indigestion caused by the earlier binge of cloud spending is largely a thing of the past. According to Andy Jassy, Amazon’s chief executive (and former head of the cloud division), this is a return to the status quo ante, when the move to the cloud was fuelled by a desire to drive down IT costs. With only 15 per cent of corporate IT workload in the cloud, he argues this trend has a long way to run.
AI is not the main force here — though, at the margin, it is certainly becoming a factor. Most clearly, Microsoft’s annualised revenue from generative AI is now generally put at about $4bn, while Jassy also said it has become a “multibillion-dollar” business for AWS.
It is unclear how quickly these AI revenues will grow, or how big the market will be. There has been a stampede by customers to train new AI models and to try out the new services these make possible. But until this period of mass experimentation passes, it is hard to predict how much value the new technology will create — or how much customers will be willing to pay for it.
While the timing of the pay-off is uncertain, the costs are very real. Alphabet, Amazon and Microsoft are on course for combined capital spending of more than $150bn in their current financial years, more than $40bn above what they spent the year before. These are massive downpayments on the promise of a coming tech boom.
Cutting depreciation charges by extending the expected useful lives of all this new data centre gear has taken some of the edge off this vast build-up in investment at all three companies. Alphabet, for instance, boosted its operating profits by nearly $4bn last year after it increased the expected life of its servers and networking gear to six years, spreading the cost of buying new equipment over a longer period.
Another factor offsetting some of the pain is the cloud companies’ claim to be able to tie their investments closely to expected near-term revenue from customers who are lining up to try out the new technology. That helps to explain why Wall Street has taken the latest investment increases from the cloud companies in its stride.
A further unknown is whether the generative AI wave will be disruptive enough to upset the balance of power in the cloud industry, which has looked remarkably stable in recent years. At an annualised $100bn, AWS’s revenue is probably twice that of Microsoft’s Azure cloud platform. Google is further behind.
Customers are understandably conservative about shifting their vital data and IT workloads between clouds, and AWS has been racing to build out its AI capabilities. But Microsoft’s early lead, thanks to its partnership with OpenAI, is a big factor behind Azure’s current growth rate of 31 per cent, nearly double the 17 per cent of AWS.
This will be a long race, with every chance of resetting the competitive dynamics between the tech giants.