Cloud busting: the disruptive potential impact of AI on computing platforms - FT中文网
登录×
电子邮件/用户名
密码
记住我
请输入邮箱和密码进行绑定操作:
请输入手机号码,通过短信验证(目前仅支持中国大陆地区的手机号):
请您阅读我们的用户注册协议隐私权保护政策,点击下方按钮即视为您接受。
FT商学院

Cloud busting: the disruptive potential impact of AI on computing platforms

Industry sees a rebound but new technology will change competitive dynamics in the long term

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.

richard.waters@ft.com

版权声明:本文版权归FT中文网所有,未经允许任何单位或个人不得转载,复制或以任何其他方式使用本文全部或部分,侵权必究。

我们是否正处于核复兴的边缘?

为什么小型核电站会对可再生能源构成巨大威胁?

内塔尼亚胡与伊朗的战争:“对他来说,这是私人恩怨”

在哈马斯于10月7日发动袭击后,这位以色列总理的政治生涯似乎已经走到尽头。但如今,他正推动着一场自己多年来一直主张的冲突。

玛格丽特•米切尔:通用人工智能不过是“氛围和蛇油”

人工智能伦理领域的先驱之一解释了为何人类需求应成为科技发展的核心驱动力。

谁能在伊朗问题上影响特朗普?

从JD•万斯到“猩猩”,MAGA忠诚支持者和军方领导人正争夺在椭圆形办公室的影响力。

为什么华尔街害怕一个33岁的政治局外人

进步派候选人佐赫兰•马姆达尼搅动了纽约市长选举,城市精英们想要阻止他。

以色列空袭伊朗伊斯法罕核设施,特朗普权衡是否介入战争

美国总统认为欧洲领导的停火谈判无效。
设置字号×
最小
较小
默认
较大
最大
分享×