AI driving cloud service providers CAPEX

• Microsoft and Amazon are among the highest spenders as they invest in data center development.
•  Microsoft will spend over 13% of its capex on AI infrastructure.
• Chinese hyperscalers’ capex is decreasing due to their inability to access NVIDIA’s GPU chips, and decreasing cloud revenues.

As the focus on artificial intelligence continues to dominate conversations and investments all around the world, it is not surprising that cloud service providers are also looking to be part of it. Cloud service providers already have some form of automation and AI in them and are now looking to increase their capital expenditure (capex) in AI.

In cloud computing, AI is embedded into the right infrastructure to automate routine processes and streamline workloads. For example in a hybrid cloud environment, AI tools can be used to monitor, manage and self-heal individual public and private cloud components.

According to research from Coutnerpoint’s Cloud Service, global cloud service providers are expected to grow their capex by an estimated 7.8% YoY in 2023. The biggest spenders are expected to be Microsoft and Amazon, who will account for 45% of the total capex. In fact, US-based hyperscalers will contribute to 91.9% of the overall global capex in 2023.

Based on Counterpoint estimates, Microsoft will spend proportionally the most on AI-related infrastructure with 13.3% of its capex directed towards AI, followed by Google at around 6.8% of its capex. Microsoft has already announced its intention to integrate AI within its existing suite of products. For context, Microsoft’s cumulative investment earlier this year in OpenAI has reportedly swelled to US$13 billion.

Interestingly, in their latest earnings call, Microsoft called for lower revenue guidance than analysts had predicted, partly because of weakness in the segment that contains Windows. According to a report by CNBC, the company’s research and development costs declined for the first time since 2016.

The report also mentioned that Microsoft’s Intelligent Cloud segment contributed US$23.99 billion in revenue, up 15% and above the US$23.79 billion consensus of analysts surveyed by StreetAccount. The unit comprises the Azure public cloud, SQL Server, Windows Server, Visual Studio, Nuance, GitHub and enterprise services.

Over at Google, the news report stated that revenue from Google’s cloud products, which includes Google Workspace productivity apps in addition to Google Cloud Platform, increased by 28%.

Counterpoint also estimates that around 35% of the total cloud capex for 2023 is earmarked for IT infrastructure including servers and networking equipment compared to 32% in 2022. AI infrastructure can be 10x-30x more expensive than traditional general-purpose data center IT infrastructure.

“Hyperscalers are increasingly focusing on ramping up their AI infrastructure in data centers to cater to the demand for training proprietary AI models, launching native B2C generative AI user applications, and expanding AIaaS (Artificial Intelligence-as-a-Service) product offerings,” explained Senior Research Analyst Akshara Bassi.

According to Counterpoint’s estimates, around 35% of the total cloud capex for 2023 is earmarked for IT infrastructure including servers and networking equipment compared to 32% in 2022.

It’s a different scenario in China though. Chinese hyperscalers are spending less due to slower growth in cloud revenues amid a weak economy and difficulties in acquiring the latest NVIDIA GPU chips for AI due to US bans. The scaled-down version – A800 of the flagship A100/H100 chips – that NVIDIA has been supplying to Chinese players may also come under the purview of the ban, further reducing access to AI silicon for Chinese hyperscalers.

While Chinese players are investing a larger portion of their spending towards AI, the amount is significantly less than that of their US counterparts due to a lower overall capex. Chinese cloud service providers like Alibaba Cloud, Huawei Cloud and Tencent continue their R&D with AI in the cloud.

Alibaba Cloud, for example, has also added Meta’s Llama 2 large language model to ModelScope, its open-source “model-as-a-service” platform that makes available more than 700 AI models. According to SCMP, the model covers various fields from computer vision to natural language processing and audio to local developers.

---

---

---

---