Liftr Insights data show the rise and growing influence of Arm in global cloud
In Q4 of 2024, 1 of every 4 new processor types deployed in the major public cloud providers were on Arm-based CPUs, as shown by Liftr data. This was the highest percentage of Arm deployments since the first sighting of Arm in the clouds over half a decade ago. Enterprises are taking notes as they manage their costs.

Microsoft Azure, whose Cobalt 100 processor became available in October, heavily influenced these new Arm-based deployments. Google also introduced their Axion Arm-based instance in October 2024. This strong Q4 places Arm-based deployments at just below 10% of the total processors deployed in the major public clouds as of December 31, 2024. As more cloud vendors release or improve their own Arm-based cloud-specific processors, will we see a continued disruption of the x86 leadership position in 2025?

Google also introduced their Axion Arm-based instance in October 2024.
We have been closely watching the rise of Arm in the public cloud, and we are seeing Arm’s rate of growth accelerate as more derivatives become available.
So, why did cloud vendors pursue an Arm-based strategy?
-
Roadmap control: Relying solely on X86 processors also means relying on Intel and AMD’s roadmaps and ability to deliver. This translated to little opportunity to differentiate from other cloud vendors.
Cost savings to the cloud provider have also translated to reduced customer prices.
-
Product margins: By licensing Arms IP and taking on design and manufacturing responsibility, cloud vendors can reduce the unit costs of processors.
-
Sustainability: In addition to product margins, power costs are an increasingly difficult part of cloud operating expenses, and it’s no longer just about cost as even power supply is becoming an issue. Arm processors are typically understood to require significantly less power than their x86 counterparts.
These advancements translate to very tangible customer benefits.
-
Reduced costs: Cost savings to the cloud provider have also translated to reduced customer prices. For instance, Liftr data show that Arm instances are on average 49.2% cheaper per core than X86. A Cobalt instance on Azure is 40.9% less per hour than a similarly specced Intel X86 instance. While assessing a comparable AMD instance was more difficult, a similar Cobalt instance that included temporary storage was 20% cheaper than a comparable AMD instance without temporary storage. Additionally, some of the Cobalt instances have more premium features available. More on Cobalt in a future article.
Source: Liftr Insights Source: Liftr Insights - Choice: Customers now have more choices for specific types of workloads. For example, Liftr data includes many Arm options across all the major cloud providers.
- Sustainability: While the cloud vendor’s sustainability benefits result in lower costs, many customer IT departments also have sustainability KPIs. Selecting more power-conscious processors as can be tracked with Liftr Insights will help maintain or exceed these goals.
We expect the architecture's upward trend to continue, especially since multiple cloud providers announced new and improved Arm-based solutions in the fall of 2024.
Liftr will examine the different cloud providers and their Arm strategies in greater depth in additional articles, including specifics on their offerings, competition, and historic market performance. The next article we will dive into Microsoft Azure and Cobalt.
Calculations using Liftr data, supported by more than half a decade of trends, demonstrate that the benefits of using Arm architectures can be significant.
Corporations and organizations managing their bottom line and sustainability goals can keep up-to-date with Liftr data to forecast and plan. One simple solution is leveraging Liftr’s Supporting Data for Global AI & Cloud reports.
Feel current (and informed) with first notice of articles like these; subscribe to the Liftr Newsletter.