Arm unveils two new products for cloud and edge applications, the Neoverse N1, which focuses on compute speed, and the Neoverse E1, focusing on throughput
For the longest time, Arm was basically synonymous with chip designs for smartphones and very low-end devices.
Context & Ripple Effects
This 2019 launch is the founding move of Arm's infrastructure business: after years of being synonymous with smartphone and low-end chip designs, Arm splits its first dedicated cloud-and-edge line into two cores — the compute-oriented Neoverse N1 and the throughput-oriented E1 — rather than adapting phone silicon. The split matters because it treats server and edge workloads as distinct design targets from day one.
Everything in the related coverage descends from this decision: the follow-on Neoverse V1 and N2 platforms claimed at least 40% better processing over their predecessors, V2 was positioned to power Nvidia's Grace CPU, and by 2024 the V3 and N3 designs were shipping up to 128 cores to customers for server, cloud, and infrastructure use.
First-order effects
- Cloud and edge chip designers licensing Arm IP immediately get two purpose-built starting points instead of one — an N1 path optimized for raw compute speed and an E1 path optimized for throughput-heavy workloads like networking and storage.
- Hyperscale customers evaluating Arm-based servers gain a credible alternative to repurposed mobile cores, giving Arm's licensees a concrete product to take into datacenter design cycles.
Second-order effects
- Incumbent server CPU vendors face competition specifically at the throughput-per-watt tier where E1-class cores are aimed, pressuring pricing and roadmap priorities in edge and networking silicon.
- The N1's compute focus sets up the accelerator pairings that define the line's later trajectory — most visibly Nvidia's Grace CPU on Neoverse V2, and eventually NVLink Fusion integration letting Neoverse CPUs couple with AI accelerators.
Third-order effects
- If the cadence holds — a new Neoverse generation roughly every two years through V3/N3 — infrastructure compute consolidates around licensable, workload-specific core families rather than vertically integrated proprietary designs, reshaping who captures value in the server supply chain.
- The compute-versus-throughput split established here becomes the template for how Arm segments future platforms across HPC, hyperscale, and edge, making core-selection strategy a first-order architectural decision for cloud builders.
The trend: Arm is executing a decade-long expansion from mobile IP licensor to full-line infrastructure compute vendor, with each Neoverse generation pulling it deeper into the datacenter and toward direct integration with AI accelerators.