2026-09-24 Changes confirmed medium confidence

AMD and Perplexity Bring Portable Computer to Ryzen AI Max PCs

Perplexity subscribers can run supported local-agent tasks and recurring workflows on Windows machines using AMD's high-memory Ryzen AI Max platform.

Edited by Tyronne Panaino

AMD announced on September 24 that Perplexity's Portable Computer can now run on supported Windows PCs powered by Ryzen AI Max Series processors, including AMD's Ryzen AI Halo developer platform. The official release says Perplexity subscribers can execute local AI tasks and recurring workflows on their own hardware without spending Perplexity Computer credits on that local inference.

The change matters to developers and advanced PC users because it gives a consumer-facing agent workflow a local execution path on a defined hardware family. It does not make every step private or offline by default, and the announcement does not document the local models, compatible applications or memory requirements in enough detail to generalize beyond supported Ryzen AI Max systems.

Portable Computer gains a local execution option

AMD describes Portable Computer as the layer that brings models, agent tools and user controls together. On the newly supported machines, some tasks can run directly on the PC rather than sending every inference request to Perplexity's cloud service. The release also says the product can still use frontier-scale cloud reasoning when needed.

That hybrid design is the material change. A local path can reduce dependence on remote inference for selected work and lets recurring activity use the user's machine. A cloud path can remain available for workloads that the local system or chosen model cannot handle. The source does not identify which decisions are automatic, which are controlled by the user, or how clearly the interface distinguishes local from remote execution.

The no-credit statement is also narrow. AMD says local inference does not consume Perplexity Computer credits; it does not say that every feature, tool call, connected service or cloud escalation is free of plan limits. Readers should treat eligibility and usage boundaries as product settings to verify in their own Perplexity account.

Ryzen AI Max supplies shared compute and memory

AMD positions Ryzen AI Max as a combination of CPU and GPU performance with a large shared-memory pool. That arrangement can help a local model and ordinary Windows applications remain available in the same workflow without requiring a separate accelerator card. The Ryzen AI Halo developer platform is named as one supported machine, but the announcement does not provide a complete device list.

The hardware claim is about availability, not a comparative performance result. AMD publishes no task-level latency, energy-use, accuracy or throughput measurements in this release. It also does not establish how Portable Computer behaves on different Ryzen AI Max configurations or how much memory a particular workflow consumes. Those details are important before teams infer that all systems in the family will deliver the same experience.

Local execution does not answer every privacy question

Running inference on the PC can reduce the amount of work that must leave the device, but local execution alone does not prove that a workflow is offline or that no data reaches a remote service. Agent tasks may use cloud reasoning, external websites, connected applications or synchronized account state. The AMD announcement does not provide a data-flow diagram, retention policy or per-tool disclosure for Portable Computer.

Organizations considering sensitive workflows should therefore test the actual product configuration. Useful checks include observing network activity, confirming when cloud reasoning is invoked, reviewing connected-app permissions and verifying where recurring-task state is stored. Those checks would turn a broad local-compute claim into evidence about a specific deployment.

What to watch next

The clearest next checkpoint is Perplexity documentation that lists supported Ryzen AI Max systems, plan eligibility, local model choices, memory requirements and the boundary between on-device and cloud work. Independent measurements of latency, power use and task reliability would also show whether the local path changes practical agent use rather than merely adding another deployment option.

For now, the supported claim is precise: Portable Computer is available to Perplexity subscribers on supported Ryzen AI Max Windows PCs, local inference can avoid Computer-credit use, and the workflow may still reach cloud reasoning. Claims about universal compatibility, complete offline operation or superior performance would go beyond the source.

Status

Confirmed. AMD's official newsroom documents the September 24 availability change and the stated local/cloud workflow. Internal confidence is medium because the release is a single vendor source, Perplexity's linked product page could not be fetched in this run, and the feature was not independently tested.