M5 Max buyers split on 2TB, 4TB storage
This Reddit discussion asks a practical question for M5 Max buyers with 128GB unified memory: is 2TB internal storage enough for LLM-heavy workflows, or does 4TB become the safer buy once model files, datasets, and tooling start accumulating? The post frames the tradeoff well, balancing Apple’s faster internal SSDs against the cost and complexity of relying on external TB5 NVMe storage. The strongest early responses lean toward 4TB for peace of mind, while still acknowledging that 2TB can work if the workflow is disciplined and external storage is part of the plan.
The angle is solid because it moves past spec-sheet hype and into real ownership friction: capacity, bandwidth, port usage, and workflow discipline.
- –The post is strongest where it distinguishes “can I make 2TB work?” from “will 4TB remove friction?”
- –For LLM users, storage pressure usually comes from model files, caches, datasets, and local project sprawl, not just raw app installs.
- –The argument for 4TB is less about absolute speed and more about keeping the whole workflow local, simple, and portable.
- –The 2TB case is still credible if the user is comfortable offloading cold data to external NVMe and treating internal storage as a fast working set.
- –The post could be even sharper if it explicitly names the kinds of files that will actually live on disk in an LLM setup.
DISCOVERED
45d ago
2026-04-26
PUBLISHED
45d ago
2026-04-26
RELEVANCE
AUTHOR
_derpiii_