This user's experience is a perfect example of how infrastructure, not data complexity, becomes the bottleneck in modern workflows. Their blank spreadsheet, which responds instantly when saved locally, turns into a sluggish interface the moment it's stored on an overseas SharePoint server. That half-second lag on every keystroke isn't caused by formulas or large datasets, it's caused by distance. And the frustrating part is that most troubleshooting advice assumes the problem is the file itself, not the network path it travels.
What this reveals is a fundamental mismatch between how collaborative tools are designed and how they actually behave at scale. The user correctly notes that Excel is supposed to save a local copy and sync changes in the background, and that has worked for years. But when the SharePoint server is physically far away, the sync layer stops being invisible. Every action, typing a character, clicking a cell, triggers a handshake with that distant server, and the latency becomes tangible. Half a second per action may not sound severe, but it destroys the feeling of direct manipulation that makes a spreadsheet feel responsive. The user is no longer working in their file; they are negotiating with it.
The practical takeaway is blunt: cloud-based file storage is not a neutral upgrade. It introduces physics into your workflow, and physics does not care about your organization's IT policies. If your SharePoint server is on another continent, the speed of light alone imposes a delay that no amount of local caching can fully mask. The user's real question, what can they do to stop this behavior, has no satisfying answer within their current setup. Disabling autosave didn't help because the lag isn't about saving; it's about the constant polling and validation that SharePoint performs even during idle editing.
This is where the conversation needs to shift. Instead of searching for workarounds inside a tool built for a different era, users should ask whether their spreadsheet application should be handling real-time collaboration over high-latency links at all. The problem isn't that the user has too much data. The problem is that their tool is trying to behave like a local application while being forced to act like a remote one. A truly modern approach would separate the editing experience from the storage layer entirely, making the file feel local regardless of where the server sits. That is the direction the industry should be moving, and it's the only path that fixes this kind of lag without asking users to move offices or change continents.