PEBBLE ACADEMY · Partnership
Why Carbon-Aware Compute Matters
Not every kilowatt-hour is created equal. Depending on where and when a workload runs, the same job can emit two-to-five times more CO₂ than its lowest-impact alternative — purely because of the energy mix flowing through the grid at that moment.
That's a huge optimization surface, and most platforms ignore it entirely. Carbon-aware compute is the practice of routing or scheduling workloads to consume electricity when and where the grid is cleanest.
What WattTime Brings
WattTime maintains a global, real-time view of marginal grid emissions: which generating sources are responding to incremental demand right now, and what that means for the carbon intensity of any new compute you spin up.
Their Marginal Operating Emissions Rate (MOER) signal is the gold standard for grid-aware scheduling. It tells you, minute by minute and region by region, the marginal CO₂ per kilowatt-hour your next job would emit.
How the Integration Works
Pebble pulls WattTime's MOER signal continuously and feeds it into our orchestration layer. When a workload is flexible (batch jobs, ML training, data pipelines, analytics), we evaluate three things in real time: the carbon intensity now, the forecast over the next several hours, and the user's deadline.
If shifting the job by 30 minutes cuts its emissions in half without missing the deadline, we shift it. If it has to run now, we run it — but we report the emissions impact transparently.
The Result
Customers using Pebble + WattTime typically see 20–40% reductions in workload-attributed CO₂ with no code changes and no measurable hit to user-facing latency. The shift is invisible to the application and visible to the sustainability dashboard.
Carbon-aware compute used to require building a custom scheduler. Now it's a switch.