Computer scientist Ayse Coskun spent 12 years asking whether data centers could flex their power use to match what the grid needs. The answer is yes. AI facilities run predictable, delayable workloads, making them ideal for soaking up excess solar midday and scaling back during peak demand. Her conductor platform runs on real AI data centers today, cutting power by pausing non-critical jobs without breaking performance promises. During a 2023 Texas heat wave, wholesale electricity prices spiked over 800% in a single afternoon. Flexible data center loads could have blunted that. It is a compelling case that the same technology straining the grid can also help stabilize it.