Id: arxiv/2611.23771 Url: https://arxiv.org/abs/2611.23771 Tags: eess.SY, math.OC Score: 0 Title: Mitigating the Grid Impact of Gigawatt-Scale AI Data Centers via High-Fidelity Model Flexibility Domain: eess.SY Source: arxiv Authors: Eihab E. E. Ahmed, Saeed Manshadi Reviews: 0 Subject: eess.SY Category: research Comments: 0 Disputed: false Division: eess Mentions: 0 Discussed: 0 Published: 2026-11-18 Page: https://delphi.science/r/arxiv/2611.23771 Gigawatt-scale artificial intelligence data centers (AIDCs) are being connected at a single point of interconnection (POI) faster than transmission can be reinforced, yet studies of their flexibility rely on DC power flow and simplified load envelopes that omit voltage, cooling dynamics, and workload deadlines. This paper addresses this gap by embedding a high-fidelity AIDC model, with separate inference and training pools, hall thermal mass, PV, a battery energy storage system (BESS) with reactive capability, and a STATCOM, in a five-minute multi-period AC optimal power flow. Simulations on…