GridTech is the domain that surprises people. It's also the one where the next decade is going to be most consequential, and least well-understood by the broader tech industry.
A few framing observations:
AI is colliding with the grid in two directions. AI workloads — training and inference — are pushing data-center demand into curves no utility planner sketched a decade ago. At the same time, AI techniques are starting to be applied inside the grid: forecasting, optimization, anomaly detection, asset health. The first collision is operational; the second is methodological. Both are real. Most of the public conversation only covers the first.
Distributed Energy Resources are not a sidecar. Solar PV, storage, EV charging, demand response, and increasingly behind-the-meter generation have crossed the threshold where they shape grid operations rather than perturb them. That changes what "the grid" even means: from a centrally-dispatched generation-and-transmission system into something closer to a coordinated mesh. The architecture implications are large and underdiscussed.
OT/IT convergence is real but uneven. Modernization initiatives keep finding the same friction: control-system protocols designed for trusted networks, IT security paradigms that don't fit operational tempo, vendor stacks with decade-long lifecycles, and regulatory frameworks calibrated for the prior era. The convergence is happening; pretending it's mostly an IT problem is the failure mode.
Industry standards have lagged the technology curve, then caught up in bursts. NERC CIP, IEC 62443, IEC 61850 — each was a meaningful step at the time, and each is showing seams the next round will need to address. The interesting question isn't whether the next revision will arrive, but whether it can keep up with the pace of OT/IT convergence and AI adoption.
Posts under this section will mostly be industry-perspective writing: grid modernization patterns, the AI/energy collision, OT/IT convergence as I see it, and selected reflections on protocols and standards.
Recent posts
AI Compute Is a DER, Not a Crisis
Most takes on AI's collision with the grid treat compute as fixed demand and ask how to grow supply. The more interesting question is what happens when you treat AI workloads as a Distributed Energy Resource — dispatchable, time-shiftable, geographically mobile, and pairable with on-site storage.