Elon's Former Battery Chief: AI Data Centers Will Make Electricity Cheaper

Gradient Dissent · 2026-08-18

This item is no longer on the current shelf. Its page remains available, and editorial corrections may be made. See what is featured now.

Why it earned a slot

Drew Baglino — the engineer who scaled battery technology at Tesla — is arguing that the grid-to-chip pipeline needs a fundamental redesign, not incremental upgrades. This is the infrastructure counterpoint to the political narrative: the physics is solvable, but it requires rethinking how power reaches the silicon, from substations to the rack level.

The short version

The battery engineer who electrified Tesla just told you the power grid can't handle AI — and what to do about it.

Why it matters

As inference demands scale exponentially with agentic workloads, the assumption that 'just build more data centers' breaks down against physical reality. Baglino's perspective from inside Tesla's energy operation — and now at Heron Power — brings a credibility to infrastructure questions that pure AI commentary lacks.

My take

The bottleneck in agentic systems that push inference hardware is increasingly not the model but the power delivery chain. Baglino's argument that this requires a ground-up overhaul — not bigger transformers — is the engineering reality the AI community is only beginning to price into its roadmaps.

How it connects

• Directly pairs with the data center opposition episode: Baglion offers the engineering pathway while the AI Breakdown documents the social resistance — both are constraints on the same deployment problem.
• The grid-to-chip redesign conversation connects to inference-optimization work — efficiency gains at the model level matter less if the power infrastructure can't deliver the electrons.

Bottom line

When planning compute infrastructure for agentic systems, start with the power delivery chain, not the GPU spec sheet.