PG&E awards $30 million to six community microgrids and names one of them

PG&E announced on 17 September 2026 that it had selected six projects through the second application window of its Microgrid Incentive Program. The awards run between $2 million and $6 million each, total $30 million, and serve more than 2,200 customers across Humboldt, Sonoma, San Mateo, Butte, Placer and El Dorado counties. The release names one of the six.

The named one is in Pescadero, on the San Mateo County coast. It is led by WestLight Energy, a community choice aggregator, pairs about 1.5 MW of solar with 2 MW of battery storage, serves more than 200 customers, and is designed to run islanded for at least 24 hours through a wildfire or a storm. That is a specific, checkable description of a community-led asset. For the other five, the microgrid trade publication mgrid.org reported on 18 September that PG&E published no names, no capacities, no individual award amounts, no construction schedules and no islanding durations. Five counties, and nothing else.

The programme is neither small nor new. It originates in Senate Bill 1339, passed in California in 2018, which directed the state to develop microgrid policy. The California Public Utilities Commission authorised $200 million for the Microgrid Incentive Program, split $79.2 million to PG&E, $83.3 million to Southern California Edison and $17.5 million to San Diego Gas and Electric. PG&E awarded $43 million to nine projects in a first round in 2025. With this round it has committed roughly $73 million of its $79.2 million, and the announcement does not account for the remaining $6 million or so.

Eight years after the statute, then, the second and near-final tranche of the largest Californian utility's community microgrid money reaches something over 2,200 customers. That number is worth stating plainly rather than as a share of anything, because the useful comparison is not with California's population. It is with what the same money buys when it is not routed through a competitive grant round administered by a utility.

The disclosure problem compounds it. mgrid.org's practical objection is that a developer preparing a bid into any future window now has "one data point and five blanks". A grant programme that does not publish what it bought cannot be evaluated by the people it is inviting to apply, and it cannot be evaluated by the ratepayers funding it either. There may be good reasons for the silence: projects at an early stage, agreements unsigned, hosts who have not consented to being named. PG&E has not given a reason, and an unexplained redaction and a considered one look identical from outside.

Underneath the transparency point is a structural one, and it is why this belongs in a decentralization publication rather than only in a California trade sheet. The Microgrid Incentive Program exists because the distribution system fails. It is a response to public safety power shutoffs and storm outages on poles and wires that a utility owns, operates and is paid to maintain. The grants that let a community build its way around those failures are administered by that same utility. The CPUC designed it that way deliberately, and there are defensible reasons: the interconnection expertise sits there, and so does the outage data.

But it does mean the naming question carries the whole weight. A community microgrid and a utility-sited resilience asset with local branding look the same in a press release and are completely different things in a blackout, because they differ in who decides when to island, who is inside the boundary, and who owns the equipment afterwards. The Pescadero entry answers those questions. The other five, which between them account for most of the $30 million, do not answer any of them.