1. SGIP Step 7 closes in eleven days; equity-resiliency budget still half-spent

PG&E’s large-storage incentive bucket is on track to exhaust before the end of Q2, while the equity-resiliency carve-out remains structurally undersubscribed for the sixth consecutive month.

The Self-Generation Incentive Program large-scale storage budget for PG&E territory at Step 7 is being consumed at roughly $2.1M per business day, against $23.4M of headroom as of Friday’s Stats Tool refresh. At that pace the bucket closes in eleven business days — on or about May 13 — six weeks earlier than program staff projected at the December 2025 workshop.¹ Three implications for developers running active interconnection processes:

SGIP Step 7 closes May 13, on a straight-line of the last 10 business days. Equity-resiliency does not close.

• Projects already past Reservation Confirmation are unaffected. The closure does not retroactively void approved reservations — only new applications.
• Projects sitting at “Application Received” without an Initial Engineering Review are at material risk. Of the 47 PG&E projects in that state at last refresh, six are ours-tracked at >5 MW and have been pending IER for more than 60 days.
• The equity-resiliency carve-out, which sits in a separate budget line, is unlikely to absorb redirected demand because of its CalEnviroScreen 4.0 75th-percentile threshold and Tier 3 fire district overlay.² Eligible territory is geographically constrained, not budget-constrained.

The asymmetry between large-storage exhaustion and equity-resiliency under-utilization is now structural. Equity-resiliency closed Q1 2026 with $84.2M unobligated, against $19.7M obligated — an obligation rate of 19.0%, the lowest in the program’s post-2020 history.³ Program staff have advanced two interpretations: the Tier 3 fire district overlay narrows the eligible parcel set faster than CalEnviroScreen widens it, and the documentation burden — specifically the Customer Vulnerability Letter requirement — effectively gates participation to organized non-profit and tribal applicants.

What we’re watching: whether CPUC Energy Division re-opens R.20-08-020 to relax the Customer Vulnerability Letter requirement before the August workshop. The Cal Advocates filing of March 27 explicitly recommends a 60-day notice to Joint Parties on this question, and the ALJ has not ruled.⁴

Bucket	Budget remaining	Burn rate (last 10 BD)	Projected close
PG&E Large Storage Step 7	$23.4M	$2.1M/BD	May 13, 2026
PG&E Small Residential	$11.1M	$0.4M/BD	Jun 26, 2026
PG&E Equity-Resiliency	$84.2M	$0.2M/BD	Open through 2027
SCE Large Storage Step 7	$41.0M	$1.4M/BD	Jul 8, 2026

2. CPUC proposed decision would reclassify exporting fuel cells under NEM-3 successor

A draft proposed decision in R.20-08-020 reframes “net-importing fuel-cell systems” for tariff treatment — a change with material payback consequences for parallel-operation deployments at hospitals, data centers, and biotech sites.

The Administrative Law Judge in R.20-08-020 issued a draft proposed decision Friday morning that would reclassify any fuel-cell or microturbine system designed to export above a 5% nameplate threshold as a Net Energy Metering 3.0 successor (“NEMS”) generator, rather than under the existing parallel-operation tariff Schedule NEMVMASH.⁵ Comments are due May 19, with reply comments June 2.

The mechanical effect of the change is to subject occasional export — the kind that happens when a 1.4 MW Bloom array oversupplies a hospital chiller cycling off — to NEMS export-credit math rather than the avoided-cost rate. For a typical PG&E B-19 site, that delta is on the order of $0.06 to $0.09 per exported kWh, depending on hour-of-day. At sub-5% export volumes, the annual revenue drag is small, but the documentation burden of NEMS interconnection (separate Form 14-708 filings, IEEE 1547-2018 inverter recertification for fuel-cell power conditioning) is non-trivial.

The change is not about revenue at <5% export. It’s about reclassifying the interconnection paperwork, which adds 60–90 days at the front of every project.

Three things matter for developers in this proceeding:

• The 5% threshold is the load-bearing parameter. SEIA’s March 12 reply comment proposed a 10% threshold; CALSSA proposed a no-threshold blanket exemption for fuel cells under 1.5 MW.⁶ The ALJ split the difference. Expect industry pressure on the threshold itself, not on the categorical reclassification.
• The grandfathering language is narrow. Projects with executed interconnection agreements before the effective date are protected; projects in pre-IEA review are not. There are 39 fuel-cell projects in PG&E’s pre-IEA queue today.⁷
• NEMS export-credit math interacts with B-19 demand-charge timing. A site that exports during 5–9 PM peak periods will see modest gains under NEMS; a site that exports midday will see losses. Run the per-site arithmetic; the headline number from your OEM is not predictive.

Bcal Intel will publish a comment-template note for paid subscribers on May 8 with a four-page filing scaffold, the existing Joint Parties contact list, and the relevant prior-decision citations. If you intend to file in R.20-08-020 and want the scaffold, watchlist the docket from your dashboard or reply to this email.

3. Bloom Energy 10-Q signals 38 MW of California behind-the-meter bookings; deposits up 19% Q/Q

The Q1 2026 10-Q line-item disclosure reframes the company’s California position as load-driven, not subsidy-driven. Three numbers warrant pulling out of the filing.

Bloom Energy filed its Q1 2026 10-Q on Wednesday after market close. Three line items are worth surfacing for behind-the-meter analysts:

1. California BTM bookings are disclosed at 38 MW for the quarter, against 29 MW in Q4 2025 and 21 MW in Q1 2025. The year-over-year delta is +81%, well above the company’s blended global YoY of +34%.⁸
2. Customer deposits sit at $191M, up from $160M in Q4 2025 (+19% Q/Q), the highest balance since Q3 2022. Deposit duration to revenue recognition averaged 7.3 months in Q1 2026, slightly extended from 6.8 months in Q4. The deposit-to-bookings ratio is 1.4×, suggesting either a price-mix shift toward larger projects or pre-payment terms tightening on the customer side.⁹
3. Operating-leverage commentary in the MD&A section explicitly cites “California PG&E and SMUD demand-charge environments” as the leading geographic driver of the deposit acceleration. This is the first quarter in which California, rather than South Korea, is named first in that paragraph.

California, not South Korea, is the named leading geography in Bloom’s Q1 2026 demand-driver paragraph. That language has not appeared since Q2 2021.

The MW figure here is a forward-disclosure of bookings, not a delivered-asset count. Bloom recognized 14 MW of California revenue in Q1 2026 against the 38 MW of new bookings. The implied California backlog is 161 MW as of March 31, 2026.¹⁰

What this implies for the rest of the OEM stack: at Bloom’s implied California pricing, the 161 MW backlog represents roughly $640M of contracted revenue. If Mainspring, Plug, and Doosan follow the same demand-driver playbook in their respective Q1 disclosures (Mainspring is private; Plug and Doosan file in May and August respectively), the California BTM bookings consensus across the OEM stack would print between 95 MW and 130 MW for Q1, against a CAISO interconnection queue throughput that has averaged 60 MW per quarter for the technology cluster.¹¹ In other words: bookings are running ahead of physical interconnection capacity. Either the CAISO cluster cycle accelerates, or 30–60 MW of pre-IEA backlog enters 2027.

For lenders: the 19% Q/Q deposit growth, against backlog disclosure that is only +12% Q/Q, means deposits are growing faster than bookings. That gap usually closes one of two ways — deposit terms loosen (BTM customers stop pre-paying), or backlog catches up. We’ll watch this in the Q2 print.

4. Forward calendar — the next 30 days

Date	Event	Source
May 5	FERC Open Meeting — Order 2222 implementation review	FERC
May 8	Bcal Intel R.20-08-020 comment-template note (paid subscribers)	Bcal Intel
May 13	Projected SGIP PG&E Step 7 large-storage close	SGIP Stats Tool
May 14	Plug Power Q1 2026 10-Q expected	SEC EDGAR
May 19	Comments due, R.20-08-020 ALJ Proposed Decision	CPUC
May 22	VA Palo Alto on-site generation RFI responses due	SAM.gov
May 23	Travis AFB resilient-generation RFI responses due	SAM.gov
May 27	CEC LDES Round 4 program design workshop	CEC
Jun 2	Reply comments due, R.20-08-020	CPUC
Jun 4	SMUD board — GS-T schedule revision second reading	SMUD

END OF ISSUE 04 · NEXT ISSUE TUESDAY MAY 5, 2026 · 06:00 PT

Disclosures

Bcal Intel is published by Bcal Energy, an active behind-the-meter fuel-cell developer. We do not write up Bcal Energy projects, customers, or pending deals. Bloom Energy is a competitor of Bcal Energy at the OEM layer; this issue’s coverage of Bloom is editor-flagged. We hold no securities positions in any company referenced.

Nothing in this issue is investment, legal, regulatory, or engineering advice. Verify primary sources before making decisions.