PG&E Q3 2026 Earnings: Core EPS Falls 18% to $0.41 as 2027 Capex Is Cut $2B

On October 22, 2026, PG&E Corporation (NYSE: PCG) reported third-quarter GAAP net income of $776.0 million — down 8.7% year over year — or $0.34 per diluted share, while non-GAAP core earnings fell further, to $928.0 million or $0.41 per share, down 19%. Both declines landed in a quarter where operating income actually grew 3.3% to $1,249.0M: the damage sits below the operating line and inside two cost lines the regulator, not the customer, controls. Management reaffirmed full-year core guidance of $1.64–$1.66 anyway, cut $2 billion from the 2027 capital plan, and filed a 5,000-mile undergrounding plan it says will deliver $117 billion of long-term wildfire-risk benefits. This is the quarter that asks whether PG&E is a rate-base compounder with a cost problem or a cost story with a rate-base problem.

The Headline Numbers​

Every figure below comes from the Form 10-Q and the earnings release. (Consolidated Net Income on the face is $776.0M; $749.0M is available for common shareholders after the $27.0M preferred dividend requirement, and per-share figures divide that line.)

Read the table top to bottom and the quarter inverts twice. Revenue slipped 1% while operating income grew 3.3% — costs fell faster than sales — yet net income fell 8.7% and core earnings fell 19%. Start with the GAAP walk, because it is exact: pre-tax income actually rose $51.0M (to $681.0M), then the income-tax benefit shrank by $125.0M (to $95.0M from $220.0M) and that single swing erased the operating gain and more. The release names the culprits in order: a lower CPUC return on equity in effect during 2026, the one-time catch-up from the tax methodology change implemented in 2025, and higher Wildfire Fund expense — partly offset by earnings on a higher rate base and net O&M savings. Nine-month figures confirm the quarter is the weak leg, not the trend: revenue $18,971.0M (+4.6%), operating income $3,982.0M (+13.0%), net income $2,422.0M (+19.1%), GAAP EPS $1.06 (+19.1%).

The core number deserves its own caution. Core earnings exclude $156.0M after tax of non-core items this quarter versus $295.0M a year ago — the adjustments shrank, yet core still fell nearly 19%. When both GAAP and the adjusted metric fall by double digits while operating income rises, the story is the cost of being PG&E: taxes, wildfire fund amortization, and interest, up $55.0M (+7.1%) on a bigger debt pile.

Revenue Deep Dive​

PG&E is a single-utility company, so the revenue split is electric versus gas — and both moved on pass-through mechanics more than on demand.

Electric fell $95.0M, but the 10-Q's own framing says not to read that as weakness: procurement costs the Utility passes straight through to customers — electricity and gas purchases, public-purpose programs — move revenue and expense together and "do not impact net income." Cost of electricity fell $111.0M (−10.9%) in the same quarter, more than the electric revenue decline. The top line shrank because the fuel bill shrank, which is the mechanism working as designed.

Natural gas grew $33.0M (+2.2%) while cost of natural gas fell $22.0M (−16.8%) — the favorable direction on both lines, though gas is a quarter of revenue and cannot carry the quarter.

The demand language in the release is all load growth, and it is specific enough to check against the ledger later. "Connected over 4,240 electric customers and 3,000 EV charging ports to the grid" is the base business; "Energized 174 megawatts of new data center projects since 2024" is the growth vector, with the striking claim that "with appropriate pricing, every one gigawatt of new data center load could help customers save 1% or more on their monthly electric bill." That sentence is doing regulatory work — new load as a rate-relief argument — and it is the closest the release comes to a demand theme: no "demand exceeds supply," no backlog figure, just megawatts energized and a bill-math promise. The product-ramp slot goes to the "first-of-its-kind virtual power plant demonstration with Google and Rewiring America," targeting "over 20,000" customer devices and "12 megawatts of distributed energy resources by 2028" — a pilot, honestly labeled as one. Note the pricing absence too: there is no pricing momentum anywhere in the release, because a regulated utility does not take price — it takes allowed return, and the allowed return went down this year.

The Margin Story​

Operating margin is the honest series and it improved nearly a full point, extending a two-year climb from 18.3%. The mechanics: O&M fell $98.0M (−3.7%) to $2,543.0M — the "net O&M savings" the release cites, tracking the 2–4% non-fuel O&M reduction target — and procurement savings ($133.0M combined) more than covered the two structural cost risers, depreciation (+$74.0M, +6.5%, the growing rate base announcing itself) and Wildfire Fund expense (+$55.0M, +64.0%, to $141.0M). One caution on the O&M trend: nine-month O&M is actually up 0.5% ($8,191.0M vs $8,147.0M), so the savings are a second-half story the fourth quarter has to confirm.

Net margin fell 1.1 points on two lines that both live below operating income: the $125.0M tax-benefit comp and the $55.0M of extra interest, partly offset by $66.0M of higher interest and other income — that arithmetic is the whole $74.0M GAAP decline. (The $27.0M preferred dividend requirement sits below even this line, as a distribution of the $776.0M.) A utility trading on earnings yield has to be read this way — operating margin for the franchise, net margin for the capital structure — and this quarter the two pointed in opposite directions.

The One Big Question: What Does $0.41 of Core Earnings Leave Out?​

When management guides only to core EPS — $1.64–$1.66 for 2026, $1.78–$1.82 for 2027 — and cannot reconcile it to GAAP "without unreasonable effort," the reconciliation it does publish becomes the most-read table in the release. Here it is, consolidated from the release's own bridge (post-tax; adds to $929.0M before rounding, prints $928.0M):

Two lines deserve scrutiny, because "non-core" does not mean "non-recurring." The $102.0M of Wildfire Fund amortization is the scheduled burn of a 20-year asset — it will be in every quarter's bridge for a decade and a half, and it grew 65% year over year as the liability accretes. Excluding it from core is defensible (it is a pass-through-ish legacy cost), but it is the opposite of one-time. The $42.0M of strategic repositioning is new this quarter — $37.0M of workforce severance plus $21.0M of consulting, legal, and advisory "including those related to the Strategic Review" — and it is the cost of the review that is supposed to fix the balance sheet. One genuine credit: bankruptcy and legal costs printed a $29.0M benefit (a Wildfire-Related Securities Claims recovery), the Chapter 11 era paying a small dividend six years later.

Then do the guidance math the release invites but does not show. Nine-month core EPS is $1.24; the full-year guide is $1.64–$1.66, so the fourth quarter must deliver $0.40–$0.42 — essentially a repeat of this quarter's $0.41. That is a hittable number, not a heroic one, which is presumably why the guide was reaffirmed rather than raised: after a quarter where core fell 19%, promising a flat repeat is the credible call.

Tracking a $147 Billion Company in Plain Text​

A regulated utility is the hardest company to read in GAAP and the easiest to audit in Beancount: every dollar of rate base has to sit somewhere on the balance sheet, and double-entry forces the wildfire costs, the securitizations, and the preferred dividends to reconcile instead of hiding in footnotes. Our books follow how we model every company. The quarter's income-statement transaction, exactly as pushed (income negative, expenses positive, so the whole thing sums to zero):

; FY2026Q3 Income Statement — quarter ended September 30, 2026
; Check: −6188 + 904 + 109 + 2543 + 1206 + 825 − 95 + 36 + 0 + 141 − 104 − 153 + 776 = 0 ✓
2026-09-30 * "PG&E Corporation" "FY2026Q3 Income Statement"
  Income:Revenue                          -6188 MUSD
  Expenses:CostOfRevenue                     904 MUSD
  Expenses:CostOfRevenue                     109 MUSD
  Expenses:OperationsAndMaintenance         2543 MUSD
  Expenses:DepreciationAndAmortization      1206 MUSD
  Expenses:Interest                          825 MUSD
  Expenses:IncomeTax                         -95 MUSD
  Income:OtherNet                             36 MUSD
  Income:OtherNet                              0 MUSD
  Income:OtherNet                            141 MUSD
  Income:OtherNet                           -104 MUSD
  Income:OtherNet                           -153 MUSD
  ; preferred stock dividend requirement of subsidiary, 27: a distribution of Net Income, not an expense (no posting)
  Equity:Adjustments                         776 MUSD

Two readings worth pointing at. First, Expenses:IncomeTax is negative — a $95.0M benefit, because deferred tax credits exceed the current provision in every period of this ledger. Second, the Income:OtherNet lines carry the whole wildfire apparatus as labeled postings: the $36.0M SB 901 charge, the $141.0M Wildfire Fund expense, and a $0 wildfire-claims line that prints "—" on the face this quarter — the dog that didn't bark, nine months after the 2025 quarters each carried nine-figure claims.

The one balance-sheet number that tells the narrative is reinvested earnings: positive $1,357.0M at September 30, against a $650.0M deficit at New Year and a $2,966.0M deficit at the end of FY2024. PG&E earned its way out of a deficit that dated to the bankruptcy era — while paying $330.0M of common dividends in nine months, nearly quadruple the FY2024 pace. Behind it sits the machine that did it: net PP&E of $102,555.0M, up $14,326.0M (+16.2%) since FY2024, funded by $65,026.0M of total debt at a steady 65% of capitalization.

The Multi-Year Arc​

Three columns, one compounder with a governor on it. Revenue grows low-single digits, operating income grows faster (+6.5% in FY2025, +13.0% on a nine-month basis), and PP&E compounds mid-teens as capex runs far above depreciation — 9M capex of $9,621.0M (+11.5%) went mostly to electric transmission and distribution hardening, capacity, and interconnection. The dividend tells the confidence story in one line: $86.0M → $220.0M → $330.0M in nine months, alongside a no-new-equity pledge through 2030. Against this stands the governor: the 2027 capital plan cut from $13.4 billion to $11.4 billion, with the 2028–2030 forecast withdrawn pending the strategic review. For a utility, capex is future rate base is future earnings — a 15% cut to next year's investment is a 15% cut to the growth engine, whatever it does for the credit metrics. The review's stated objective, "becoming financially strong and securing an investment-grade credit rating," names the trade explicitly: PG&E is choosing balance-sheet repair over growth, and the $117 billion undergrounding plan is the growth story it gets to keep while it does.

The Verdict: Bull vs. Bear​

Bull Case

• Rate base compounds regardless of the quarter: net PP&E up $14.3B (+16.2%) since FY2024, and 9M capex of $9.6B (+11.5%) keeps the machine fed even after the 2027 trim.
• O&M discipline is real and measurable: Q3 O&M down 3.7% while operating margin expanded to 20.2%, tracking the 2–4% non-fuel reduction target.
• The wildfire overhang is shrinking on the balance sheet, not just in prose: wildfire-related claims payable down from $916.0M (FY2024) to $173.0M, and Q3 claims printed nil.
• The deficit is gone: reinvested earnings positive $1,357.0M for the first time in the ledger's window, while common dividends nearly quadrupled off the FY2024 base.
• The guide requires no heroics: Q4 needs $0.40–$0.42 of core EPS against Q3's $0.41, and the company pledges no equity issuance through 2030.

Bear Case

• The CPUC cut the allowed return, and nothing in the release says by how much or for how long — the single input a regulated utility cannot control moved against earnings power.
• Core earnings fell 19% in a quarter the company calls solid, and the $102.0M Wildfire Fund add-back (+65%) recurs every quarter for 20 years — "core" flatters by construction.
• The $2 billion 2027 capex cut plus the withdrawn 2028–2030 forecast is deferred growth: 15% less investment next year means 15% less future rate base, whatever the credit benefit.
• Leverage keeps climbing into the review: $65.0B of total debt (+13.6% since FY2024) at 65% of capitalization, with 9M operating cash flow down 13.6% to $5,840.0M while capex rose.
• Data-center load is 174 MW energized — a rounding error on a 70,000-square-mile grid — and the "1% bill savings per gigawatt" math cuts both ways if the load does not come.

Our Take: Buy the rate base, discount the core number. PG&E put $14.3 billion of steel in California ground in 21 months, erased a bankruptcy-era deficit, and held operating margin expansion while cutting O&M — that is a franchise working. But $0.41 of core EPS that needs a $102 million recurring add-back to exist, a cut ROE, and a shrinking capital plan is not a growth print, and the strategic review reads as management agreeing with the bears about the balance sheet. The undergrounding plan is the swing factor: if 5,000 miles at $117 billion of promised benefits survives the CPUC process, the capex cut is a pause; if it does not, it is the growth story. Until the review concludes, this is a 65%-debt utility paying a ramping dividend to wait — fairly priced patience, not a compounder's discount.