PG&E Corporation announced on September 2, 2026 that its board had established a Strategic Review Committee of four independent directors to evaluate, in the company’s words, “the full range of regulatory, financial, operational and strategic alternatives reasonably available, including the full range of options related to how the Company is organized and financed.” Becoming a financially strong, investment-grade company is named as a foundational objective.

At the same time PG&E revised its 2027 capital plan, deferring approximately $2 billion of planned investment to leave roughly $11.4 billion, and cutting its debt financing needs by the same $2 billion. It reaffirmed 2026 core earnings guidance of $1.64–$1.66 per share and initiated 2027 guidance of $1.78–$1.82, while stating it will re-evaluate its long-term earnings growth rate and its 2028–2030 capital investment and rate base outlooks as part of the review.

Chief executive Patti Poppe said California’s wildfire liability framework “continues to create financing risks that drive higher costs.” The context: California’s 2026 legislative session closed on August 31, 2026 with a wildfire bill that omitted the liability protections utility investors had expected, after which PG&E fell about 18% and Edison International about 23%. PG&E serves 16 million people across a 70,000-square-mile territory and carries a market value near $31 billion.

Opportunity

The intuitive reading of open-ended wildfire liability is that preventing ignitions becomes existential, so burying lines and replacing conductors stops being discretionary, hardening budgets rise, and the contractors who do that work get paid more for longer. PG&E has just done the opposite, and the mechanism explains why.

A regulated utility earns a return on its rate base; rate base grows by spending; spending is funded by borrowing; and borrowing has become expensive precisely because of the liability the legislature declined to cap. Faced with that, the rational move is not to spend more on safety — it is to spend less on everything, because capital cannot be raised on terms a regulator will let you charge customers for.

Hypothesis: the consensus response to the August 31 outcome ran one step — utilities damaged, contractors helped — and stopped, when the second step is that a utility denied affordable capital shrinks its programme and takes the contractor revenue with it. If that is right, the mispricing is not in the utility, which fell violently and publicly, but in names still carrying a California grid-hardening growth assumption nobody has revisited.

A separate and larger inference: a board reviewing “how the Company is organized and financed” at a business of this size is describing structural change, and structural change at a utility serving 16 million people is a rare and asymmetric event in either direction.

How it could play out

The deferral removes work from contractor backlogs in the single largest state market for transmission and distribution services. The 2028–2030 rate base outlook is formally reopened, so the deferral may be the first instalment rather than the whole of it. Meanwhile the physical system ages slightly faster than planned, which raises ignition risk, which raises liability, which raises the cost of capital again. That loop is the argument for eventual political intervention — and the strategic review is the company applying pressure toward exactly that, in public, with a committee and a timetable.

Questions worth asking

• Which listed contractors have the largest share of revenue tied to PG&E and Edison International specifically, and is that disclosed anywhere at segment level? This is the single question that decides the idea — without it the thesis is directionally right and unquantifiable.
• What does “how the Company is organized and financed” actually contemplate — separating the gas utility, ring-fencing legacy wildfire claims, a holding-company restructuring, or sale of assets to public power?
• The 2027 earnings guidance implies roughly 9% growth on a shrinking capital plan. If that comes from cost reduction and lower financing expense rather than rate base, the earnings algorithm of the business has quietly changed.
• Does Edison International follow with its own capital reduction? Two utilities retrenching simultaneously is a different order of magnitude for the supply chain than one.
• California still has to serve rising electricity demand. If both large investor-owned utilities are pulling back, who builds the system — and does that push large customers toward generating their own power?

Where to look

• Quanta Services, MYR Group, Primoris Services — transmission and distribution contractors whose California work is the thing being deferred
• PG&E — the equity is now a wager on what the review concludes, the bonds a different wager entirely
• Edison International — the same liability exposure, no capital reduction announced yet
• Sempra — California utility exposure in a different service territory
• suppliers of on-site generation, if utility-served load growth in the state stalls

Thesis check

The strength is that this is arithmetic rather than narrative: PG&E has disclosed a dated, quantified reduction in the spending that produces both its own earnings and its suppliers’ revenue, and rate base is the entire mechanism by which a regulated utility grows. The weakness is scale — $2 billion of deferral is meaningful to PG&E and small against the national revenue of a contractor like Quanta Services, so demonstrating real damage requires California-specific disclosure that these companies may simply not provide. The strategic review, meanwhile, could run for a year and conclude with nothing.

Timing

Announced Sept 2; updates at quarterly results

Sources

PG&E Corporation, Sept 2 2026 · Investing.com, Sept 2 2026 · 24/7 Wall St, Aug 31 2026

On the final night of California’s 2026 session, lawmakers passed AB 2383 — the Fair Share in Energy Act, authored by Assemblymember Rick Chavez Zbur — and SB 886, authored by Senator Steve Padilla, and sent both to Governor Newsom, who has until the end of September 2026 to sign or veto them.

AB 2383 requires electrical corporations, community choice aggregators and electric service providers to adopt separate generation and transmission tariffs for new large-load customers taking service on or after January 1, 2027, and is drafted to take effect only if SB 886 also becomes law. The measures direct the California Public Utilities Commission to create distinct rates and updated interconnection rules for data centres so that other customers do not fund the generation and grid upgrades built to serve them.

This is not isolated: more than 300 data-centre bills were introduced across 30 state legislatures in the first six weeks of 2026, and at least 18 states have introduced bills creating special rate classes for very large users.

At federal level, the Ratepayer Protection Act (H.R. 9340), sponsored jointly by Representatives Gabe Evans and Kathy Castor, cleared the House Energy and Commerce Committee by 52 votes to nil; it directs state utility commissions to consider requiring data-centre loads of 100 megawatts and above to cover the full cost of the grid upgrades needed to serve them, and it awaits a floor vote. Separately, on September 2, 2026 FuelCell Energy reported its first data-centre power agreement, disclosing that a customer option for up to 350 megawatts of on-site fuel cells added roughly $2.4 billion to its awarded capacity backlog.

Opportunity

Read as a cost story this is trivial: a somewhat higher power price is immaterial against the capital cost of an AI data centre, and the operators would pay it without blinking. That framing misses what the binding constraint actually is.

What limits a data centre is not the price of electricity but the time it takes to get connected — interconnection queues in the major markets run for years, and every one of these measures adds a further layer of tariff design, cost-allocation proceedings and commission process on top of that wait. Hypothesis: the effect of the ratepayer-protection wave is therefore less financial than architectural.

It makes the regulated grid slower and more politically contingent at precisely the moment a customer needs certainty, and the rational response is to stop asking permission — to put generation behind the meter and treat the utility connection as backup.

If that is what happens, the beneficiaries are not the utilities and not the hyperscalers but the people who sell megawatts you can install on your own land in eighteen months, and the timing is set by a signing deadline and a January 1, 2027 effective date rather than by anything in the technology cycle.

How it could play out

Large-load tariffs become standard across a dozen or more states within a year, each with its own commission proceeding and its own timetable. Developers stop treating grid interconnection as the default and start pricing on-site generation into the base case. Orders concentrate on whatever can actually be delivered quickly — reciprocating gas engines, aeroderivative turbines, fuel cells — while large frame turbines remain sold out for years.

Utilities, counter-intuitively, end up in a better position: a large customer contractually obliged to pay the full cost of its own upgrades is a safer counterparty than one whose costs a commission may later disallow. And states that decline to legislate market themselves on speed instead of price.

Questions worth asking

• Who can actually supply baseload power at data-centre scale on a two-year timeline, and which of them has uncommitted manufacturing capacity? This decides the idea. A large gas turbine ordered now arrives near the end of the decade, so the winner is whoever is not in that queue.
• Does AB 2383’s January 1, 2027 trigger create a visible rush to energise before the deadline, and would that show up in California interconnection filings and utility large-load pipelines?
• Is full cost responsibility actually bullish for the utilities? If a large-load tariff removes the risk that a commission later disallows recovery, these laws hand utilities cleaner, safer growth than they had before — the opposite of how the headlines read.
• Which states are moving the other way and openly courting the load, and has land, water rights or generation capacity there repriced yet?
• If the bills are vetoed, does the issue die or move to the commission and the ballot? A veto may delay the cost shift without removing it.

Where to look

• FuelCell Energy and Bloom Energy — on-site generation sold explicitly as a way around the interconnection queue
• Caterpillar and Cummins — reciprocating gas gensets, the unglamorous option that can actually be delivered
• GE Vernova and Mitsubishi Heavy Industries — the turbine suppliers, whose problem is that their slots are already sold
• American Electric Power, Dominion Energy and DTE Energy — utilities whose large-load tariffs would be validated rather than threatened by these rules
• Vistra and Constellation Energy — merchant generators able to contract directly with a data centre
• Digital Realty and Equinix, as the operators who have to solve for power either way

Thesis check

The strength is the sheer simultaneity and the absence of partisan resistance — a 52-to-nil committee vote and 300 state bills describe a settled political direction, not a contested one, and California supplies a dated decision point at the end of September 2026.

The weakness is that the escape route is constrained by the same physics as the problem: fuel cells, engines and turbines all have finite production capacity and multi-year queues of their own, so “they will just generate their own power” may describe a 2029 outcome rather than a 2027 one. Bloom Energy in particular has already had a large move on this theme, which is a reason to look at the less obvious suppliers rather than the famous one.

Timing

California signing deadline end of Sept 2026

Sources

CalMatters Digital Democracy, AB 2383 · FOX 11 Los Angeles, Sept 2026 · Daily Energy Insider, 2026 · ArentFox Schiff, 2026 · FuelCell Energy, Sept 2 2026

A Section 232 proclamation issued in April 2026 imposed a 100% duty on imported patented pharmaceuticals and the active ingredients that go into them, structured in tiers. The default rate is 100%. A company with a Commerce Department-approved onshoring plan pays 20%. A company that signs both an onshoring agreement with Commerce and a most-favoured-nation pricing agreement with the Department of Health and Human Services pays nothing at all, through January 20, 2029.

The 100% rate took effect on July 31, 2026 for the seventeen large firms named in the proclamation’s Annex III, and applies to all other companies from September 29, 2026 — a staggered window of 120 days for the larger firms and 180 days for the smaller ones. Generic drugs and biosimilars are excluded for now, though the proclamation directs Commerce to review within one year whether to extend the tariff to them.

On August 31, 2026 the administration announced that nine further manufacturers — Alcon, Astellas Pharma, BeOne Medicines, BridgeBio, CSL, Kyowa Kirin, Sun Pharma, Teva Pharmaceuticals and UCB — had agreed to most-favoured-nation pricing for Medicaid, bringing the total number of companies inside the programme to twenty-six. Announced US manufacturing commitments across the industry now exceed $480 billion across 22 sites.

Opportunity

Almost all the coverage treats this as a drug-pricing story and asks whether patients save money. Read structurally it is something else. A company paying zero while a direct competitor pays one hundred percent on the same imported molecule is holding one of the largest cost advantages ever created by a single administrative act, and it is contractually fixed until January 2029.

Hypothesis: the market appears to be scoring these agreements as concessions — a company giving up price to Medicaid — when for a firm with modest Medicaid exposure and heavy import exposure the arithmetic may run strongly the other way, making the deal closer to cheaply purchased tariff immunity than to a price cut. Nobody seems to have worked that trade-off company by company.

The second and less examined side is the losers: on September 29 the full rate simply arrives for every remaining importer of patented product, and the list of who is not inside the programme has not been assembled anywhere I can find. That absence is itself the opportunity, because it is a finite, researchable list with a hard date attached.

How it could play out

The deadline passes and a set of mid-sized importers discover their US gross margin has been halved on products they cannot reprice mid-contract. Some capitulate into agreements on worse terms; some exit US launches; some become acquisition targets for companies that already hold exemptions. Meanwhile the signers must actually build the plants they promised, which means real orders eventually landing on bioprocessing, cleanroom and fill-finish suppliers — and if Commerce extends the tariff to generics inside the review year, the same sorting happens again to a much larger and far more fragile supply chain.

Questions worth asking

• Who is not on the list? Building the roster of listed mid-cap and specialty pharmaceutical companies that import patented product into the United States and hold neither a pricing nor an onshoring agreement is the piece of work that decides this idea.
• What does a most-favoured-nation Medicaid price actually cost a given company, set against a 100% tariff avoided? For some the exchange is plainly favourable and the disclosure to test it exists in their filings.
• The exemption runs to January 20, 2029 — the end of a presidential term. Have the signers bought protection that lasts exactly as long as the administration that granted it, and is that priced into anything?
• If Commerce extends the tariff to generics within its review year, which companies are structurally unable to survive it, and does that make the generic supply chain a policy short rather than a value long?
• Where the tariff does bite, who absorbs the cost? Hospitals and health systems buying physician-administered drugs cannot reprice inside a contract year, and that exposure sits well outside the pharmaceutical sector.

Where to look

• Teva Pharmaceuticals, Sun Pharma, CSL, UCB, Astellas Pharma, Alcon, BridgeBio, Kyowa Kirin and BeOne Medicines — the newly signed group, each now holding a tariff position their unsigned peers do not
• Eli Lilly, Pfizer and Johnson & Johnson — already inside the regime and furthest along on American plants
• Thermo Fisher Scientific, Danaher, Sartorius, Repligen and West Pharmaceutical Services — the equipment, consumables and containment suppliers that any genuine onshoring has to buy from
• hospital operators and drug distributors, on the receiving end of whatever cost is not exempted

Thesis check

The strength is that the mechanism is statutory, dated and extreme — a hundred-point tariff differential between two competitors selling the same class of product has to show up in somebody’s margins. The real constraints are three: this administration’s tariff deadlines have a poor record of arriving unamended, generics and biosimilars are excluded so the affected revenue pool is far narrower than the headline rate suggests, and the companies most brutally exposed may turn out to be private or listed abroad, leaving a correct thesis with nothing clean to express it through.

Timing

Full rate applies to remaining firms Sept 29 2026

Sources

White House proclamation, Apr 2026 · Crowell & Moring, 2026 · White House fact sheet, Aug 31 2026 · STAT, Sept 1 2026