---

Three signals this week share a structural thread: the gap between announced energy capacity and physical energy capacity is widening across every fuel type and every asset class. Data centers that exist on paper but never ordered transformers. Clean energy projects sprinting toward a tax deadline that, once it passes, removes the economics that justified them. Oil inventories that look healthy until you trace the pipeline feeding them. In each case, the market is learning to distinguish signal from substance.

- Oil in transit collapsed while stockpiles climbed — framing OECD inventories as a countdown rather than a buffer, with IEA's Birol warning April arrivals drop to zero. Read →

- $650 billion in AI spend, half the 2026 data centers won't get built — distinguishing phantom load from real load as utilities force developers to put skin in the game. Read →

- £1 billion saved, billions paid, same month, opposite grids — UK wind and solar displaced 21 TWh of gas imports while US households absorbed added fuel costs. Read →

- US clean energy buildout faces a policy cliff — the OBBBA tax deadline is pulling capacity forward, then a policy cliff replaces the demand cliff. Read →

- Renewables passed gas in March 2026 — Ember data confirmed the crossover while IRA credits are being clawed back and federal permitting slows. Read →

Each signal below traces practitioner debate and audience movement on the week's most-debated posts: what got challenged, who composed the audience, and what that pattern indicates.

Phantom Data Centers: The 2026 Bottleneck Is Not Demand. It's Whether the Transformers Were Ever Ordered.

Sightline Climate put hard math on the AI infrastructure gap. Of the 16 gigawatts of US data center capacity slated for 2026, only about 5 GW is under construction. Thirty to fifty percent will miss the year, delayed or cancelled outright. The post reframed this not as a cooling AI market but as the moment physical reality caught up with phantom load. For two years, developers filed identical interconnection requests across multiple utilities to hedge power availability, then walked away from whichever didn't clear fastest. AEP Ohio cut its pipeline from 30 GW to 13 after a new tariff forced developers to commit to 85 percent of requested capacity whether they use it or not. More than half the queue vanished overnight. PG&E revised 400 MW out. Exelon says only 22 percent of its 65 GW 2040 pipeline is real. The free optionality era is ending.

The real bottleneck, though, is boring: transformers, switchgear, and batteries all carry 18-to-24-month lead times on equipment nobody was ordering in 2023 when the press releases started flying. The companies bringing real capacity online in 2026 are not the ones with the biggest announcements. They are the ones who locked equipment contracts 18 months ago.

The practitioner response confirmed the thesis with operational specifics. The CEO of a consulting firm endorsed the post's framing directly: "We're seeing clients who secured transformer orders in late 2023 getting delivered on schedule while those who waited are looking at mid-2027 at best. The phantom capacity wash-out was inevitable once utilities started requiring skin in the game." A founder advising power generation operators endorsed the conclusion from a different angle, arguing that modular deployment is the structural answer: deploy where power is already available, start generating value faster, scale as additional power comes online, and retain the flexibility to redeploy if site economics change. Neither commenter disputed the core thesis. Both added specificity that sharpens it.

Send activity ran at 5.88x the 90-day average rate. CXO and VP-level readership was 24.0%, against a baseline of 12.74%. IT Services and IT Consulting viewers composed 4.0% of the audience against a topic baseline of 1.22% for data center content, a 3.28x shift. Construction viewers composed 4.0% against a topic baseline of 0.0%.

> If AEP Ohio's 85-percent-commitment tariff model spreads to PJM-wide adoption within 12 months, does that accelerate real capacity by clearing phantom load, or does it freeze mid-tier developers out of the queue entirely and concentrate buildout power among three to four hyperscalers?

Read Jamie's full post →

Reply if any of this is playing out at your company, or contradicting what you're seeing on the ground. Every reply goes directly to Jamie and feeds our intelligence.

The Clean Energy Tax Cliff: 70 GW Pulled Forward, Then the Floor Drops Out

Every developer in America is sprinting to beat July 4, 2026. That is the OBBBA deadline: begin construction before that date, or place in service before December 31, 2027, and you still qualify for the 45Y and 48E tax credits. Miss it, and wind and solar lose the economics that made them work. So the market is pulling forward everything it can. Seventy gigawatts of new solar in 2026 and 2027 alone. ERCOT battery storage tripling from 15 GW to 37 GW by end of 2027. Then look at 2029 and 2030: 19,000 MW, then 12,000. That is an 87 percent collapse, and it is a policy cliff, not a demand cliff.

Send activity ran at 3.76x the 90-day average rate, with save activity at 2.47x the 90-day average rate. Utilities viewers composed 9.0% of the audience against a topic baseline of 5.76% for battery storage content. CXO and VP-level readership was 17.0%, against a baseline of 12.74%. The pushback from operators was persistent but ran tangential to the core tax-credit thesis. A founder advising energy sector operators challenged the framing on multiple fronts, arguing that battery costs added to renewables will make electricity more expensive for end users once tax credits expire, and separately questioning whether the grid's purpose survives if the only cost-avoidance strategy is going off-grid entirely.

Read Jamie's full post →

The practitioner pushback, 1 additional signal, and Field Notes from oil and gas and clean energy continue below for paid subscribers.