⚛️

TLN

💡
Thesis. Holds the largest single power contract in the category — an ~$18B, 1,920 MW nuclear PPA with AWS running to 2042 — and raised 2026, 2027 and 2028 guidance. But the entire thesis rests on one customer and one plant.
Reviewed 19 Sep 2026, after Q2 2026 results. Next review after Q3 2026 earnings on 6 Nov 2026.

What it does

Talen is an independent power producer with roughly 10.7 GW of generation, including the 2.5 GW Susquehanna nuclear plant in Pennsylvania and about 6.3 GW of dispatchable capacity. CEO Mark "Mac" McFarland, HQ Houston.
One structural detail explains most of this page. The AWS arrangement was originally behind-the-meter — the datacentre drawing power directly from the plant, bypassing the grid. FERC rejected that structure. Talen restructured it as front-of-the-meter, delivering the same power across the grid through PPL Electric Utilities instead, which removed the regulatory obstacle entirely and is why this contract exists while co-location rules remain unresolved elsewhere.
Talen did not win by owning more nuclear than its peers. It won by finding a contract structure the regulator would accept, and then signing the largest one in the category.

Bull case

  • The AWS agreement is the largest single power deal in this category: ~$18B of expected lifetime revenue. A 1,920 MW PPA through 2042, with 2% annual price escalators from 2028. At full contract quantity it is expected to generate up to $1.4B a year
  • Q2 2026 revenue $747M. Adjusted EBITDA rose $284M YoY to $374M; adjusted free cash flow $212M
  • FY26 adjusted EBITDA guidance raised to $2.03–2.23B, with the 2027 and 2028 outlooks also increased following the Cornerstone acquisition
  • Management projects the contract drives a ~50% increase in after-tax cash flow per share versus 2026 guidance — more than $8 per share by 2030–2032, a ~20% CAGR from 2024
  • The front-of-the-meter restructuring avoided the FERC approval that had previously been rejected, removing the major regulatory obstacle
  • Exploring small modular reactors with Amazon in the Pennsylvania footprint, plus uprates adding net-new capacity to PJM
  • The contract substantially reduces merchant market risk and reliance on federal nuclear production tax credits

Bear case

  • Extreme single-customer concentration. AWS is the overwhelming counterparty. One renegotiation or deferral reshapes the entire thesis
  • Concentrated fleet. Susquehanna is the crown jewel, and any unplanned outage there has outsized impact — exposing both Talen and AWS to spot prices
  • The staged ramp to the full 1,920 MW takes years. Near-term revenue is far below the $1.4B headline, so the contract is mostly a future claim
  • Smaller scale and less diversification than
    ⚛️
    CEG
    and
    ⚛️
    VST
  • Reconfiguration required transmission upgrades and a scheduled refuelling outage in spring 2026
  • Same PJM co-location regulatory backdrop as its peers
  • The growth outlook beyond the PPA depends on uprate execution and speculative SMR development

Major customers

  • Amazon Web Services — the dominant relationship. 1,920 MW front-of-the-meter PPA through 2042, ~$18B of lifetime revenue, serving datacentre campuses adjacent to Susquehanna and across Pennsylvania. Previously a 300 MW behind-the-meter co-location arrangement
  • PPL Electric Utilities — delivers the power across the grid under the restructured framework. A delivery partner, not a customer
  • PJM wholesale market — merchant sales and capacity revenue from remaining uncontracted output
🔗
Read-through. Talen is the most concentrated expression of the hyperscaler-nuclear trade in this file: Amazon's datacentre buildout in Pennsylvania is effectively Talen's revenue forecast. More usefully, the AWS–Talen deal is widely regarded as having established the procurement template that
⚛️
CEG
(Microsoft, Meta) and
⚛️
VST
(Meta) subsequently followed — so Talen is both a participant in and a bellwether for the whole nuclear-PPA theme. The front-of-the-meter workaround is the specific thing peers copied, and it is why the unresolved PJM co-location question matters less here than at Constellation.

What would change the view

  1. The staged ramp toward the full 1,920 MW — the pace determines when the $1.4B of annual revenue actually arrives
  2. Susquehanna capacity factor and any unplanned outages
  3. PJM co-location ruling — ~November 2026 filing, clarity Q1–Q2 2027
  4. Progress on uprates and any firm SMR commitment with Amazon
  5. Cornerstone acquisition integration
  6. Whether a second large counterparty is added to dilute the AWS concentration — the single thing that would most change the risk profile

Update log

19 Sep 2026 — Converted to the standard template. Current through Q2; no new disclosures since. Moved the behind-the-meter to front-of-the-meter restructuring into "what it does," since it is the mechanism the rest of the category copied rather than a footnote.
Q2 2026 — Revenue $747M; adjusted EBITDA $374M (+$284M YoY); adjusted free cash flow $212M. FY26 adjusted EBITDA guidance raised to $2.03–2.23B, with 2027 and 2028 outlooks also raised after the Cornerstone close.
Ranking note — this page was originally written on incomplete data and ranked 8 of 10 in the old Energy Grid ordering. On these figures it belongs materially higher. Rank lines have been removed from converted pages; if ranking returns it should be a database property, not body text.
Probabilities: strengthen ~58% / weaken ~27%.

Research and education only — not investment advice.