Open in interactive viewer → charts, metric popovers & call review

NiSource's Microsoft Discount

A $40B seven-year plan, a third hyperscaler, and $1.75B handed back to ratepayers — and the shares are still 10.6% off their June high.
NI · Earnings Call · 2026-10-28

The ninety days that changed the pitch

NiSource reported $0.27 of third-quarter adjusted EPS, took year-to-date to $1.49, narrowed 2026 to the upper half of $2.02–$2.07, initiated 2027 at $2.22–$2.30, stretched the capital plan to $40 billion through 2033 and lifted the consolidated adjusted EPS CAGR to 10–11%. That is a lot of new commitment for a $21 billion regulated gas utility. The tape's response was to shrug: the stock is down 7.2% over 90 days and sits 10.6% below its June 26 peak of $49.08. Two things explain the gap. First, rates. This quarter's global keyword board is a rate-shock board — Fed rate hike at rank 3, "Fed rate increased" at 9, "rising rate environment" at 29, "hiking cycle" at 59. A levered, capex-heavy utility is precisely the wrong instrument in that tape, and management knows it: “our plan did not need to modify as a result of the dislocation we have seen over the last, call it, 2 months” — Shawn Anderson, Executive Vice President and Chief Financial Officer · 2026-10-28. Second, the AI-infrastructure complex is being de-rated — AI data center names are off 5.4% over 90 days with fifty decliners against six advancers in that basket. So the tension is clean: management is raising the long-duration growth rate at exactly the moment the market is cutting the multiple it pays for long-duration growth.

Selling capacity, buying political cover

Microsoft takes a 20-year term starting in 2028, with roughly 1.1 gigawatts added to the Genco pool — about 500MW of combustion turbines and 600MW of battery storage. That makes three of the largest technology companies in the world as center customers, and five agreements in just over a year.

Our signed agreements are now expected to deliver approximately $1.75 billion in savings to existing NIPSCO electric customers.

Lloyd Yates, President and Chief Executive Officer · 2026-10-28
That is up from $1.4 billion last quarter, roughly $140 a year for an average residential customer at full ramp, with bill credits potentially starting in January pending IURC approval. “we start to roll back $1.7 billion to our customers” — Lloyd Yates, President and Chief Executive Officer · 2026-10-28, Lloyd Yates put it. The mechanism matters more than the megawatt count. The Genco assets sit outside NIPSCO's rate base, so retail customers carry no construction risk, and the benefit lands as a credit line on the bill rather than a volumetric rate cut. That is NiSource's answer to the affordability politics that cost it a TDSIC rider in July. The broader wave is consensus — CenterPoint is pitching rate base CAGR and its Batch Zero/Batch One data-center tranches, AEP is building an infrastructure business, and data-center demand is the most repeated phrase across this week's utility reporters. What is not consensus is the ring-fence-plus-credit structure. That is the company-unique piece.

What management stopped saying

The language shift this quarter is itself the signal. Last quarter's keyword board was defensive and regulatory: "recovery mechanisms" at rank 1, "regulatory environment" at 3, "CapEx plan" at 4, plus "affordability discussion," "array of options" and "pursue recovery." This quarter's board is plan arithmetic — capital plan first, plan horizon second, ramp rate third. Even "special contract," a top-three company keyword as recently as early 2026, slid into the decliner list. The posture moved from defense to offense in one quarter. Partly earned: Ohio's first three-year forward-looking rate case under Senate Bill 103 should compress regulatory lag, and Kentucky and Virginia remain on schedule. But the Indiana dockets are unresolved, the IURC is running hearings on the August 11 derecho response, and the Microsoft contract still needs an order in the first half of 2027. On the storm, the company commissioned an independent third-party review and concedes a resiliency capital number has not been sized yet — an unquantified upside to the plan, but also an unquantified ask to regulators. And note the persistence of the Microsoft thread. Two quarters ago the answer was flat: “I'm not going to comment on that specific opportunity” — Lloyd Yates, President and Chief Executive Officer · 2026-08-05. Today Michael Luhrs will still not discuss the site — “we cannot, you know, comment on any of our customers sites. Or the size of those sites” — Michael Luhrs, Executive Vice President, Technology, Customer, and Chief Commercial Officer · 2026-10-28 — which leaves the natural question open: is this the first phase or the whole thing?

The arithmetic that has to work

Revenue of $1.29B grew 3% year over year, but operating income fell 13% to $228M and net income halved to $46M. The line that matters is capital: capex of $1.1B, up 67%, against capex at about 85% of quarterly revenue. Operating cash flow of $680M, up 37%, is what makes the whole thing financeable. Which brings the equity question. A year ago the guidance was $300M to $500M of ATM equity annually; today it is “$400 million to $600 million of maintenance equity annually in the near term” — Shawn Anderson, Executive Vice President and Chief Financial Officer · 2026-10-28, stepping to minimal levels from 2032 as fully ramped customers fund the business. On a percentage-of-funding basis the plan looks more efficient; in absolute dollars the ask grew with the plan. That is arithmetic, not spin — but it is a different sentence than "lower equity." The same applies to bills. Management cut its outlook for residential bill increases by more than 100 basis points to under 4% average annually through 2031, on 9–10% investment growth. Higher growth, lower bills, same balance sheet. That only closes if Genco revenue arrives roughly on time — and one of Amazon's two combustion turbines has slid to 2030.

It is not a delay in schedule. it is actually an optimization.

Michael Luhrs, Executive Vice President, Technology, Customer, and Chief Commercial Officer · 2026-10-28
Maybe. But the fixed capacity charges are what pay for the equity, and the ramp rate is what pays for the patience.

Bottom line

NiSource is running one of the more politically sophisticated affordability hedges in the US utility sector: bring in creditworthy hyperscalers, ring-fence their assets, hand the savings to the residential base as a visible bill credit, and use that cover to grow the FFO to debt base from 16.1% at the end of 2025 back toward a 14–16% commitment. The plan is credible; the balance sheet has a 13% downgrade threshold of cushion, and $3 billion of base-business upside plus up to 4 gigawatts of Genco pipeline sit outside guidance. The market is not paying for it yet, and the reason is not the Microsoft contract. Interest coverage of 1.1x plus 23.1x trailing earnings is an awkward combination with a Fed that is hiking. NiSource is asking investors to underwrite eight years of execution in a tape that is repricing duration down. That is the trade, and it is not yet resolved.