This Boring Pipeline Stock Just Signed a Deal to Power AI Data Centers

ONEOK signed a deal to supply gas to a 1-gigawatt power plant serving data centers, positioning it to win additional AI-power related contracts. The $100 million project is small versus its $2.7B–$3.2B 2024 capex guidance, but management cited a “very nice return” and hinted at “a couple of other opportunities.” Separately, net income rose 13% in Q2, and ONEOK boosted full-year net income guidance to $3.6B at the midpoint, supported by strong volumes and ongoing growth projects—potentially extending long-term expansion capacity through H1 2029.

Analysis

The key mechanism is not incremental EBITDA from a single project; it is proof that midstream can monetize AI load growth with low capex and fast payback. That matters because the winning bottleneck in this cycle is speed-to-power, and molecules can be deployed faster than wires, which shifts some economics away from utilities and toward localized gas logistics. If this becomes a repeatable sourcing channel, OKE can earn a higher growth multiple without needing a step-up in leverage.

Second-order beneficiaries are upstream gas names and pipes with dense regional footprints that can supply firm molecules into power projects; the overlooked loser is the set of utilities and renewable developers whose load-growth cases depend on slower grid interconnects. The first deal is economically immaterial to OKE’s near-term EPS, but repeated wins could justify selective lateral/storage builds over 12-24 months and create a cleaner organic growth narrative than the market typically assigns to pipelines. The market may be underestimating how much of the AI capex cycle is really a utility-infrastructure arbitrage.

Contrarian view: this is still mostly an option on follow-on announcements, not a fundamental inflection yet. Near term, sentiment can outrun cash flow, so the trade only works if management converts pipeline discussions into booked backlog over the next 1-3 quarters. Falsifiers are simple: no additional AI-related wins by the next two earnings prints, a faster-than-expected grid solution, or gas basis/power economics that make behind-the-meter generation less attractive. If those fail to materialize, this is just a nice narrative overlay on an already solid but not re-rating-worthy business.

Key Decisions for Investors

• Long OKE on pullbacks over the next 1-3 weeks; treat this as a 6-18 month multiple-support story, not a one-day earnings trade. Risk/reward improves if management converts the 'couple of other opportunities' into disclosed backlog before the next two quarterly calls.
• Pair trade: long OKE / short XLU for 3-6 months to express the view that AI load growth will monetize through molecules faster than regulated grid expansion. Falsify if utilities start announcing a material cadence of faster-to-service AI load contracts.
• Watch EQT and CNX as second-order beneficiaries; if data-center gas demand broadens, upstream gas names should see tighter regional pricing and better basis. Add only after confirmation from multiple midstream announcements, not on this headline alone.
• No options chase here; implied volatility is unlikely to reward a one-off contract. If you want convexity, consider a small starter long in OKE and add only on evidence of recurring deal flow at the next earnings update.