Why Pipeline M&A Is Suddenly an AI Power Trade
TL;DR: I count five pipeline acquisitions closing in the past few weeks, worth more than $14 billion combined, and all of them trace back to one driver: AI data centers need gas fired power, and gas …

What happened, as I read it
Five buyers moved on five pipeline and midstream targets in the span of a few weeks, and the total comes to more than $14 billion. Tulsa-based ONEOK paid $4.42 billion for Brazos Midstream's Permian Basin assets. Williams paid $5.5 billion for Momentum Midstream's Texas and Louisiana gathering and processing facilities. Western Midstream paid $1.6 billion for Brazos's Delaware Basin assets. Enbridge paid $2.55 billion for Tallgrass Energy's crude assets, and Plains All American paid $585 million for Silver Creek Midstream in Wyoming. Fortune reported all five on October 1.
I don't normally cover midstream M&A for this audience. I'm covering it because the buyers behind this wave aren't just pipeline companies chasing scale. They're chasing AI power demand, and that changes what the asset actually is.
Why does this matter to an accredited investor?
Gas-fired power plants need gas. Gas needs pipelines. Pipeline operators are expanding capacity specifically to serve new data centers: Williams overbuilt the capacity of a pipeline serving one of its Meta-affiliated power plants so it could "be an energy artery along which other projects could be developed," according to WIRED.
That's the shift behind the Fortune deals. A pipeline asset tied to a hyperscaler's long-term power contract behaves like a different credit than one tied to basin production volumes. Downside first: the question isn't whether AI power demand is real. It's whether the specific contract backing a specific asset is real, too.
Who's funding the buildout — and can you get in?
Not all of this capital is coming from the strategic buyers Fortune named. Alternative investors — infrastructure funds, sovereign wealth vehicles, private credit shops, have been involved in $20.35 billion of LNG and midstream transactions so far in 2026, more than double all of 2024, according to data provider Infralogic, as Reuters reported.
Some of the larger pieces: Sempra Infrastructure's Port Arthur LNG facility drew a $7 billion investment, Williams' power projects drew $5.34 billion, and ONEOK's financing included $9 billion in backing, per the same Reuters reporting. Stonepeak took a 40% stake in Woodside Energy's Louisiana LNG project and committed $5.7 billion toward development costs. Blackstone Credit & Insurance bought 49% of a midstream joint venture from EQT.
None of that moved through a brokerage account, and that's the point. Most of the capital described here, Stonepeak and Blackstone Credit & Insurance, moves through institutional vehicles. For an investor deploying capital from a liquidity event, this is exactly the kind of deal flow that shows up pre-packaged in a private credit fund facing its own spread pressure or an infrastructure-debt vehicle long before it reaches you directly. I wrote about what that packaging costs in the liquidity illusion inside private credit ETFs, and in more depth in Infrastructure Debt Investing: How Accredited Investors Can Fund the AI Data-Center Boom. Verify before you trust the wrapper. The fee layer is where the quiet losses happen, not the headline deal.
Before you write a check into any of this, ask the sponsor three things: What length is the underlying power contract, and with whom? Is the pipeline regulated, or merchant-exposed to basin production volumes? What happens to your capital if the project slips a year?
Deals named in the reporting
| Buyer | Target | Price | What it is |
|---|---|---|---|
| ONEOK | Brazos Midstream (Permian assets) | $4.42B | Midstream gathering/processing |
| Williams | Momentum Midstream (TX/LA) | $5.5B | Gathering and processing |
| Western Midstream | Brazos Midstream (Delaware Basin) | $1.6B | Midstream facilities |
| Enbridge | Tallgrass Energy (crude assets) | $2.55B | Crude pipeline assets |
| Plains All American | Silver Creek Midstream (WY) | $585M | Midstream facilities |
Source: Fortune, reported October 1, 2026.
What's not settled yet
Fortune also reported that the pace of these deals is uneven. Sellers are asking for prices built on inflated valuations. Buyers are underwriting more conservatively. That gap has slowed dealmaking elsewhere in energy even as the midstream and data-center-linked segment accelerates. Read that as the market still pricing how much of the AI power story is real demand versus forward promise. I'd rather flag that once, plainly, than pretend the gap doesn't exist.
Common mistakes investors make reading this story
- Treating every "AI infrastructure" deal as the same risk. A regulated pipeline with a long-term hyperscaler power contract is not the same credit as an unregulated midstream gathering asset exposed to basin production volumes.
- Assuming retail access means direct access. The capital behind this wave moves through institutional vehicles, not a brokerage account.
- Treating the wrapper as free. A single-deal SPV or an interval fund built around one of these assets carries its own fee stack and liquidity terms, on top of the risk in the underlying pipeline or plant.
- Ignoring the valuation gap Fortune flagged. A deal that doesn't close, or closes below what the seller wanted, tells you something about where the market actually stands on AI-driven energy demand.
FAQ
What major pipeline projects are tied to this AI-driven deal activity? Williams has already built out pipeline capacity for Meta's gas-fired power plants, intending it as "an energy artery along which other projects could be developed," according to WIRED. Sempra Infrastructure's Port Arthur LNG expansion is drawing billions in private capital, per Reuters.
What companies are investing in AI infrastructure? On the pipeline and power side: ONEOK, Williams, Enbridge, and Western Midstream. On the capital side: Stonepeak and Blackstone Credit & Insurance, according to the sourcing above.
How do accredited investors actually get exposure to this? Most of this deal flow runs through private infrastructure debt and credit funds, not public markets. If you're evaluating one, the sponsor and the structure matter more than the AI narrative attached to it. Ask who's underwriting the contract, not the headline.
Why is infrastructure getting less attention than AI chip stocks? Pipelines and gas plants are less glamorous than chipmakers. Kiplinger makes the same point: the infrastructure behind the AI buildout gets a fraction of the investor attention the technology itself does.
The Bottom Line
If you're deploying capital from a liquidity event into private infrastructure or private credit, ask any sponsor pitching an "AI power" angle to show you the actual contract backing the asset. A long-term hyperscaler power purchase agreement is a different investment than a bet on basin production volumes. No matter how the deck frames it, the contract is the asset, not the narrative.
That's the kind of judgment I write up every week in the free AIN briefing. Subscribe below and get the next one.
Educational content only. Not investment, tax, or legal advice. Not an offer or solicitation to buy or sell securities. Past performance does not guarantee future results. Private-market investments are illiquid and involve risk of loss, including total loss of capital. Consult qualified advisers. Angel Investors Network is not a broker-dealer or investment adviser.
Topics
Part of Guide
Looking for investors?
Browse our directory of 750+ angel investor groups, VCs, and accelerators across the United States.
About the Author
Jeff Barnes, MBAContinue Reading

Nasdaq Hits a Record High as AI Stocks Outperform

Why the 10-Year Treasury Yield at 5.23% Matters to You

Investing Fads That Didn't Pay Off (and Some That Did)

Dollar Index Holds Near Two-Month High as Fed Turns Hawkish

SEC's Fair-Value Reminder on Private Asset Valuations
