Powerstack: Blackstone sets the standard for data center exits

Hi there,

Next Tuesday, we're hosting a briefing with our friends at Taylor Wessing on powering the data center buildout. We'll highlight our latest research on the risks and bottlenecks in the pipeline, and strategies companies are using to get their projects built...ahead of everyone else.

There are a few spaces still available, so if you're in London and this is your world, register here.

This is Sightline’s weekly newsletter on the moves and motives shaping the load growth era. Not a client yet?

Blackstone filed to IPO its REIT for data centers, revealing a new acquisition benchmark for stabilized data centers -- $15M/MW, investment-grade tenant, 10-20 year NNN lease. It gives developers a price to work backwards from and lenders a valuation to underwrite against. The 60GW of pipeline capacity that doesn't meet this standard, like Fermi America, now faces a stark choice: land a hyperscaler tenant or risk being stranded.

What happened

Two developments this week showed the difference between a stranded data center, and a financeable one.

Fermi America gives us the example of the stranded one. The co-founder and CEO of the high-profile developer departed effective immediately this week, the latest break for the company. Fermi's pitch was speed and scale. It aimed to have 1GW online by end-2026 and $1.5bn in revenue per gigawatt, ultimately landing at 17GW on 5,000 acres in the Texas Panhandle. Execution has missed every mark. A tenant pulled out in December, the company still has no publicly confirmed anchor hyperscaler, and the CFO said there will be no further constructionon its massive data center until a definitive tenant agreement and financing are in place. The stock sits more than 80% below its October peak.

The lesson isn't that power-first was wrong, just that power without a tenant is an expensive hole in the ground.

Blackstone, also this week, defined the standard of what’s acquirable, and by extension, financeable. ‘Derisked’ has been a subjective term in data center finance, where everyone seems to have a definition. Blackstone Digital Infrastructure Trust (BXDC) filed an S-11 to IPO a public REIT that will buy a specific kind of data center – not too big, not too small, but solid and stable. The buy box looks like an industry spec checklist, defining every buyable criterion:

BXDC also has priority acquisition rights over qualifying assets across Blackstone's platform for 24 months post-IPO, with $25bn in near-term opportunities already under review. For developers, the practical implication is that "derisked" now has a definition with a number and signature attached to it.

Shown in the chart above, Sightline tracks 710 data center projects in the pipeline. Through 2030, developers have announced 102.3GW of capacity, but only 40.8GW qualifies as derisked (Sightline clients can access the analysis and data here). The 40.8GW is BXDC's addressable market today. The remaining 60GW is who, through the S-11, Blackstone is telling to shape up.

Mark's take

BXDC is the missing piece.

In Blackstone's data center stack, it had development via QTS and AirTrunk (in APAC), a debt vehicle with $10bn+ deployed, and an opportunistic equity fund that invested up and down the value chain. But until now with BXDC, it had nowhere to park the stabilized, income-generating assets at public-market cost of capital. QTS develops (and aims to limit power exposure and go grid-connected) leases to a hyperscaler, stabilizes, and sells the asset on to BXDC. Capital recycles into the next build. And with that, wipes hands, Blackstone’s vertical integration across the data center stack is complete.

BXDC isn’t limited to just buying QTS assets. Obviously. With the S-11, it defines the types of assets it’ll seek, and really what derisked means in data center finance. It set the standard. We, of course, were quite pleased to see the criteria (see above) map almost 1:1 to our own credibility methodology. The difference is that Blackstone introduces one more layer of transparency to it – it puts numbers to the expected return (5.75–7.00%), and the acquisition price ($15m/MW). Now, developers have an exit price they can work backwards from, and lenders have a valuation they can finance against. So when we look at Fermi, it never really had a locked-in IG hyperscaler tenant, so it didn’t quite make it, and would get filtered out of BXDC’s screen. So if you’re a developer hoping to exit into the Blackstone-iverse, the tenant matters just as much as the power.

Stay calm, stay positive, never give up...and win, baby, win.

Who this helps:

Who should be nervous: 

A quick read on the numbers shaping the market. The capex, the contracts, the regs, all anchored in the so-what.

1GW // Meta's capacity reservation for Noon Energy’s multi-day storage tech. An initial 25MW/2,500MWh demonstration phase is planned for 2028 and would be a big scale-up from Noon’s pilots. Still unclear: whether the demonstration will use Noon’s core tech (reversible carbon dioxide fuel cell) or its more mature but less energy-dense hydrogen-based variant. Even though Meta hasn’t committed to buying the full GW yet, hyperscalers have become LDES market-makers as Crusoe (Form) and Meta (Noon) follow Google’s lead (Form, Energy Dome).  

$691m // Japan's subsidy commitment for next-gen geothermal to cover up to two-thirds of drilling costs through 2030. Japan sits on the third-largest geothermal resource on earth but has barely touched it, relatively speaking, because the best sites are under onsens, and public opposition has blocked development for decades. Next-gen geothermal doesn't need hydrothermal reservoirs, so now, with high power prices, an energy security approach, and a newly pro-geothermal government, Japan could be the most exciting market outside the US.

$380m // Blue Energy's project finance raise to fund its first modular nuclear-plus-gas power plants. Blue Energy's thesis is that nuclear's problem isn't the technology, it's the massive cost overruns and delays. Its workaround is to separate the plant into components and de-risk; first, build the conventional balance-of-plant with gas, then drop in an SMR reactor later, pre-fabricated offsite at a shipyard where nuclear-on-site risks don't apply. The bet is that each piece is bankable on its own terms, so no single overrun can sink the whole project.

13 // Previously-canceled grid projects that DOE has quietly reinstated, as part of a list of nearly 2,000 Biden-era awards the agency told Congress it plans to "retain or modify." They are mostly transmission and distribution projects funded under the Bipartisan Infrastructure Law that had been killed, apparently for political geography reasons, and are now back. But implementation remains a question, after staff cuts and payout delays.  

5GW // The compute capacity Anthropic committed to consume from Amazon as part of a deepened partnership that could reach $25bn in investment and $100bn in AWS cloud spend. As Anthropic grows, it’s been visibly constrained by compute scarcity, and this deal is effectively Anthropic outsourcing its power problem to Amazon. And Amazon’s power strategy seems to be working - it’s the world's largest corporate renewables buyer, holds nuclear PPAs, funds utility grid upgrades directly, and simply outspends everyone else in the interconnection queue (with $200bn in planned capex in 2026).

Explore more Signals on Sightline here.

The policies, rulings, and company moves worth watching.

IEEPA tariff refunds are now open for claims. Tariffs embedded in equipment procurement costs may have inflated tax basis, ITC calculations, rate base filings, and tax equity structures -- meaning refunds could trigger downstream adjustments across project finance and regulatory positions. The importer of record is the only party that can file directly, so developers and utilities need to audit EPC and supply contracts now to determine whether tariff costs flow back to the project owner. 

FERC’s June deadline to act on DOE's data center interconnection reform, deciding whether FERC sets a federal floor or deferring to state-by-state frameworks. The commission punted past DOE's original April 30 deadline, citing the fast-moving regulatory landscape -- PJM, SPP, ComEd, and Tri-State have all had proposals decided since October. It’ll determine interconnection cost exposure for every large load developer in an RTO footprint.

Texas PUC weighing new interconnection rules for large loads. The state's grid operator just forecast demand could quadruple to 367.8GW by 2032 -- up from an 85.5GW record in 2023, and regulators are now mulling new requirements for data centers and other large users seeking faster grid access, with guardrails to protect ratepayers and reliability. Texas is essentially stress-testing the same tensions playing out at FERC and in PJM.

The latest research, features, and data drops on the Sightline Climate platform.

Sightline Climate's latest research answers the question data center procurement teams keep asking us: "how can I get new clean megawatts online the fastest?" This report benchmarks four pathways by timeline and risk using project-level interconnection data: buying projects from the ERCOT queue, repowering retiring coal/gas sites, building greenfield, and going off-grid. The report comes with a filterable dataset of 544 ERCOT queue projects and 76 brownfield sites. Clients can read it here.

Where the market is meeting, and where to find us

📅 Powering the data center buildout // London // April 28, 9-10AM // We're hosting a briefing with our friends at Taylor Wessing on powering the data center buildout. We'll highlight our latest research on the risks and bottlenecks in the pipeline, and strategies companies are using to get their projects built, ahead of everyone else.

📅 Innovation Zero World Congress // London, 28-29 April, 2026 // The UK's leading low-carbon innovation congress to accelerate the transition to a net-zero economy. Let us know if you'll be in town to join our team on site.

📅 Boston Climate Week // Boston, MA, 3-10 May, 2026 // Boston's inaugural climate action week, featuring academic and research-driven climate solutions from New England's innovation ecosystem. Let us know if you'll be in town to join our team on site.

Attending an event? Connect with our team.

Interested in diving deeper? Talk to our team and leverage the tactical intelligence that hundreds of energy and investment decision-makers like Southern Company, Tokyo Gas, Jefferies, Galvanize, B Capital and others use to stay ahead in the energy transition.