On May 6, 2026, Marshall & Stevens convened a group of senior practitioners for the 2026 Energy Forum: an unscripted, practitioner-level conversation about where the energy market is heading and what it means for the people financing, developing, advising on, and valuing assets in it.
The forum brought together five senior leaders from across the energy sector:
The discussion was moderated by John Geraghty, National Practice Leader for Energy and Infrastructure at Marshall & Stevens.
What follows is a combination of the key insights from that conversation. The full analysis, including Marshall & Stevens’ forward-looking valuation perspective, is available in the 2026 Energy Outlook Report.
The panel discussion suggested the current environment is increasing consolidation pressure among developers whose projects struggle to stand on their own economics.
With tax incentive scaffolding now stripped away for solar and wind technologies, and FEOC (Foreign Entity of Concern) compliance adding a new layer of supply chain uncertainty, the question for energy developers in 2026 is no longer whether to build. It is whether their projects are bankable without the federal support structures that garnered so much activity in those sectors.
“Developers who are well-capitalized, who have operating assets, are in a better position than developers that are not well-capitalized, who are develop-andflip shops who used to command a premium in the market.”
The bankability test has reset. Ken Malik described what that looks like in practice:
“What we’re seeing is not so much the prospect of acquiring a project, but rather getting inundated with a mix of good projects that can stand alone on their own, based on those supply-and-demand economics versus ones that were heavily reliant on safe harboring.”
The discussion also pushed back on the idea that any single technology would dominate the next phase of development. “I don’t think there will be one silver bullet that we’ve been hearing for so long,” said Siddiqui.
Developers still advancing projects appear increasingly focused on region-by-region analysis, prioritizing sites with secured interconnection, existing offtake structures, and communities that can be brought along.
And on what ultimately differentiates the ones succeeding: “Money is not going to replace expertise.”
— Ken Malik
The investment dynamics across the energy stack came into sharp focus at the forum. Active sell pressure, a segmenting buyer landscape, and a clear divergence between firms that can hold and firms that cannot.
Bobby Majumder was direct about what he is seeing in deal flow:
“With respect to renewables, we’re definitely seeing sell pressures, because the deals aren’t economical without the tax incentives.”
From the buyer’s side of the same dynamic, Ken Malik described what that creates: “From our side, absolutely, it’s definitely a buying opportunity. We’re somewhat seeing that as a buyer’s market now, quite frankly. But what makes it an issue with this is there’s so many projects that were heavily reliant on the tax incentives but not relying on true supply and demand economics.”
Capital flowing toward fossil fuel assets is centered more on consolidation and operational efficiency than large-scale new generation.
The firms still doing deals appear increasignly focused on fundamentals. Fahad Siddiqui identified the distinguishing characteristic: “It’s patient capital. It’s who can be patient here. And this will thin out the herd.” Brent Nelson reinforced it: “We said this is going to be a good time for patient capital, and it’s going to be a really tough time for people who need to recycle capital.”
On whether the AI build-out will cause reallocation or expansion of total capital committed to energy, the panel’s answer was consistent: Expansion.
“We haven’t had to build new stuff at this scale in a while,” Nelson observed. Ken Malik added nuance: “It will be both reallocation and expansion, with an edge toward expansion.”
A key challenge remains evaluating long-duration projects while tax, regulatory, and policy assumptions continue to evolve.
The panel spent a significant portion of the forum on the legal and compliance questions affecting transactions right now.
On FEOC, Bobby Majumder’s position was unambiguous:
“There are no workarounds for FEOC.”
Strategies being attempted across the market like establishing clean manufacturing lines outside flagged countries, reducing foreign subsidiary control, or shifting headquarters may not fully eliminate FEOC-related compliance risk.
Majumder cautioned: “You do not want to run the risk that you’re relying on certifications from your suppliers and have all your tax credits at risk.” The discussion also highlighted concerns that IRS audits years later could invalidate supplier representations.
The panel broadly agreed that FEOC guidance remains unsettled. Fahad Siddiqui: “I don’t think there’s any developer who has clarity. I don’t think there’s any bank who has clarity. We’re all looking at it in real time, trying to figure it out.”
On the question of whether Congress will reinstate ITC and PTC, Ken Malik’s framing was the one the full panel endorsed: “Possible, just not bankable. Treat any of this as upside optionality, not something you can underwrite in your base case.”
Majumder pointed to the grid itself as an underhyped variable in energy today. “Our grid in the United States hasn’t had any meaningful upgrades since the 1970s.”
The discussion highlighted how aging grid infrastructure increasingly intersects with transmission constraints, reliability concerns, and broader grid-security considerations that may shape long-term project and investment risk.
The forum produced a pragmatic discussion on the current policy landscape. The summary, in Brent Nelson’s words:
“If you’re going to hope for anything, hope for permit reform.”
FEOC guidance remains unsettled. The OBBBA has produced a new level ofuncertainty around incentive durability, and the discussion reflected growing caution around projects heavily dependent on long-term incentive assumptions.
The discussion repeatedly returned to the challenge of evaluating 30-year assetsin a policy environment that can shift materially over short political cycles.
Nelson offered a very useful framework for managing political risk at the forum: “It’s easy to get caught into what today’s political climate is. Gas plants built today won’t be online for five years. What does the political outlook look like then? Could be very different from today.”
The panel’s consensus was that project economics must be stress-tested across a range of policy
outcomes over a 15-to-20-year hold. The current administration’s priorities are not the permanent market signal.
The missing policy conversation, according to both Bobby Majumder and Ken Malik, is transmission. Generation investment alone may not resolve the reliability challenges until transmission infrastructure becomes a larger part of the policy discussion.
The data center power discussion produced some of the most striking exchanges of the forum, in regard to how much of the announced pipeline is real, what it actually needs in terms of power, and what infrastructure constraints nobody is currently pricing.
Brent Nelson:
“ERCOT is talking about 400 gigawatts of data center demand. That’s just not happening. Ten percent, maximum.”
The binding constraint is not enthusiasm. “There’s not the capital, there’s not the labor, there’s not
the chips.”
The panel’s estimates for what percentage of the national announced pipeline will actually come online ranged from 25% to 50%, offering a significant gap from headline numbers. For investors underwriting energy exposure to the AI build-out, the announced pipeline number may be less important than the percentage that can be realistically delivered.
A distinction that became increasingly important during the discussion: training-model and inference-model data centers are not the same energy problem. Training facilities need speed-to-power above everything else and the competitive cost of falling behind in AI development is significant.
Inference facilities, which serve queries at scale, will become progressively more costconscious, willing to shift compute geographically, and latency-tolerant as AI matures. “I think 10 years from now, we’re going to be very much more in that space,” Nelson said. Infrastructure being over-engineered for training today may be misaligned with the inference-heavy world of 2032.
Two constraints were discussed as increasingly important considerations. The first is transmission, and the structural gap between where generation is available and where data centers are located. The second is water. “Nobody is talking about cooling. Nobody is talking about water,” said Bobby Majumder. Hyperscale facilities in water-stressed geographies may face greater scrutiny around siting and water availability.
The data center demand story is real. The timeline , geography, and specific power requirements are all more complex and more constrained than the headline pipeline numbers suggest. Secured power access repeatedly emerged as a key differentiator.
