The Politics of Powering AI
The U.S. faces a significant energy infrastructure challenge to support AI data center growth, with a projected 57 gigawatt power shortfall through 2028. The bipartisan American Affordability and Jobs Act proposes permitting reforms and requires data centers to cover transmission costs, shifting from voluntary commitments to federal requirements to protect residential electricity affordability.
Summary
Ariana Salvatore, Head of Public Policy Research at Morgan Stanley, discusses the intersection of AI infrastructure demands and energy policy in the U.S. The core challenge is substantial: Morgan Stanley's projections estimate a 57 gigawatt power gap through 2028 needed to support data center operations, nearly a 50% increase from their previous 38 gigawatt estimate.
The American Affordability and Jobs Act represents a significant policy response with two main components. First, it aims to accelerate infrastructure development through federal permitting reforms, including streamlined environmental reviews and greater project certainty. However, Salvatore notes that permitting reform alone cannot address all constraints, including equipment availability, labor shortages, and grid capacity limitations.
Second, the legislation addresses cost allocation by requiring data centers to pay for associated transmission infrastructure upgrades. This represents an evolution from earlier voluntary commitments, such as the Ratepayer Protection Pledge signed by hyperscalers and utilities, which committed companies to securing their own power and paying for upgrades without burdening residential consumers.
Salvatore identifies this cost-sharing shift as notable because it appears to have bipartisan support around electricity affordability concerns, even where broader energy policy disagreements exist. The legislation could influence investment economics by making projects more attractive through faster approvals, while potentially redirecting development based on transmission cost requirements and available power solutions.
Looking forward, Salvatore highlights two policy catalysts: whether Congress can achieve permitting reform and whether lawmakers establish formal cost-sharing requirements for data centers. She emphasizes that for investors, the constraint on AI deployment may increasingly be the speed and cost of securing power infrastructure rather than technology innovation or capital availability.
About this episode
<p>Energy costs are a key election issue and AI data centers are adding to the debate. Our Head of Public Policy Research Ariana Salvatore looks at policy discussions in Washington that could determine how the U.S. powers its next technological boom.</p><p>Read more <a href="https://www.morganstanley.com/insights?cid=mg-SM_CORP-insights-17607" rel="noopener noreferrer" target="_blank">insights</a> from Morgan Stanley.</p><p><br /></p><p>----- Transcript ----- </p><p> </p><p><strong>Ariana Salvatore:</strong> Welcome to Thoughts on the Market. I'm Ariana Salvatore, Head of Public Policy Research at Morgan Stanley. </p><p>Today, I'll discuss the growing energy demands of AI and what the government can do in response. </p><p>It's Monday, October 12th at 10am in New York. </p><p>Markets and policymakers are preparing for a potentially important debate following the November elections. The future of American energy infrastructure. And at the center of that debate is an increasingly urgent question: How can the U.S. build enough power infrastructure to support the AI investment cycle – without passing the costs on to everyday electricity consumers? </p><p>Late last month, lawmakers in Congress introduced the bipartisan American Affordability and Jobs Act. That legislation represents a significant effort to reform federal permitting. But importantly, it also includes provisions that require data centers to cover the transmission costs associated with their projects. </p><p>We think that's a notable development, particularly as we think about the legislative agenda after the November elections. Of course, the path forward still remains uncertain given substantial policy differences and a limited legislative calendar after the elections. </p><p>But why might this matter for markets? </p><p>Well, first, it's worth considering the scale of the energy challenge. My colleague Stephen Byrd and his team recently updated their projections for the U.S. data center power shortfall. They now see a potential 57-gigawatt gap through 2028, and that's before accounting for new solutions that could accelerate access to power. That's up from their previous estimate of 38-gigawatts. </p><p>And that's why the permitting proposal is so important. There are two elements that I'd highlight. </p><p>First is the effort to accelerate infrastructure development. The bill would reform federal environmental reviews and provide greater certainty for projects that have already received approvals. We think that can help reduce delays and uncertainty around new energy projects, although permitting reform alone won't resolve constraints involving equipment, labor, or grid capacity. </p><p>The second element is the growing focus on who pays for the infrastructure, and here the proposal actually builds on an existing policy initiative. Earlier this year, a number of hyperscalers signed on to the Ratepayer Protection Pledge. </p><p>That's a voluntary agreement under which companies commit to securing the electricity needed for their data centers and pay for associated power infrastructure upgrades rather than passing the costs on to residential consumers. The pledge was subsequently expanded to include a broader group of utilities, developers, and state participants. </p><p>Now, what's different about the permitting reform bill is its potential shift from voluntary commitments toward federal requirements. In particular, the proposal would require data centers to cover their associated transmission costs. </p><p>That's an area where we think policymakers from both parties appear to share concerns about electricity affordability, even though they may disagree over the broader approach to energy policy.</p><p>Looking beyond the elections, we think that creates two related policy catalysts for investors to watch. </p><p>The first is whether Congress can reach an agreement on permitting reform that helps accelerate the construction of energy infrastructure. And second, whether lawmakers move to establish a more formal cost-sharing requirement for data centers. </p><p>The eventual policy mix, we think, could influence both the timing and the economics of new projects. Faster approvals could make some investments more attractive. But requiring developers to absorb additional transmission costs could also change where projects are built and which power solutions they pursue. A changing policy framework could also reinforce the importance of alternatives. </p><p>Of course, legislation is far from guaranteed. The elections will shape the political environment in which these negotiations continue, and any final agreement could differ substantially from the current proposal. </p><p>But stepping back, I think there's a broader takeaway. </p><p>For much of the AI investment cycle, the focus has been on the pace of technological innovation and capital spending. Increasingly, though, policy decisions around electricity infrastructure, permitting, and affordability could become just as important in determining how quickly that investment translates into operational capacity. </p><p>The permitting bill is an early indication of how D.C. may approach those challenges after the elections. And for investors, the central question may be less about how much companies are willing to spend on AI and more about how quickly and at what cost they can secure the power needed to support it. </p><p>Thanks for listening. If you enjoy the show, please leave us a review wherever you listen and share Thoughts on the Market with a friend or colleague today.</p>
Key Insights
- Morgan Stanley's updated projections show the U.S. data center power shortfall has grown to 57 gigawatts through 2028, a 50% increase from the previous 38 gigawatt estimate, indicating the energy challenge is accelerating faster than previously anticipated.
- The American Affordability and Jobs Act proposes shifting cost responsibility for power infrastructure from residential consumers to data center operators through federal requirements, building on voluntary commitments like the Ratepayer Protection Pledge.
- Permitting reform and cost-sharing requirements could fundamentally alter AI investment economics by making certain projects more or less attractive based on approval speed and transmission cost burdens, potentially becoming as important as technological innovation in determining AI deployment capacity.
Topics
Transcript
Welcome to Thoughts on the Market. I'm Ariana Salvatore, Head of Public Policy Research at Morgan Stanley. Today, I'll discuss the growing energy demands of AI and what the government can do in response. It's Monday, October 12th at 10 a.m. in New York. Markets and policymakers are preparing for a potentially important debate following the November elections, the future of American energy infrastructure. And at the center of that debate is an increasingly urgent question. How can the U.S. build enough power infrastructure to support the AI investment cycle without passing the costs on to everyday electricity consumers? Late last month, lawmakers in Congress introduced the bipartisan American Affordability and Jobs Act. That legislation represents a significant effort to…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Thoughts on the Market
Will High Yields Crack the Market’s Resilience?
Andrew Sheets argues that current interest rate levels are not obviously too high, supported by robust global growth, persistent inflation above targets, and a non-inverted yield curve. However, beneath aggregate market resilience, weaker credits and lower-rated companies face disproportionate stress, suggesting a quality bifurcation similar to 2005.
High Mortgage Rates and a Stuck Housing Market
Jay Backow and Jim Egan from Morgan Stanley discuss how mortgage rates have surged to 7.5% (150 basis points since February), driven by rising treasury yields and inflation concerns. This has created affordability challenges for new homebuyers and reduced housing market activity to 40-year lows, though existing homeowners with locked-in rates below 4.5% remain largely unaffected.
Will Midterms Test the AI Investment Cycle?
Morgan Stanley's Head of Public Policy Research examines how the midterm elections could affect AI infrastructure investment sentiment, arguing that while the elections may influence the pace and friction of data center development, they are unlikely to fundamentally reshape AI regulation. The key distinction is that state and local governments, rather than federal outcomes, control most of the policy levers affecting data center projects.
Japan’s Banks Enter a New Era of Opportunity
Morgan Stanley's Miya Nagasaka argues that Japan is entering a once-in-30-year investment cycle where companies are shifting from excess savings to increased capital spending, creating substantial new opportunities for banks through lending, debt capital markets, and M&A services. Japanese megabanks' return on equity could rise from 10-11% to approximately 15% over the medium term as corporate financing demand grows.
Canada’s Next Growth Phase
Morgan Stanley's Arunima Sinha argues that despite near-term headwinds around trade, population growth, and productivity, Canada's economy is positioned for medium-term acceleration through capital deepening and technological diffusion, with potential growth rising from 1.5% to 1.75-2% by the 2030s.