Navigating AI: Risks, Responsibility & Opportunity

Artificial intelligence has the potential to be one of the most transformative forces for human progress in history. It holds the promise of solving some of our most complex scientific and technical challenges while perhaps also being able to unlock creativity on a massive scale. With AI innovation moving at breakneck speed, tasks that previously took hours now take seconds. However, the computing power this technology requires is voracious and that has consequences both local and global.

While corporations and certain municipalities seek out more land, energy and water to support this technology's growth, signs of pushback to ‘progress at all costs’ are starting to percolate. From state legislatures pressing pause on AI-driven energy projects, to growing concerns about user data rights and corporate governance, to the enormous strain AI is placing on the electric grid, we are watching this space closely, with caution and cautious optimism.

Our ESG framework is built precisely for moments like this. Because we look beyond financial statements - examining environmental impact, governance quality, and social accountability - we are equipped to evaluate AI not just as an investment theme, but as a set of real-world risks and opportunities that demand engagement. In this newsletter, Longwave lays out what we are watching, why it concerns us, emerging guardrails and where we see opportunity.

WHAT WE ARE WATCHING: CONCERNS THAT DEMAND ATTENTION

Energy Demand & Grid Strain AI data centers are among the most power-hungry facilities ever built, running 24 hours a day without idle. The International Energy Agency estimates global data centers consumed approximately 415 terawatt-hours of electricity in 2024, roughly 1.5% of all global power.[i] By 2030, that figure is projected to more than double. In the United States, a Bloomberg analysis found wholesale electricity costs near major data center clusters are as much as 267% higher than five years ago, with residential utility bills in some regions already climbing as a result.[ii

Environmental Impact As the energy industry scrambles to meet demand, much of the new electricity output comes from the grid as data centers plug in. Currently the US grid is 60% fossil fuel-based, according to the US Energy Information Administration.[iii] This increased reliance on the grid means more reliance on fossil fuels. Fossil fuel energy production is more on demand than renewables, meaning the quick need for more energy is likely to be met by ramped up emissions from fossil fuels, reversing a meaningful trend of decarbonization here in the US. However, many AI companies are taking it upon themselves to self-supply their energy, turning to new technologies such as nuclear as well as large solar farms.

Maine’s AI Moratorium: A Push Back Gains Steam In early 2026, Maine became one of the first states to pass legislation halting new large-scale AI data center development pending further environmental review.[iv] The legislation cited concerns about grid capacity, water usage, and the disproportionate burden on rural ratepayers. Maine is unlikely to be the last. As communities around the country feel the real costs of AI infrastructure, higher utility bills, strained water supplies, and altered land use, we expect legislative and regulatory friction to increase. These environmental risks can also represent investment risks too. Companies that have not planned for these challenges will face delays, cost overruns, and reputational exposure.

User Data Rights & Privacy AI systems are trained on, and increasingly dependent on, vast quantities of personal data. The legal and regulatory landscape around how that data is collected, used, and protected is shifting quickly. Proposed and enacted privacy legislation at the state and federal levels is creating compliance complexity, and companies that have been casual about data governance face meaningful legal liability. One way our fund partners mitigate these risks is through scrutiny of internal data practices at AI companies. For example, our partners evaluate and report on how companies disclose data practices and whether their governance frameworks are keeping pace with the regulatory environment.

Corporate Governance: The Question of Who Is in Charge Some of the most powerful AI companies have faced serious questions about leadership accountability and board independence. For instance, the turbulence around OpenAI’s governance structure and Sam Altman’s role, including questions about oversight mechanisms, conflicts of interest, and the speed at which massive capital commitments are made, underscores a pattern that requires watching: concentrated decision-making power without adequate checks. We believe the right approach is vigilance in making sure commitments to safety, environmental stewardship, and all stakeholders are honored.

OUR LENS: WHY ESG LETS US SEE MORE

The upshot is that while we must consider the effects of AI on health and safety, as investors, we must also think about what happens if portfolio companies don’t take these risks seriously.

Traditional financial analysis might look at an AI company’s revenue growth and see opportunity. On the other hand, a framework that considers Environmental, Social and Governance (ESG) risks, might ask harder follow-up questions: What is the climate impact of greater energy usage? Who governs the technology? What happens to ratepayers when the grid strains? Are user data rights being protected? Is energy consumption being honestly disclosed? These questions are not peripheral to the investment thesis. They are central to it. Companies that cannot answer them well carry risks that do not show up on a balance sheet.

WHERE WE SEE OPPORTUNITY: COMPANIES BUILDING SOLUTIONS

Not every story in AI is one of concern. As investors, we must be open to new technologies like AI that hold the promise to re-shape the future. Luckily, there is a meaningful cohort of companies that are directly addressing the challenges AI creates. Eaton Corporation and Donaldson Company are two examples.

Eaton Corporation Eaton is an intelligent power management company whose products sit at the critical junction between the utility grid and the AI server rack. In 2025, Eaton announced a collaboration with NVIDIA to fundamentally redesign how electricity is delivered inside AI data centers, significantly reducing energy losses in transit. The company also partnered with Siemens Energy to offer modular on-site power solutions with renewables integration built in, cutting carbon dioxide emissions roughly 50% versus traditional backup generation. Eaton’s U.S. data center construction backlog has grown to nine years. It is not a company that is riding the AI wave passively. It is building the infrastructure that will determine how efficiently AI runs.[v]

Donaldson Company Donaldson is a global filtration manufacturer that operates its own business with notable discipline. In its fiscal year 2024 sustainability report, the company disclosed completing 134 energy efficiency projects across global operations, producing nearly 15,000 megawatt-hours of annual energy savings. Scope 1 and 2 greenhouse gas emissions are down 18% since 2021, and renewable energy use has increased 40% over the same period.[vi] For a company embedded in the AI hardware supply chain, that level of operational accountability reduces the carbon footprint of the products it manufactures, and signals a culture of rigor that we look for across every holding.

HOW WE NAVIGATE: ENGAGEMENT, NOT JUST ALLOCATION

Our recommended funds also own companies that could be doing more. In those cases, we would rather try to engage and have impact, than simply divest and lose our voice altogether. While there are many companies we chose not to own, divestment is usually our last resort. Managers at Calvert Research and Management and Boston Trust Walden are active shareholders who push for accountability through education of board members, shareholder resolutions, and active proxy voting. 

Calvert Research and Management has identified AI and energy as their top thematic research priority for 2026, engaging portfolio companies directly on energy sourcing, emissions disclosure, and whether AI expansion plans are paired with credible decarbonization commitments. In the past, when dialogue fell short, Calvert filed shareholder resolutions, voted against management, or collaborated with other institutional investors to amplify pressure. Companies that fail to meet Calvert’s standards are excluded from the portfolio entirely.[vii]

Boston Trust Walden voted at 199 annual meetings in 2024, opposed management on at least one item at 62% of companies, and led five shareholder resolutions, three of which were withdrawn after companies agreed to take concrete action. That kind of negotiated commitment is an often-underappreciated mechanism: it does not make headlines, but it leads to lasting change.[viii]

OUR COMMITMENT: CAREFUL, DILIGENT, EYES OPEN

One of the paradoxes of AI is that it could help us come up with solutions to the very problems it is creating. The strain on energy demand is fueling new technologies like small modular reactors. Utility companies are having to increase the speed at which they integrate new energy projects, of which 90% are renewables.[ix] 

This complexity is both a challenge and opportunity for investors. For values-oriented investors, there is further layer of analysis. You have told us that our mandate is impact and for us true impact means engagement. We look forward to continuing to engage with our clients, our fund partners and by extension the companies on the forefront of innovation. If you have any questions about how these themes affect your portfolio and financial plan, don’t hesitate to reach out.

[i] IEA Energy and AI Report (2025): https://www.iea.org/reports/energy-and-ai

[ii] Bloomberg Energy Analysis (2025): https://www.bloomberg.com/professional/insights/artificial-intelligence/ai-energy-demand-to-climb-in-2025-26-despite-efficiency-gains/

[iii] US Energy Information: https://www.eia.gov/tools/faqs/faq.php?id=427&t=3

[iv] Maine LD 2266 (2026): https://www.mainelegislature.org/LawMakerWeb/summary.asp?ID=280092516

[v] Eaton Corporation Press Release (2025): https://www.eaton.com/us/en-us/company/news-insights/news-releases/2025/eaton-accelerates-data-center-infrastrructure-in-ai-era-with-nvidia.html

[vi] Donaldson Company FY2024 Sustainability Report: https://www.donaldson.com/content/dam/donaldson/corporate-communications/annual-reports/Donaldson-FY24-Sustainability-Report.pdf?srsltid=AfmBOoq9w9qFk7y26DyiM5gBFKlbQ5QMhPHhZcrBlMY7r6WuCruoj4fo

[vii] Calvert Research and Management 2026 Outlook: https://www.calvert.com/insights/articles/2026-research-themes.html

[viii] Boston Trust Walden Proxy Season Report (2024): https://www.bostontrustwalden.com/2024-proxy/

[ix] Environment America Research and Policy Center: https://environmentamerica.org/center/updates/90-of-new-electricity-capacity-in-2024-to-date-comes-from renewables/#:~:text=Renewable energy sources accounted for,rather than adds to it.%E2%80%9D

This newsletter is for informational purposes only and does not constitute investment advice. Past performance is not indicative of future results.

Holdings referenced are based on recent fund filings and are subject to change.

This material is intended for informational/educational purposes only and should not be construed as investment advice, a solicitation, or a recommendation to buy or sell any security or investment product. Please contact your financial professional for more information specific to your situation. All indices are unmanaged, and investors cannot actually invest directly into an index. Unlike investments, indices do not incur management fees, charges, or expenses. The main risks of international investing are currency fluctuations, differences in accounting methods; foreign taxation; economic, political, or financial instability; lack of timely or reliable information; and unfavorable political or legal developments. Diversification does not assure a profit or protect against loss in declining markets, and diversification cannot guarantee that any objective or goal will be achieved. Investments are subject to risk, including the loss of principal. Some investments are not suitable for all investors, and there is no guarantee that any investing goal will be met. Past performance is no guarantee of future results. Market data and third-party research referenced herein are sourced from public market sources as of March 2026.