In this piece:

The AI infrastructure buildout is increasing demand for electricity, water, and land as large U.S. technology companies—including Amazon, Microsoft, Alphabet, and Meta—are expanding the computing infrastructure that supports cloud and AI services.

Community opposition, however, is already affecting development. Approximately $130 billion in proposed U.S. data center projects were blocked or delayed in the first quarter of 2026, compared with $156 billion for all of 2025.1

Meanwhile, the environmental and community implications extend across the AI value chain, from data center operators to chip manufacturers and providers of semiconductor packaging and equipment.

For investors, these pressures raise questions about operating costs, construction timelines, capital requirements, and the durability of expected returns.

Impacts Across the AI Value Chain

The AI buildout is highly resource-intensive, so our engagement examines how companies across the AI value chain factor resource constraints and community obligations into investment decisions and whether their plans remain viable if demand or financial conditions shift.

Our engagement with five companies—Microsoft, SK hynix, Samsung Electronics, TSMC, and Delta Electronics—focuses on the following areas:

Community Relations.

Host communities must decide whether to commit water, infrastructure, and public resources to projects with uncertain demand and returns. We advocate early consultation and transparent, enforceable agreements that define corporate obligations, name creditworthy counterparties, and protect communities if projects underperform or ownership changes occur.

Water.

Site-level assessment, disclosure, and management of water use, particularly for manufacturing facilities and data centers, are critical in water-stressed regions. For operations in such regions, we are requesting site-level reporting on freshwater withdrawals, consumption, and progress toward reducing withdrawals.

Product Design and Chemical Safety.

More energy-efficient chips can reduce data-center power needs, while safer production can protect waterways and downstream communities. We engage chipmakers on eliminating PFAS that are capable of entering wastewater in manufacturing. These persistent chemicals, with known toxic effects, can contaminate rivers and groundwater.

Energy and Emissions.

Expanding data centers and semiconductor facilities are increasing electricity demand, straining local grids, and undermining existing climate commitments. We ask companies to disclose their electricity procurement strategies, energy sources, responses to grid constraints, and, where applicable, power usage effectiveness (PUE), a data center efficiency metric. We also seek absolute emissions-reduction targets, interim milestones, and credible implementation plans.

Company-Level Engagement

Across the AI value chain, our engagement seeks evidence that companies can translate commitments into operating practices that address resource constraints and community obligations. Targets, disclosures, and implementation plans help us assess how companies manage risks that can affect operating costs, project timelines, and expected returns as AI buildout expands.

The global software, cloud, and AI provider remains committed to becoming carbon negative by 2030 and to matching all of its annual global electricity consumption with renewable energy. It has also launched its Community-First AI Infrastructure Plan, the first of its kind among hyperscalers. The plan ends nondisclosure agreements with local governments and seeks to terminate existing agreements worldwide.2

Our engagement evaluates whether these commitments can deliver measurable results as Microsoft expands its data center footprint. We seek clearer plans to reduce absolute water withdrawals in stressed regions, fuller PUE disclosure, the share of new data centers using behind-the-meter generation, and details on board oversight of the Community-First framework.

The semiconductor manufacturer supplies memory and storage chips for AI and other computing devices. It has set efficiency targets for select products and secured carbon-footprint certification for 15 products in 2025, but lacks a company-wide efficiency target. Scope 1 emissions rose over 60% in 2024 amid increased AI demand, highlighting the need for absolute reduction targets.3

We recommended company-wide product efficiency and absolute emissions-reduction targets, plus a timeline to eliminate PFAS. We also asked the company to assess whether its water-reuse target at the Cheongju, South Korea, manufacturing site is adequate amid increased production and water stress.

The semiconductor, smartphone, and consumer electronics manufacturer faces substantial resource demands in chipmaking. Its Device Solutions division accounts for 94.5% of company-wide Scope 1 emissions and 90.5% of water withdrawals.

We asked Samsung to adopt division-wide efficiency targets and emissions disclosure; apply interim 2030 and 2035 decarbonization targets; set site-specific water targets for its semiconductor facilities in Xi’an and Suzhou, China; and formalize a timeline to eliminate PFAS.

The semiconductor manufacturer produces chips designed by other companies, including those used in AI systems. It recently published a Responsible AI policy, discloses product carbon footprints by fabrication facility and manufacturing process, and has set an absolute emissions-reduction target aligned with the Science Based Targets initiative (SBTi).4

TSMC’s consolidated Scope 1 emissions rose by more than 20% in 2025 despite its stated zero-growth goal, and its water recycling rate has declined for three consecutive years. We have asked TSMC about its future emissions expectations and how it plans to address declining water recycling rates, particularly at its water-stressed sites in Arizona and China.

The manufacturer supplies power management and cooling systems for data centers and other industrial applications. It has SBTi-validated targets, a board-level ESG Committee, and has reported a 67% reduction in Scope 1 and 2 emissions from its 2021 baseline.

However, water withdrawals rose by nearly 18% in 2025, and workplace safety indicators worsened year over year. We have asked for interim emissions-reduction targets, a roadmap toward 100% renewable electricity, site-level water targets for its highest-risk locations, and a plan to address the deterioration in safety performance.5

Sustained Dialogue

Goodman Group’s Expanding Data Center Business

We met with Goodman Group, a developer, owner, and manager of logistics properties and data centers held in our portfolios, to discuss how it is managing the risks and resource demands of its expanding data center business.

Goodman’s customers are primarily hyperscalers—large cloud and technology companies—with their own sustainability requirements and renewable electricity procurement arrangements. Goodman described efforts to incorporate resource efficiency into building design, including rooftop solar and improvements to overall building efficiency. Securing sufficient power remains a central challenge as the business grows.

Drawing on its experience in Australia and Europe, Goodman described its approach to community consultation, including direct discussions with residents and collaboration with local governments. The company favors smaller developments on existing industrial land with established power and water infrastructure, an approach it believes can help reduce development impacts and community concerns.

Our discussion also identified areas for further work. Goodman describes its greenhouse gas emissions targets as aligned with the SBTi, while Scope 3 reporting and assessment of emissions across the full building lifecycle remain areas for development. These will be priorities for follow-up as its data center business expands, alongside how Goodman incorporates resource efficiency and community considerations into development decisions.

Responsible AI

Governance, Safeguards, and Accountability

Alongside our engagement on the environmental and social impacts of AI infrastructure, BCAM continues to engage companies on AI governance and risk management through the World Benchmarking Alliance’s (WBA) Ethical AI initiative. Recent collaborative outreach includes Alibaba and Tencent, providers of digital platforms, cloud computing, and AI services; Spotify, an audio-streaming platform; Naspers, a consumer internet investment group; TSMC, a semiconductor manufacturer; Nvidia, a designer of AI chips and computing systems; and Microsoft, a provider of software, cloud computing, and AI services.

TSMC recently published a Responsible AI policy, providing a basis for further dialogue on implementation. We also co-lead a collaborative dialogue with Microsoft on human rights due diligence and impact assessment. We are asking for KPIs and case studies that demonstrate how its AI safeguards work in practice, how it monitors emerging risks, and how it provides or supports remedies when harm occurs.

Building on Microsoft’s 2025 assessment of its most significant human rights risks and its generative AI human rights impact assessment summary, we are encouraging clearer disclosure of:

  • Implementation: Actions taken to address identified risks, associated commitments, and the teams accountable for delivery.
  • Effectiveness: How Microsoft evaluates whether those actions and safeguards are achieving their intended results over time.
  • Oversight: How the board and senior executives oversee implementation and respond to identified shortcomings.

In the coming months, we plan to expand engagement across the AI value chain, with greater attention to cybersecurity risks affecting AI infrastructure. Future updates will examine companies’ responses and the evidence they provide that governance policies and safeguards are working in practice.

1 Q1 2026: Data Center Watch Report | Data Center Watch

2 2026 Environmental Sustainability Report | Microsoft

3 SK Hynix Sustainability Report 2026 | SK Hynix

4 Responsible Artificial Intelligence Policy | TSMC

5 Delta Electronics

Published On: October 2, 2026

This piece is intended as an editorial review of a survey of risk considerations and does not constitute investment advice, a solicitation, or an offer to buy or sell any security. Past performance does not guarantee future results. All investments involve risk, including the risk of loss of principal.

Company developments described in this report occurred during the period of our engagement but should not be assumed to have resulted from Boston Common’s activities. Where we believe our engagement contributed to a development, we describe the basis for that assessment.

References to third-party research, publications, and data sources are provided for informational purposes only. Boston Common Asset Management does not warrant the accuracy or completeness of third-party information cited herein. Readers should consult original sources and qualified professional advisors before making any investment decision.