The U.S. public turned against the artificial intelligence industry in 2026, rapidly and decisively opposing construction of new data centres. Big Tech’s plans to vastly expand the reach of AI met fierce pushback across the country, in red and blue states alike. Hundreds of communities launched campaigns to stop new data centres. The State of New York imposed a one-year pause on data centre development, and more than 500 municipalities across the United States blocked or curbed new data-centre construction. Most of these restrictions were put in place this year.
Polling for Heatmap, a climate news platform, shows that Americans were evenly split on how they would feel about a hypothetical data centre in their community in August 2025 (43% supportive, 42% against, 15% not sure). By August 2026, a big majority had swung against such a development (61% strongly and 14% somewhat opposed). A mere 15% signalled support.
The thirst for electricity, natural gas and water by these computing behemoths has animated the backlash. But there’s a deeper layer to the resistance: the public’s profound suspicion of the artificial intelligence industry and how it governs what we see, hear and read on the internet.
All of this has created a big problem for the sustainable- and responsible-investment sector. As an industry committed to financing a sustainable future, it faces a reckoning over its investments in AI.
Since the early 2000s, funds and portfolios invested under environmental, social and governance guidelines have accumulated large tech holdings. Investment managers assumed that tech’s pay rates, positive community relations and minimal (until recently) climate problems raised no major ESG obstacles. But as tech adopted AI over the years, assumptions crumbled about its supposedly benign nature. Ballooning carbon dioxide emissions, a generation of young people addicted to social media, threats of job losses through economic transformation and questions about whether humans can completely control AI are causing the public to demand a slowdown in AI development.
This rethinking of tech comes as two of the largest AI companies – OpenAI and Anthropic – issue initial public offerings, shifting from private ownership to the public stock markets. Anthropic’s IPO anticipated this fall is expected to value the company at an eye-popping US$1.5 trillion.. Like conventional investors, ESG investors are weighing the pros and cons of committing more funds to the tech sector.
With so much money at stake and with the public increasingly questioning the very basis of AI, should the responsible-investment industry be invested in AI at all? And if so, which companies?
Anthropic as the ethical choice
Much of this debate centres on some of the newest and most powerful entrants into the industry, the so-called frontier labs. As frontier labs, OpenAI and Anthropic occupy the most sophisticated layer of the artificial intelligence industry, operating with the highest machine-learning capabilities at the very edge of AI capability. Google DeepMind, Meta AI and the SpaceX lab xAI are also large U.S. players. DeepSeek, Alibaba Qwen and Moonshot AI Kimi are major Chinese-based frontier labs.
For most sustainable investors who restrict their investments to the public markets, frontier lab investing hasn’t been possible since the major players were privately controlled. But two years ago, three U.S. philanthropic investors with a mission to invest in community-impact companies made a bold move to acquire a small stake in this fast-growing industry. The trio of funds – Ford Foundation, Omidyar Network and Nathan Cummings Foundation – felt that the mission of one of the companies – Anthropic – was well aligned with their impact-investing philosophy and represented a break from the profit-only ethos of the AI industry. In 2024, each bought a portion of 250,000 Anthropic shares at a total price of US$7.5 million as part of a larger auction of company shares. The Impact Alpha news platform reports that the price of these shares has soared 50-fold as Anthropic’s share price has exploded.
Anthropic has a “constitutional AI” code, pledging to avoid toxic or discriminatory outputs, avoiding engagement in illegal or unethical activities, and seeking to create a “helpful, honest and harmless” AI system. Anthropic is also registered as a “public-benefit corporation,” creating a legal obligation and governance mandate to balance social and financial interests.
Over the years, Anthropic has implemented policies and controls to manage potential catastrophic risks from its models. It says it has established best practices on cybersecurity and corporate governance as part of its status as a public-benefit corporation. It conducts research into its societal and economic impacts. And it says it has imposed safeguards on its models to reduce instances of terrorism, extremism, sexual abuse, child harm and election interference. These policies were tested earlier this year when Anthropic resisted a request from the U.S. Pentagon to remove guardrails preventing its models from selecting targets autonomously.
“As philanthropic leaders who manage our grants and our capital for social good, we invest in generative AI that protects, promotes and prioritizes public interest and the long-term benefit of humanity,” wrote Roy Swan, head of Ford’s mission investments, and Mike Kubzansky, then-CEO for Omidyar, in a 2024 commentary for Observer, a tech news platform. Swan and Kubzansky didn’t suggest that Anthropic is the only path to investing in responsible AI. They argued that a portfolio of private investments in small, socially oriented tech companies could also work toward responsible AI.
Anthropic’s ESG record isn’t spotless, however. Climate critics have pointed to its lack of disclosure on carbon dioxide emissions and the absence of emission-reduction targets. The company has also been dogged by copyright disputes from authors who say their work has been unlawfully used to train AI models. But Swan and Kubzansky said Anthropic’s ESG policies stand out. Anthropic’s founders “started the company with trust and safety at its core,” they wrote. (Corporate Knights reached out to the Ford Foundation to speak with Swan, but a spokesperson said he wasn’t available for an interview by the deadline.)
Anthropic’s major competitor, OpenAI, is even less consistent on ESG issues. It is governed by a public-benefit corporation under a non-profit foundation but lacks Anthropic’s comprehensive safety and governance policies. So it tends to be reactive rather than proactive on ESG issues, as, for example, when it recently released a version of ChatGPT for teenagers with protections on self-harm and sexual chats. The company, like Anthropic, also is weak on carbon dioxide emissions and targets and faces lawsuits by authors for copyright infringement.
A third company, SpaceX, which operates the xAI frontier lab as well as the company’s satellite and space programs, is indisputably the frontier lab with the lowest ESG ranking. This was confirmed by business rating company MSCI when the stock started trading in June. The ranking reflects the barrage of public controversies surrounding the company and its poor ESG governance and management under CEO Elon Musk.
A full ESG assessment will have to wait until the flood of information under public disclosures is released by Anthropic and OpenAI and fully digested by the markets. In the meantime, early indications point to Anthropic as the best choice for public market ESG investors in the frontier AI sector.
But apart from individual company performance on ESG issues, there’s a growing consensus that regulators around the world need to step in and provide greater oversight of the industry. After one of OpenAI’s models autonomously hacked into a small AI company, Hugging Face, this summer with a swarm cyberattack, Anthropic also disclosed that it has been responsible for similar less serious incidents. The attack accelerated public fear of the dangers of AI, prompting Anthropic CEO Dario Amodei to propose a slowdown in AI development and that third-party external reviewers should be embedded in AI companies. Musk and OpenAI CEO Sam Altman endorsed Amodei’s proposal. The incident has also sparked calls for full-fledged regulation of the tech industry, a sector that has received only light-touch oversight to date.
A rebel alliance for tech
Next to the frontier labs sits a vast ecosystem of AI tools that help individuals, businesses, entrepreneurs, and large and small organizations function in the modern world. These include chatbots for information and agents tasked with specific jobs, such as debugging a computer program, renting a car, booking a flight or buying music tickets.
One person who has thought a lot about how to deliver these tools responsibly is Mark Surman, president of the Mozilla Foundation. Surman believes there should be a “rebel alliance” of individuals and organizations fighting to supply these tools in easily accessible open-source formats that are responsive to human needs, not Big Tech’s business ambitions. “What we want to do is to make sure there is technology out in the world that keeps human flourishing at the heart of its design, as opposed to building the biggest company or the biggest AI model,” he says.
Surman has worked his entire career to create a free and open internet. He worked with some of the first web companies at the inception of the internet 30 years ago, moving to Mozilla as executive director and later as president. Mozilla developed the free Firefox browser and Thunderbird email tools that were instrumental in disrupting Microsoft’s grip on the browser market.
Now, Surman, a Canadian, wants to do the same thing for the basic tools of AI. Through its capital arm, Mozilla Ventures, Mozilla has invested in about 40 companies fostering a more secure and trustworthy AI industry. The portfolio includes Open Universal Machine Intelligence, a Seattle-based start-up helping companies build custom AI models; Palo Alto–based Credo AI, a governance service-provider enabling companies to enforce AI compliance and manage risks; and New York- and Paris-based Hugging Face, target of the inadvertent cyberattack from OpenAI, which hosts a library for the AI community to share machine-learning models.
There is a growing sector of responsible AI companies directly addressing the recent explosion of public fears over AI, Surman says. He wants to build profitable companies by meeting apprehension over job losses, business threats, privacy concerns and corporate power. “We invest in companies that are building real tech that is going head-on at those worries,” he says in an interview.
Surman would like to expand Mozilla’s network of co-investors to boost capital flowing to innovative tech firms at all stages of development – from start-ups to medium-size public companies. The key is to enable “responsible AI and responsible tech to stay mission-aligned and become profitable without getting bought and sucked into Apple or Google or Microsoft.” Notably, only a few weeks after we spoke, one of Mozilla’s stable of companies, Hugging Face, announced it had signed an agreement to be acquired by the world’s largest microchip maker, Nvidia, for US$13 billion.
The choice for the future
Throughout the turmoil of the last year, most of the sustainable- and responsible-investment industry has sat on the sidelines, watching nervously over its tech company holdings, which were accumulated at a time when the public was not so laser-focused on artificial intelligence.
So, what does the future look like in the relationship between ESG investors and the artificial intelligence industry?
First, the ESG industry needs to acknowledge the level of anxiety over artificial intelligence felt by the public and its own investors. Fears of a tech industry meltdown have investors worried about their own portfolios as well as fearing what artificial intelligence is doing to the economy, their children and their communities. ESG investors – who invest in their funds partly because they expect them to reflect their values – need to hear these fears voiced by their ESG funds, even if they are relatively small players in global capital markets.
In terms of specific action, it appears that Anthropic is emerging as an ESG leader and should command a position in ESG portfolios once reserved for giants like Nvidia, Apple and Microsoft. But this investment should also be accompanied by strong engagement as an active investor, supporting a slowdown in AI development across the industry. At the same time, ESG investors could become leaders in the debate over data centres, pushing AI companies to reduce carbon dioxide emissions and water use, and divesting from AI companies that are abandoning climate targets or ignoring their social harms.
ESG-oriented private equity and small-company investment funds should commit significant capital to help finance open-source AI tools and agents, prioritizing easy access, security and trustworthiness. ESG investors should also take a strong position in favour of AI public policy, making their voices heard in supporting laws and regulations to reduce the control risks of AI models.
Without these actions, ESG investment funds and portfolio managers will manage as they have in the past, investing in the shadow of conventional capital, mirroring the tech portfolios of Wall Street banks and investment firms, and operating without regard to the anxieties of their investors.
This is the choice facing the sustainable- and responsible-investment community. It needs to finance an AI industry that understands and urgently acts on growing public fears of AI and comes to grips with the perils posed by increasingly powerful artificial intelligence models.
Eugene Ellmen writes on sustainable business and finance. He is a former executive in the Canadian responsible investment industry.
The Weekly Roundup
Get all our stories in one place, every Wednesday at noon EST.
