The Visionary CEO’s Guide to Sustainability
|
Executive Summary
This article is part of Bain’s 2026 CEO Sustainability Report Ask senior executives what share of global energy AI will consume three years from now, and the typical answer is likely to be around 11%. Consumers’ average estimate is even higher at 19%. But Bain’s proprietary climate-economic modeling tool, IntersectSM, forecasts 0.7%. The gap between perception and forecast is enormous: nearly 16-fold among executives and almost 30-fold among consumers (see Figure 1). Their calculations may be off, but their concern is real. The share of Americans more worried than excited about AI rose from 37% in 2021 to 50% in 2025, according to data from the Pew Research Center. In Bain’s consumer research, nearly two in three respondents report taking concrete action due to their concerns about AI, such as limiting what they share, switching platforms, dropping certain tools, or speaking out publicly.
Figure 1
This concern is not mirrored inside companies. In our second annual survey of 400 senior professionals responsible for AI and sustainability decisions across industries and regions, AI’s environmental intensity ranked tenth out of thirteen barriers to sustainable AI. Only 22% of respondents placed it among their five greatest challenges. Companies are focused on other challenges: proving the value, building the capabilities, and creating the data and standards needed to scale. With two years of data, companies’ direction of travel is becoming clear, and increasingly divergent. Two divides are opening: one between companies, the other within them. The first is between companies that are learning and companies that are losing faith. We identified a small group of sustainable AI leaders we call “shapers,” representing about 20% of the sample, based on their AI maturity, breadth of adoption, reported value, and overall sustainability maturity. Among these companies, belief in AI’s sustainability potential remains close to 90%, down just four percentage points from last year. Confidence among “laggards,” the companies at the opposite end of the spectrum, was far lower and fell further from last year, to 41% from 57% (see Figure 2).
Figure 2
Shapers’ modest decline reflects the calibration of experience. This group has actively adopted 86% of the sustainable AI use cases in our survey, compared with just 31% for laggards. That breadth of deployment gives shapers a more grounded view of what AI can deliver. Their expectations may have narrowed, but they are now anchored in tangible results. Laggards have fewer results against which to judge. They started slowly, and their confidence is falling even before they have built the capabilities needed to test the opportunity. They are only about half as likely as shapers to invest in quality sustainability data or upskill their people, and less than half as likely to codevelop tools with partners. There’s a danger that laggards lose conviction just as leading companies accumulate the experience and evidence to move faster. A year ago, shapers and laggards were separated mainly by adoption. Today, they're separated by belief and capability as well. In addition to this divergence between companies, there is a second divide within companies, between the people building the case for sustainable AI and those making the business decisions needed to scale it (see Figure 3). In our survey, the central concern for general managers and C-suite executives is return. Seventy-five percent of this group identify an unclear business case or ROI as a leading barrier to deploying sustainable AI. When evaluating sustainable AI investments, they rank financial return first. The sustainability professionals we surveyed identify different constraints. For this group, data quality and the absence of clear standards rank ahead of ROI. When evaluating sustainable AI investments, they place regulatory compliance and risk management first, and financial return fourth. These groups also approach risk from different directions. Business leaders are more alert to workforce disruption, employee anxiety, and trust, while sustainability professionals focus on energy, emissions, and disclosure.
Figure 3
One group is asking about revenue, margin, and return. The other is focused on data, impact, and compliance. The case is being argued in one language and being judged and funded in another. Neither perspective is wrong, but each is incomplete. Sustainability teams need to connect materiality, data, and compliance to revenue, margin, and resilience—and give greater weight to workforce and trust. Business leaders should account more fully for environmental impacts and the data and standards required to manage them. Across companies and functions, the clearest place to start is where sustainability and financial value are already aligned. Shapers use sustainable AI to make better, sell better, and protect better, and support those applications with an operating system that makes the results repeatable. The highest-value use cases for sustainable AIShapers don’t simply use more AI. They are more than three times as likely as laggards to adopt the highest-value AI use cases: those that improve operations, create commercial advantage, and strengthen resilience (see Figure 4).
Figure 4
Make better: Prove value through operationsFor companies struggling to prove the AI opportunity, operational applications offer the fastest route from ambition to value. Across the companies surveyed, AI-driven sustainability starts with four things: energy efficiency, asset and process efficiency, demand forecasting, and emissions monitoring. These apply across industries, from energy and heavy industry to agriculture and consumer goods. Consider some of the ways technology supplier Bosch has deployed AI across its manufacturing and logistics operations. Through applications including predictive, risk-based maintenance and agentic orchestration of both inbound supplier workflows and delivery sequencing, the company has reduced unplanned downtime and streamlined logistics, generating significant efficiency gains. The potential for such gains is broad and significant. According to the IEA, by optimizing production processes and cutting usage in energy-intensive industries, well-documented AI use cases could save more than 13 exajoules of energy by 2035, equivalent to 3% of global final energy consumption. Sell better: Turn sustainability into commercial advantageHere, the translation gap between business and sustainability begins to close. AI helps companies design more compelling sustainable products, identify the customers most likely to value them, and translate sustainability into growth. Our 2025 B2B survey found that revenue growth leaders are 1.6 times more likely to use sustainability as a top-line growth lever. They develop new sustainable products and services, gain share, reach new customer segments, and earn green premiums. Most companies, however, do not sell sustainability well. AI can help convert a broad sustainability proposition into a targeted commercial one by analyzing disclosure data at scale, identifying customers for whom sustainability matters most, benchmarking competing propositions, and generating sales plays tailored to a buyer’s commitments and unmet needs. A leading manufacturer piloted an AI-powered sales intelligence tool in one of its business units, automatically surfacing customer-specific sustainability data to help reps tailor value propositions to each buyer's needs. The team saw 80% growth in its opportunity pipeline during the pilot and secured three major retailers wins—demonstrating what embedding AI into sustainability-focused commercial sales processes can unlock. Protect better: Put climate risk on the balance sheetProtecting better brings the two perspectives together: Sustainability teams identify physical risks, while business teams translate those risks into capital allocation and financial value. As the climate warms, physical climate risk is increasingly showing up in corporate financials through insurance cost inflation, asset value erosion, and supply chain damage. Companies with high physical risk exposure face a 22-basis-point premium in their weighted average cost of capital, according to Bloomberg. Companies are beginning to use AI to help identify and manage these risks. When a leading global pension fund used AI-powered climate risk data to assess its portfolio, executives could identify which assets offered the greatest potential return from adaptation. Targeted measures could increase the value of selected properties by mid-single-digit percentages, through a combination of lower insurance and energy costs, stronger rents, and higher valuations. AI can provide the regional and asset-level granularity needed to quantify and manage climate exposure. Shapers in our survey have adopted use cases that help them protect better at four times the rate of laggards. Building the operating system for scaleFinding value is only the first step. Our study shows that leading companies build an operating system that institutionalizes sustainable AI decisions from investment scoping through deployment. Shapers are twice as likely as laggards to make sustainability mandatory in AI scoping, twice as likely to upskill employees on sustainable AI, and nearly four times as likely to consider sustainability risks consistently during deployment. They connect AI to the company’s material priorities; bring business, technology, and sustainability leaders into the same decisions; and assess value and risk together as applications scale (see Figure 5).
Figure 5
Note: CSO is Chief Sustainability Officer Source: Bain Sustainable AI Survey 2026 (n=400)That operating discipline also helps close the translation gap inside companies. Sustainable AI scales when business and sustainability leaders work from a shared definition of value, combining revenue, margin, and resilience with environmental impact, workforce effects, and trust. The divergence in our data is likely to widen. Shapers are turning deployment into evidence and capability, while laggards risk losing conviction before they have fully tested the opportunity. Public scrutiny of AI’s footprint will continue, and companies must manage it carefully. Visionary leaders will go further: aligning their organizations around a shared case for its value and building on each deployment to strengthen the next. That is how sustainable AI becomes part of the way companies make better, sell better, and protect better at scale. Read our 2026 CEO Sustainability ReportMore from the report |