Research Report
Powered for Change 2025
Industrial decarbonization in the age of gen AI
10-minute read
August 15, 2025
In brief
From isolated projects to multigenerational transformation
- One-off pilots won’t decarbonize industry fast enough. With heavy industry lagging on 2050 climate goals, companies must shift from isolated projects to a multigenerational, portfolio-driven approach that compounds learnings, standardizes and modularizes components and encourages innovation.
- Scaling up unlocks cost efficiencies. Our research confirms that early, low-carbon investments, though expensive, become cost-competitive as continuous learning and efficiencies take hold.
- AI drives compounding efficiencies. Beyond accelerating projects, AI preserves institutional knowledge and ensures every initiative builds on past insights to drive sustained cost reductions.
Reinventing net‑zero infrastructure, at scale
Every year to 2050 counts—for the planet and for the industries reshaping it. Heavy industry, oil and gas and power companies must significantly reduce emissions while continuing to meet growing demand and sustain profitable growth.
Powered for Change 2025 focuses on the how. It sets out a roadmap for reinventing the build-out of net-zero infrastructure, including renewables, nuclear, green hydrogen, carbon capture, lower-carbon manufacturing and the transmission and distribution networks that connect them.
At the center is a multigenerational approach: a shift from isolated projects to connected systems where each generation of infrastructure builds on the last—technically, financially and strategically (see Figure 1).
Figure 1. Costs decline and savings rise when you take a multigenerational approach.
Proportions based on an inverse S-curve and NPV calculation for levelized cost of green hydrogen in Europe in base and optimized "multigenerational" scenario. Source: Accenture S-curve model.
01
Why this matters now
Ambition alone is no longer enough. Volatile markets, trade disruption, supply chain fragility and policy uncertainty are making capital projects harder to deliver. Our research reveals rising skepticism towards capital projects, both traditional and green. Negative sentiment grew from less than 35% in 2024 to 50% in 2025. At the same time, companies face persistent barriers beyond economics, including infrastructure inflexibility, missing community consent and a scarcity of talent.
Figure 2. Companies have growing concerns about the viability of capital projects. Negative sentiment is on the rise.
Percentage of companies in utilities, oil and gas, chemicals and mining and metals that mention cancelling capital projects or cutting or delaying capital investments—measured as a share of all companies referring to capital projects or investment plans. Source: Accenture Research analysis augmented with AI, using data from earnings calls and company publications from January 1, 2024 to April 15, 2025.
A multigenerational, repeatable approach helps organizations hedge against uncertainty, safeguard access to capital and sustain momentum by turning learning into compounded advantage.
02
From bespoke projects to a multigenerational approach
Today, most net-zero infrastructure projects are still planned and executed as one-off efforts. This bespoke model increases cost, risk and delivery time.
A multigenerational approach replaces reinvention with iteration. Companies develop modular designs, standardized ways of working and repeatable supply chains, then improve them with each deployment. Over successive generations, accumulated learning drives sharp cost reductions and faster time to value.
03
Repeatable delivery redefines the cost curve
Accenture’s inverse S-curve modeling shows how costs fall as organizations move from first-of-a-kind projects to repeatable delivery. Early projects deliver modest savings, followed by a tipping point where learning and scale drive rapid cost reductions, and then sustained gains as learning compounds across portfolios.
Green hydrogen illustrates this effect, as shown in Figure 3, below. A multigenerational approach can deliver earlier cost parity, significant cumulative savings and material net present value by capturing learning across successive projects rather than treating each build in isolation.
Figure 3. Taking the multigenerational approach and modeling it on green hydrogen, using an optimized inverse S-curve.
7% WACC, discounted from 2025-2050, for Europe. The cost of fossil-fuel based gray hydrogen is expected to increase in line with EU carbon tax, at $70-80/t CO2 today, $150/t in 2037 and $300/t in 2050. Source: Accenture S-curve model.
04
AI as a force multiplier
AI’s greatest impact is not accelerating single projects but embedding continuous learning across portfolios. By analyzing vast amounts of structured and unstructured project data, AI captures insights beyond human scale, accelerating learning curves and improving decision quality.
Companies that build this capability can outperform today and create the foundation for outperformance in future generations of projects, as shown by modeling on green hydrogen in Figure 4, below.
Figure 4. Taking a multigenerational approach with green hydrogen is a >$60 billion opportunity.
H2 demand based on IEA WEO 2024 NZE 2050 scenario. NPV calculation based on delta LCOH between base and optimized scenario, 7% WACC, discounted from 2025-2050. The cost of fossil fuel-based gray hydrogen is expected to increase in line with carbon tax, at $70-80/t CO2 today, $150/t in 2037 and $300/t in 2050. Source: Accenture S-curve model.
Four levers that enable the multigenerational approach
Industry lens: Electricity networks in the age of AI
Electricity networks are where the pressures described in Powered for Change 2025 are converging the fastest. Rapid electrification, AI-driven data center demand, aging assets and rising resilience risks are stretching network utilities beyond models built for stable, predictable growth.
Historically, networks have been optimized for static efficiency: lean teams, fragmented processes and limited end-to-end integration managing a fixed asset base. That model is now hitting scalability limits. Fragmentation across organizations, platforms and data makes it harder to compress timelines, optimize under uncertainty or connect capacity at pace.
Incremental fixes and point AI solutions are not enough. Infusing AI into existing processes still leaves outcomes constrained by human-paced orchestration and manual hand-offs. Electricity networks instead need to become AI-native, reinventing end-to-end processes around what AI makes possible.
Figure 5. An example of an AI-native electricity network.
Source: Accenture research, 2026.
In an AI-native network, AI agents, guided by humans and underpinned by platforms and data, dynamically optimize and orchestrate work across planning and delivery. This compresses cycle times, improves decision quality and enables continuous optimization as conditions change, while freeing people to focus on judgment, accountability and system-level trade-offs.
Electricity networks show how the multigenerational approach comes to life. By shifting from fragmented delivery to AI-native optimization, utilities can scale capacity faster, design once and replicate and compound learning across generations of work rather than treating each program or project in isolation.
How Accenture can help
Accenture supports organizations across the full net-zero journey, from strategy through infrastructure transformation and operations. This includes net-zero strategy, sustainable products and markets, net-zero finance, enterprise carbon intelligence, infrastructure transformation and low-carbon operations.
About the research
Powered for Change 2025 combines AI-augmented signal analysis, executive interviews, proprietary inverse S-curve modeling and patent analysis to provide a rigorous, actionable roadmap for scaling industrial decarbonization.