Duke Energy AI Grid Upgrades, $87 B Capital Plan, Brookfield Deal, and Coal Plant Extensions (2021-2025)
AI Demand Risk, Duke Energy’s $87 B Plan to Reinforce the Grid
Utilities are aggressively shifting from predictable, incremental load growth models to a rapid expansion of generation and grid capacity, directly driven by explosive and concentrated demand from artificial intelligence data centers. This strategic pivot, exemplified by Duke Energy, introduces significant execution and financial risk as companies invest heavily to capture this new, high-value market segment.
- Prior to 2024, utility load forecasts were characterized by modest, low single-digit annual growth. In 2025, this model was upended as Duke Energy announced it needed to prepare for at least 1.5 gigawatts of new data center demand in the Carolinas alone by 2033, reflecting an unprecedented acceleration in power consumption.
- The primary driver for this shift is the hyperscale growth of AI infrastructure, highlighted by Amazon’s $10 billion investment in North Carolina for cloud computing and AI. This single move is projected to accelerate Duke Energy’s load growth in the Carolinas from approximately 2% to a range of 4-5% between 2027 and 2029.
- This localized surge is part of a national trend. Projections from Deloitte indicate that power demand from U.S. AI data centers could reach 123 GW by 2035, a more than thirtyfold increase from 2025 levels, forcing utilities to undertake capital-intensive projects to prevent grid instability.
- To address this demand, Duke Energy filed its 2025 Carolinas Resource Plan, which controversially extends the operational life of its coal plants by up to four years and proposes new natural gas units to ensure grid reliability, demonstrating the difficult trade-offs between meeting industrial demand and pursuing decarbonization goals.
Duke Energy Charts AI-Driven Load Growth
This chart quantifies the explosive load growth driven by AI data centers, which is the core risk and driver for the strategic shift described in this section.
(Source: Investing.com)
$87 Billion, Duke Energy’s Capital Plan to Power AI Growth
Duke Energy is financing its historic grid expansion through a combination of strategic asset sales and major equity partnerships, a clear strategy to fortify its balance sheet for a new era of capital-intensive growth needed to support the AI economy.
Duke Energy Boosts Capital Plan to $103B
This chart visualizes the massive scale of the 5-year capital plan discussed in the section, showing the financial commitment required to power the AI economy.
(Source: Investing.com)
- In July 2025, Duke Energy completed the sale of its Tennessee Piedmont Natural Gas business to Spire for $2.48 billion. The company allocated the net proceeds of $1.5 billion directly to funding its expanded five-year, $83 billion capital plan focused on energy transition and grid upgrades.
- Shortly after, in August 2025, the company secured a major financial partner, selling a 19.7% non-controlling stake in Duke Energy Florida to Brookfield. This deal infused $6 billion and enabled Duke Energy to increase its total five-year capital plan to $87 billion.
- To further support its infrastructure investments, Duke Energy has outlined plans for an additional $6.5 billion in equity issuance between 2025 and 2029, with $1 billion specifically targeted for 2025, signaling a long-term financing strategy to match its long-term buildout plans.
Table: Duke Energy Strategic Investments (2025)
| Partner / Project | Time Frame | Details and Strategic Purpose | Source |
|---|---|---|---|
| Brookfield Equity Investment | Aug 2025 | Brookfield invested $6 billion for a 19.7% non-controlling equity interest in Duke Energy Florida. This partnership supports a $4 billion increase in Duke’s overall capital plan to $87 billion. | PR Newswire |
| Spire Asset Sale | Jul 2025 | Completed the sale of its Tennessee Piedmont Natural Gas business to Spire for $2.48 billion, with net proceeds of $1.5 billion funneled into the capital plan for grid upgrades. | PR Newswire |
| Expanded Capital Plan | Feb 2025 | Officially raised its five-year capital expenditure plan to $83 billion (later increased to $87 billion), a 13.7% increase, to accommodate rising demand from data centers and industrial customers. | Reuters |
Duke Energy’s 3 Key Alliances for Grid and Generation (2025)
To manage supply chain risks and secure technology for its expansion, Duke Energy has forged critical alliances with key equipment manufacturers and financial players, while simultaneously facing new competition from decentralized power models targeting the same AI data center market.
Duke’s 14 GW Generation Expansion Strategy
This chart details the specific generation expansion (including gas turbines) that requires the key alliances with equipment manufacturers like GE Vernova mentioned in the section.
(Source: Investing.com)
Duke Summarizes 2025 Capital Accomplishments
This chart’s reference to completing “financial transactions to fund its capital plan” directly supports the specific investments and asset sales detailed in this section’s table.
(Source: Investing.com)
- In September 2025, Duke Energy announced a strategic partnership with GE Vernova to manage and secure additional natural gas turbine production capacity. This move is designed to mitigate supply chain bottlenecks and ensure a stable equipment supply for the new power plants required to meet AI-driven load growth.
- The August 2025 equity partnership with Brookfield was not just a financial transaction but a strategic alignment, providing Duke Energy with the capital to accelerate its grid modernization projects and directly respond to the energy needs of large industrial customers in Florida.
- A competing model emerged in October 2025 when Brookfield also partnered with Bloom Energy in a $5 billion deal to build “AI factories” with onsite power generation. This represents a significant challenge to the traditional utility-supply model, offering data centers a path to bypass the grid for reliable power.
Table: Duke Energy Strategic Partnerships (2025)
| Partner / Project | Time Frame | Details and Strategic Purpose | Source |
|---|---|---|---|
| GE Vernova | Sep 2025 | Strategic partnership to secure natural gas turbine production capacity, ensuring a stable supply chain for new power plant construction needed for AI load growth. | Duke Energy |
| Brookfield | Aug 2025 | Secured a $6 billion investment from Brookfield for a 19.7% stake in Duke Energy Florida, directly supporting an expanded capital plan to serve industrial and AI-related growth. | PR Newswire |
| Sumitomo Corporation | Jul 2025 | Representation in a 50/50 joint venture for the development, construction, financing, and operation of energy projects, strengthening its project execution capabilities. | Norton Rose Fulbright |
The Carolinas, Duke Energy’s Epicenter for AI Power Demand
Duke Energy’s strategic and capital focus has intensely narrowed on the Carolinas and Florida, regions experiencing a unique confluence of rapid economic growth and a surge in data center development that is forcing a fundamental, location-specific overhaul of energy infrastructure.
- Between 2021 and 2023, utility infrastructure planning was more generalized across service territories. By 2025, the Carolinas emerged as the distinct epicenter of load growth for Duke Energy, driven almost entirely by projected energy needs from new AI data centers.
- This geographic concentration prompted the filing of the 2025 Carolinas Resource Plan in October 2025. This regulatory filing is a direct response to the “unprecedented growth” in the region and outlines a location-specific plan for infrastructure modernization and new generation.
- Similarly, in Florida, Duke Energy proposed a plan in September 2025 to develop data centers directly, a proactive strategy to capture the rising demand from AI and cloud computing sectors within that specific service territory.
- The company’s proposal to extend the life of the Gibson coal plant in Indiana and its new infrastructure investments in Florida further underscore a state-by-state strategy to ensure grid reliability in key industrial and high-growth zones.
Technology Integration, Duke Energy’s Use of AI for Grid Management
While racing to build physical infrastructure, Duke Energy is simultaneously deploying AI-driven software and patented grid simulation tools to manage the very system being strained by AI’s growth, elevating these technologies from pilot projects to operational necessities.
AI Applications For Utility Grid Modernization
This infographic perfectly illustrates the types of AI-driven software and grid management tools that the section describes Duke Energy as deploying to manage its infrastructure.
(Source: SmartDev)
- In the 2021-2024 period, AI applications in the utility sector were largely exploratory. By 2025, Duke Energy had moved to active deployment, tapping a hybrid AI model to proactively identify grid vulnerabilities and enhance predictive maintenance in response to increasing stress.
- The company patented new grid simulation technology in January 2025, which allows it to model future grid operations with complex loads. This tool is critical for planning the integration of large, energy-intensive data centers without compromising system stability.
- To improve customer-facing operations, Duke Energy introduced an AI-powered chatbot in its mobile app in August 2025. The chatbot handled a significant volume of inquiries in its first three months, demonstrating a practical application of AI for operational efficiency.
- These internal developments are complemented by external scouting, such as participation in EPRI’s Incubatenergy® Labs, to identify and integrate cutting-edge grid modernization technologies from startups.
SWOT Analysis, Duke Energy’s AI-Driven Growth Strategy
Duke Energy’s strategy to meet AI-driven demand creates a powerful growth opportunity but is constrained by significant regulatory hurdles, massive execution risks, and a difficult, publicly scrutinized compromise between ensuring reliability and meeting decarbonization targets.
Duke Energy Forecasts Strong EPS Growth
This chart quantifies the “powerful growth opportunity” aspect of the SWOT analysis by showing the projected earnings-per-share growth resulting from the AI-driven strategy.
(Source: Investing.com)
- The company’s strengths lie in its incumbent market position and ability to raise substantial capital, while its primary weakness is its renewed dependence on fossil fuels to guarantee power for data centers.
- The clear opportunity is capturing the multidecade growth from powering the digital economy, but this is threatened by the potential for slower-than-forecasted AI adoption and competition from decentralized power solutions.
Table: SWOT Analysis for Duke Energy’s AI Infrastructure Strategy
| SWOT Category | 2021 – 2023 | 2024 – 2025 | What Changed / Resolved / Validated |
|---|---|---|---|
| Strengths | Stable, regulated utility model with predictable, low growth. Focus on incremental grid improvements and renewables integration. | Proactive capital strategy ($87 B plan), strong access to capital markets (Brookfield deal), and dominant position in high-growth territories. | The company validated its ability to pivot quickly and raise massive capital to meet a sudden, structural shift in market demand. |
| Weaknesses | Navigating pressure to accelerate coal plant retirements and meet clean energy mandates. | Increased reliance on fossil fuels by extending coal plant life and building new gas units, creating a conflict with long-term decarbonization goals and ESG commitments. | The AI power crunch exposed the tension between clean energy goals and the need for 24/7 reliable power, forcing a pragmatic but controversial reliance on fossil fuels. |
| Opportunities | Modest load growth from residential expansion and economic development. Gradual EV charging infrastructure build-out. | Massive, sustained load growth from AI data centers, creating a multi-decade opportunity to expand the regulated asset base and become a foundational enabler of the digital economy. | The scale of the AI energy demand was validated, transforming it from a future possibility into an immediate, core driver of the company’s growth strategy. |
| Threats | Regulatory pushback on rate increases; impact of severe weather events on grid reliability. | Competition from decentralized power (e.g., onsite fuel cells from Bloom Energy); risk of overbuilding if AI energy efficiency improves or growth slows; regulatory/public opposition to fossil fuel extensions. | The threat of non-utility competition became concrete with the Brookfield-Bloom partnership, validating a decentralized model as a viable alternative for large power users. |
Duke Energy 2026, Will Flexible Loads Defer New Power Plants?
The single most critical variable for Duke Energy and the U.S. utility sector in 2026 is whether they can successfully design and implement flexible load agreements with data centers. If achieved, this could manage peak demand and defer the need for billions of dollars in new fossil fuel generation.
- If flexible load arrangements gain traction: Watch for the introduction of new tariffs and contracts that allow data centers to curtail power during peak stress events. A successful implementation, validated by a Duke University study showing the grid can handle up to 126 GW of flexible demand, could significantly slow the planned build-out of new natural gas peaker plants.
- If flexible load arrangements fail or are slow to materialize: Watch for an acceleration in the permitting and construction of new gas-fired power plants as the default solution for reliability. This would indicate that the grid is being built to meet inflexible, peak demand, locking in carbon emissions and higher infrastructure costs.
- Signals to monitor include regulatory filings for demand-response programs targeted at large industrial customers, public announcements of load management partnerships between utilities and data center operators, and technology adoption that enables rapid load-shedding at data centers without disrupting critical operations.

