Chevron AI Strategy: Both Sides of the AI Equation

Chevron is on both sides of the AI equation. It uses AI to find and produce oil and gas, and it is building a business that supplies power to AI data centers, starting with a 20-year deal with Microsoft. In October 2026, it reported more than $100 million of value from AI so far in the year, a small figure next to its $3 billion to $4 billion cost-cutting program.

The CFO frames it as two sides. On the third-quarter 2026 results call on 30 October, CFO Eimear Bonner said: “We are sitting on both sides of the AI equation.” She added that “2026 is also proving to be an inflection year for AI, with more than $100 million in value capture to date” (call transcript).

Inside, AI is a productivity tool. Chevron’s 2025 annual report says it “is applying artificial intelligence (AI) to drive productivity, efficiency and value to its global operations.” Senior leadership, including CEO Mike Wirth, oversees an enterprise-wide AI strategy (Fortune, June 2026), which Chevron has not published.

Outside, AI is a market. At its November 2025 investor day, Chevron set out plans to “advance power solutions for AI data centers.” In June 2026, it signed a 20-year agreement to supply about 2.67 gigawatts of gas-fired power to a Microsoft data center in West Texas, with first power targeted for 2028 and a final investment decision expected by the end of 2026 (World Oil). Jeff Gustavson, president of Chevron New Energies, said: “AI is reshaping the global economy, and abundant, affordable, reliable energy is essential to fueling that transformation.”

Inside, the focus is the subsurface. Chevron uses AI across (third-quarter 2026 call):

• exploration: ApEX, its own AI tool, draws on four petabytes of data and 100 years of company knowledge. Scout AI builds basin assessments “in hours, not weeks” and now runs in 10 new basins.
• drilling and production: real-time drilling optimization and an agentic assistant for production engineering
• refining and LNG: predicting and reducing downtime, with teams of people and AI agents planning refinery turnarounds
• finance: investor relations analysis, audit support and forecasting (Fortune)

Chevron has catalogued about 15 enterprise workflows and use cases for AI and advanced analytics, but it has not said which come first or what it will not prioritize.

Proprietary AI for the core business. ApEX and Scout AI are built in-house. An AI “moonshots” program targets exploration and reservoir recovery, often with technology partners. A small skunkworks team created a few years ago has been scaled up (Fortune, June 2026).

Off-the-shelf tools for everyone else. Employees use Microsoft Copilot and Anthropic’s Claude. Chevron set up an Enterprise AI team around 2023 to make AI “as commonplace as something like Microsoft Word and Excel inside Chevron” (CMG, 2023).

Ownership sits at the top. Anu Krishnan, Chief Data and Analytics Officer, leads enterprise data and AI strategy; she joined from Shell in 2025 (DataIQ, 2026). Bonner, now CFO, was Chevron’s Chief Technology Officer before 2024. No AI budget has been disclosed.

Training reaches the workforce, and finance goes furthest.

• Chevron offers “tailored generative AI training for leaders, practitioners and the broader workforce” (2024 annual report).
• Bonner wants all of finance, about 3,500 people, trained in effective AI use. Her framing: “Think of AI as your partner to glean more insights” (Fortune).
• On the October 2026 call, she spoke of “equipping everyone in the organization with the capabilities that they need.”

This is a role-by-role approach to AI fluency, with no company-wide standard published.

The workforce is shrinking, but not in AI’s name. In February 2025, Chevron said it would cut 15% to 20% of its workforce by the end of 2026. Vice Chairman Mark Nelson said the work involves “leveraging technology to enhance productivity” (Axios). At its investor day, Chevron raised its structural cost reduction target to $3 billion to $4 billion by the end of 2026. Headcount fell 5% in 2025, even after the Hess acquisition. Chevron has not attributed the cuts to AI.

Governance is thinly described. Chevron says it is “employing processes to assess its capabilities, limitations and readiness” for AI and is a member of the Responsible AI Institute (2025 annual report). The sources reviewed contain no published AI principles or review process.

AI fluency, coded against the Lead with AI Fluency Matrix (September 2026 edition):

• Fluency floor: Not published. Generative AI training is offered across the workforce, and finance aims to train all of its roughly 3,500 people.
• Highest level evidenced: Level 3 (Integrating), company-claimed. AI is built into exploration, drilling and refinery planning, including an agentic assistant for production engineering. Chevron has published no process-level results, so Level 4 is not evidenced.
• Portfolio design: Specialist-led. A central team builds proprietary AI for the subsurface, while general staff use Copilot and Claude.

Chevron reports one AI value figure. Bonner put AI value captured in 2026 at more than $100 million by October and said she expects it to increase significantly (call transcript). Chevron has not said how the figure is measured or what it includes. It is company-claimed.

The figure is small next to the cost program. It is a small share of the $3 billion to $4 billion in structural cost reductions Chevron targets by the end of 2026, and Chevron has not said how much of that program comes from AI. Even loosely defined, a dollar figure is the step that moves a company from AI adoption to AI value.

The operational claims have no numbers attached. Basin assessments in hours instead of weeks, more than 150 exploration leads found by AI pattern recognition in the Gulf of America, and less downtime in refining are all company claims without published baselines. Bonner’s test for each use case: “How can AI help improve performance or overcome what’s limiting performance?”

The results release is silent. Chevron’s third-quarter 2026 release did not mention AI; the AI figures came on the call.

Next markers: the final investment decision on the Microsoft power project, expected by the end of 2026, and full-year 2026 results, which will show how far the workforce cut has gone.

Change log

• 2023: Chevron sets up an Enterprise AI team to scale AI across the company.
• February 2025: Chevron announces a 15% to 20% workforce cut by the end of 2026, citing technology among its levers.
• November 2025: Investor day sets out power for AI data centers and raises the structural cost target to $3 billion to $4 billion.
• February 2026: The 2025 annual report says Chevron applies AI to drive productivity, efficiency and value.
• June 2026: The CFO describes about 15 enterprise AI workflows, and Chevron signs a 20-year power deal with Microsoft.
• October 2026: The CFO reports more than $100 million of AI value captured in 2026.

• Disclosure: 2 of 4. Chevron answers where it will focus and, unusually, how it will know, with a dollar figure for AI value. Where it is going and what it will take are answered only in part.
• It sits on both sides of the AI equation: it uses AI, and it is building a power business for AI data centers.
• It puts a number on AI value, more than $100 million in 2026, though it has not defined it.
• Its AI is built for the subsurface first, with proprietary tools for exploration and production.
• The workforce is shrinking by 15% to 20%, with technology cited as a lever and AI not named as a cause.
• The say-do gap is scope. AI value is small next to the cost program, and Chevron has not said how AI fits into its restructuring.
• Watch the AI value figure. The test is whether it grows, gets defined, and shows up in the results release, not only on the call.

What the numbers show: Revenue fell 23% from 2022 to 2025 as oil and gas prices came down from their 2022 peak, and headcount fell 2% despite two acquisitions. Chevron plans to cut 15% to 20% of its workforce by the end of 2026, so most of that reduction is not yet in these figures. These figures do not show that AI caused any change.

Revenue, headcount and revenue per employee, FY2022 to FY2025

• FY2022: $246.3bn revenue, 43,846 employees, $5.62M per employee
• FY2023: $200.9bn revenue, 45,600 employees, $4.41M per employee
• FY2024: $202.8bn revenue, 45,298 employees, $4.48M per employee
• FY2025: $189.0bn revenue, 43,039 employees, $4.39M per employee

Change, FY2022 to FY2025: revenue down 23%, headcount down 2%, revenue per employee down 22%.

Notes: Revenue is total revenues and other income, as reported. Headcount is employees at year-end, including about 5,000 service station employees. Revenue moves mainly with oil and gas prices. The figures include PDC Energy from August 2023 and Hess from July 2025.

These figures give context on scale and productivity. They do not show that AI caused any change. Sources: Chevron 10-K filings.

What is Chevron’s AI strategy?

Chevron uses AI to improve productivity and efficiency in its operations, with a focus on exploration, drilling, production and refining, and it is building a business that supplies power to AI data centers. Its enterprise AI strategy is overseen by senior leadership but has not been published.

‍

How does Chevron use AI in exploration?

Chevron uses its own AI tools, ApEX and Scout AI, to analyze subsurface data and build basin assessments in hours instead of weeks. It says AI pattern recognition has found more than 150 exploration leads in the Gulf of America.

‍

Is Chevron cutting jobs because of AI?

Chevron is cutting 15% to 20% of its workforce by the end of 2026 and has cited technology as one lever, but it has not attributed the cuts to AI.

‍

What is Chevron’s AI data center power business?

In June 2026, Chevron signed a 20-year agreement to supply about 2.67 gigawatts of gas-fired power to a Microsoft data center in West Texas, with first power targeted for 2028.

‍

How is this profile put together?

Lead with AI profiles use public sources only, such as annual reports, earnings materials, filings and on-record interviews, dated at the time of writing. AI fluency is coded against the Lead with AI Fluency Matrix. Company figures are marked as company-claimed unless they have been independently verified.