AI Spending Will Double in 2026 - And CEOs Are Taking Control
Corporate investment in AI is set to double next year. The BCG AI Radar 2026, based on responses from 2,360 executives, shows companies plan to raise AI spend to about 1.7% of annual revenue in 2026, up from roughly 0.8% in 2025. The message is simple: AI has moved from experiment to enterprise strategy.
Decision-making has shifted to the top. 72% of CEOs now say they're the main decision-makers on AI-double last year. AI outcomes are tied to enterprise performance, capital allocation, and leadership accountability. Half of CEOs say their job stability depends on getting AI right by 2026.
Investment Will Continue, Even If Near-Term Returns Lag
Companies aren't waiting for perfect ROI. 94% of organizations will keep investing in AI even if current projects don't deliver financial returns in 2026. Only 6% plan to scale back. AI is being treated as essential infrastructure.
Agentic AI Takes Center Stage-With Bigger Risk Exposure
Agentic AI-systems that can operate autonomously-will command 30%+ of AI budgets in 2026. Nearly 90% of CEOs expect measurable returns from these systems this year. But growing autonomy raises the stakes on data privacy and cybersecurity. Without stronger governance, new vulnerabilities are likely.
Who's Moving Fastest
BCG identifies three CEO archetypes: Trailblazers, Pragmatists, and Followers. Around 15% are Trailblazers-investing more aggressively, upskilling more of their people, and deploying AI across full functions. Confidence is highest in India, Greater China, and the Middle East & Africa. In Europe and the United States, many leaders are moving forward primarily due to competitive pressure.
What This Means for Your 2026 Plan
- Set a clear spend baseline: Use the 1.7% of revenue benchmark as a reference point. Allocate 30%+ of the AI budget to agentic AI with tight ROI gates and phased rollouts.
- Make it CEO-led: Treat AI as capital allocation, not experimentation. Tie executive incentives to AI-linked KPIs: cost-to-serve reduction, cycle-time compression, revenue lift, and risk incident rates.
- Stand up agentic guardrails: Approve use cases, enforce data minimization, keep humans in the loop for high-impact actions, and monitor autonomous activities. Use established guidance like the NIST AI Risk Management Framework.
- Prioritize end-to-end workflows: Move beyond isolated pilots. Target full processes in customer operations, finance, and supply chain where time-to-value is shortest.
- Get your data house in order: Improve first-party data quality, access controls, and lineage. Poor data makes agentic systems brittle and risky.
- Adopt a multi-model strategy: Avoid lock-in. Set procurement standards for security, cost, latency, and domain performance. Always run a fallback path.
- Upskill at scale: Define role-based AI curricula for product, ops, finance, and compliance. If you need a shortcut, see curated options by role at Complete AI Training.
- Enforce a stop rule: Kill projects that don't show a path to measurable value within two to three quarters. Reallocate to proven use cases.
The Bottom Line for Executives
AI is now a CEO agenda item with P&L consequences. Budgets are doubling, autonomy is rising, and risk exposure is real. The advantage goes to leaders who set firm budgets, ship production use cases, and build governance that scales as fast as the tech.
2026 is the year to lock your operating model for AI-spend, guardrails, talent, and metrics. Commit, measure, and compound.
Your membership also unlocks:
