Archive · 25 Sep 2026
AI Infrastructure Investment May Reach $769 Billion Amid Safety Concerns
Companies are spending much more money to build artificial intelligence systems.
McKinsey estimates that spending on AI infrastructure and model designs could reach $769 billion in 2026.
That would be more than five times the amount invested in 2025.
The money is being used for items such as computer chips, data centers and electricity systems.
OpenAI, SoftBank and Anthropic are among the organizations discussing large investments.
However, some experts worry that AI is developing faster than safety rules can keep up.
AI systems also need a lot of electricity, which could create pressure on power networks.
Many businesses use AI, but fewer say it has clearly improved their profits.
The next stage may focus on using AI effectively while making it safer and more reliable.
McKinsey projects AI infrastructure and model-architecture investment could reach $769 billion in 2026.
The forecast is based on nearly $384 billion invested during the first half of 2026.
Semiconductors, data centers and power systems are driving demand for AI infrastructure.
OpenAI, SoftBank and Anthropic are pursuing major funding and investment commitments.
Energy constraints, limited grid capacity, cybersecurity risks and safety concerns could slow AI expansion.
- Who
- McKinsey & Company, major AI companies and investors, and businesses adopting AI.
- What
- Investment in AI infrastructure and model architectures is projected to rise to about $769 billion in 2026.
- When
- The projection covers 2026 and is based on investment of nearly $384 billion in the first half of the year.
- Where
- The investment and infrastructure buildup is global, with specific energy impacts discussed for the United States and California.
- Why
- Demand for chips, data centers and power systems is increasing, while companies seek to expand AI capabilities and business use.
Accelerate AI investment
Slow development for safety
Pace of AI development
Accelerate AI investment
Growing investment can expand AI infrastructure, model capabilities and business adoption.
Slow development for safety
Anthropic CEO Dario Amodei has called for a slower rollout of new AI capabilities to allow more time to address safety risks.
Regulation and safeguards
Accelerate AI investment
Investment is supporting the development of AI systems and their integration into business workflows.
Slow development for safety
OpenAI has backed mandatory national safety requirements, including independent assessments, cybersecurity measures and incident reporting for advanced AI systems.
Infrastructure expansion
Accelerate AI investment
More semiconductors, data centers and power systems are needed to meet rising demand for AI.
Slow development for safety
Energy constraints and more than 2,500 gigawatts of projects awaiting grid connections could limit further expansion.
- Projected 2026 investment
- About $769 billion in AI infrastructure and model architectures.
- 2025 investment
- $145 billion.
- First-half 2026 investment
- Nearly $384 billion.
- OpenAI potential valuation
- Around $1.2 trillion in funding discussions.
- Business AI adoption
- 89% of organizations regularly use AI in at least one business function.
- Reported EBIT impact
- 37% of organizations report a positive impact on EBIT.
- Potential energy impact
- U.S. AI data centers could use as much electricity by 2030 as California uses today.