AI Infrastructure Investment Seen Surging to $769 Billion in 2026 Amid Growing Safety Concerns

AI Infrastructure Investment Could Reach $769 Billion

Investment in AI infrastructure and model architectures is on track to surge to approximately $769 billion in 2026, according to McKinsey's Technology Trends Outlook 2026.

The estimate represents more than a fivefold increase from the $145 billion invested in 2025. Nearly $384 billion had already been invested during the first six months of 2026, with the full-year estimate assuming that the current pace continues.

This would make AI infrastructure and model architectures the largest-funded technology trend among those tracked by McKinsey.

Why Is AI Infrastructure Attracting So Much Capital?

The rapid expansion of generative AI and increasingly sophisticated AI models is driving demand for computing capacity, advanced chips, data centres and related infrastructure.

Companies developing AI models require large-scale computing resources for both training and running models. This has led technology companies and investors to commit significant amounts of capital to data centres, GPUs, networking equipment and energy infrastructure.

The International Energy Agency has also highlighted the rapid expansion of AI-focused data centres, noting that electricity consumption from such facilities increased sharply in 2025.

AI Investment Surge Comes With Safety Concerns

The investment boom is taking place alongside growing concerns about the pace at which AI capabilities are advancing.

Some AI industry leaders have called for stronger safety measures and more time to develop safeguards as increasingly capable systems are deployed.

Anthropic CEO Dario Amodei has called for greater caution around the release of new AI capabilities, while OpenAI has backed measures including independent safety assessments, cybersecurity safeguards and incident reporting for advanced AI systems.

These concerns have increased the focus on how governments and technology companies should manage advanced AI systems.

Governments Are Looking at AI Safety Rules

AI safety is increasingly becoming a regulatory issue.

Union Minister Ashwini Vaishnaw has said AI regulation will require both legal and technical safeguards and argued that companies developing AI systems should take the lead on safety because they understand the technology and its potential risks.

The debate includes questions about responsibility when autonomous AI systems make consequential decisions or cause unintended outcomes.

Power and Data Centres Are Becoming Critical

The AI investment boom is also creating pressure on the physical infrastructure needed to operate advanced models.

The IEA has identified electricity supply, grid connections, chip manufacturing capacity and data-centre construction as important constraints for the expansion of AI.

AI-focused data-centre electricity consumption grew by 50% in 2025, according to the IEA, while overall data-centre electricity demand increased by 17%.

This means the AI buildout is not only a technology investment story but also an energy and infrastructure issue.

Data Centre Investment Is Also Rising

Separate research from PwC estimates that global data-centre capital expenditure could rise from around $800 billion annually in 2026 to $1.8 trillion annually by 2050.

PwC projects cumulative global investment of approximately $31.6 trillion through 2050 to build the infrastructure required for AI and other digital services.

The estimates illustrate the potentially long-term nature of the infrastructure investment cycle, although actual spending will depend on demand, technology changes, power availability and financing conditions.

AI Financing Faces Growing Scrutiny

Despite strong demand for AI infrastructure, investors and lenders are becoming more attentive to project economics and execution risks.

A Reuters report said Oracle's move to invoke a force majeure provision relating to delays at a major New Mexico data-centre project has increased scrutiny of AI infrastructure financing. The project is being developed by a Blue Owl unit for OpenAI.

The episode has raised questions among lenders about how delays, power constraints and construction risks are allocated in large AI infrastructure projects.

Major AI Companies Continue to Seek Capital

The investment surge is also reflected in fundraising activity among leading AI companies.

According to the report, OpenAI has been in discussions with investors for additional funding at a potential valuation of around $1.2 trillion. SoftBank has also launched an $11 billion bond offering to help finance a further investment in OpenAI.

Anthropic, meanwhile, has been reported to be in discussions with Nvidia over a potential investment of up to $10 billion. These reported transactions underline the scale of capital being directed towards the AI ecosystem.

AI Investment Is Expanding Beyond Model Development

The capital flowing into AI is not limited to companies developing foundation models.

Investment is also moving into:

• AI data centres

• GPUs and advanced processors

• Networking infrastructure

• Electricity generation and grid capacity

• Cloud computing

• AI software and agents

• AI research and scientific applications

McKinsey expects several technology trends, including AI infrastructure and model architectures, agentic software development and AI for scientific discovery, to receive significantly more investment in 2026 than in 2025.

What Does the $769 Billion Estimate Mean?

The projected $769 billion figure represents the scale of investment flowing into AI infrastructure and model architectures, rather than a forecast of AI companies' revenue or profits.

It is also an estimate based on the pace of investment recorded during the first half of 2026. Actual full-year spending could therefore differ from the projection.

The rapid capital deployment also raises questions about whether infrastructure capacity, electricity supply, chip availability and safety frameworks can expand quickly enough to support the industry's growth.

Key Takeaways

• AI infrastructure and model-architecture investment could reach $769 billion in 2026.

• Nearly $384 billion was invested during the first half of the year.

• The projected 2026 figure is more than five times the $145 billion invested in 2025.

• AI data centres are driving increased demand for electricity and computing infrastructure.

• AI leaders and policymakers are raising concerns about safety and the pace of development.

• Governments are considering legal and technical safeguards for advanced AI systems.

• Financing and construction risks are also receiving greater scrutiny as AI infrastructure projects become larger.

• The $769 billion figure is a projection based on the pace of investment and is not a guarantee of actual full-year spending.