SoftBank Predicts AI Will Demand $5 Trillion a Year by 2040, Why the Biggest Technology Investment Boom Is Still Ahead

Luca Moretti
9 minutes ago
7 min read

Artificial intelligence has entered a phase where the discussion is no longer centered on whether the technology will transform industries, but on the scale of investment required to build the infrastructure capable of supporting it. As AI models become increasingly sophisticated, demand for computational power, specialized semiconductors, energy generation, networking equipment, robotics, and autonomous software continues to accelerate.

Against this backdrop, SoftBank Group Chairman and CEO Masayoshi Son presented one of the boldest long-term forecasts yet, arguing that global AI development could require annual investment of approximately US$5 trillion by 2040. He also rejected suggestions that current enthusiasm surrounding AI represents a speculative bubble, instead portraying artificial intelligence as the foundation of an entirely new economic era.

Whether or not this projection ultimately proves accurate, it reflects a broader shift taking place across technology, finance, infrastructure, and energy markets. AI is rapidly evolving from a software innovation into a massive industrial ecosystem requiring unprecedented levels of capital expenditure.

AI Is Becoming Infrastructure Rather Than Just Software

For decades, software companies scaled primarily through cloud services and digital distribution. Artificial intelligence changes this equation.

Modern AI systems require enormous physical infrastructure, including:

Hyperscale data centers
Advanced graphics processing units (GPUs) and AI accelerators
High-speed networking equipment
Massive storage systems
Reliable electricity generation
Cooling technologies
Fiber connectivity
Semiconductor manufacturing capacity

Unlike traditional software products, cutting-edge AI models require continuous investment even after deployment. Training larger models demands increasingly powerful hardware, while serving billions of AI requests each day requires extensive inference infrastructure.

This transition explains why technology companies are investing not only in software research but also in semiconductor design, energy projects, robotics, networking technologies, and cloud infrastructure.

Understanding the Scale of a US$5 Trillion Annual Investment

An annual investment figure of US$5 trillion represents an extraordinary level of economic activity.

Rather than viewing this solely as spending on AI software, it is more accurate to interpret it as investment across an entire industrial ecosystem supporting intelligent systems worldwide.

Potential areas receiving significant investment include:

Sector	Why Investment Is Increasing
Data centers	Expanding AI training and inference capacity
Semiconductor manufacturing	Producing advanced AI chips
Power generation	Meeting rapidly rising electricity demand
Networking infrastructure	Enabling faster AI communication
Robotics	Deploying autonomous systems in industry
Cloud platforms	Delivering AI services globally
Edge computing	Running AI closer to users and devices
Cybersecurity	Protecting increasingly autonomous systems

Each component reinforces the others, creating a feedback loop in which infrastructure investment enables larger AI models, which in turn create demand for even greater infrastructure.

Why AI Spending Is Rising So Rapidly

Several technological trends are driving investment growth simultaneously.

Larger Foundation Models

Modern foundation models contain dramatically more parameters than earlier generations and require vast computational resources during training.

Although improvements in algorithmic efficiency continue to reduce certain costs, overall computational demand continues expanding as organizations pursue increasingly capable systems.

AI Agents

One of the most significant emerging trends is the evolution from conversational assistants toward autonomous AI agents capable of planning, reasoning, coordinating tasks, and interacting with digital systems.

Instead of responding to isolated prompts, AI agents may eventually manage workflows, negotiate with other software, monitor operations, analyze data continuously, and execute complex business processes with limited human intervention.

If millions or even billions of AI agents become active simultaneously, infrastructure requirements could increase substantially.

Enterprise Adoption

Organizations across finance, healthcare, manufacturing, logistics, retail, education, media, and scientific research are integrating AI into core operations.

Unlike earlier digital transformation initiatives that often remained department-specific, AI increasingly influences entire business processes.

As adoption widens, demand for scalable infrastructure naturally grows.

AI and the Global Economy

One of the most ambitious elements of SoftBank's long-term outlook is the expectation that AI-generated economic activity could represent a significant share of global gross domestic product by 2040.

Regardless of the precise percentage ultimately achieved, AI is increasingly becoming a productivity technology rather than merely an automation tool.

Historically, transformative technologies have generated economic value by:

Increasing labor productivity.
Creating entirely new industries.
Reducing operational costs.
Improving decision-making.
Accelerating scientific discovery.
Expanding access to digital services.

Artificial intelligence has the potential to influence all six simultaneously.

If AI continues expanding across virtually every major sector, its contribution to global economic output could become comparable to previous industrial revolutions.

Is AI Really a Bubble?

The debate surrounding an AI bubble has intensified as valuations of leading AI companies have risen sharply and capital spending has reached record levels.

Supporters argue that current investment reflects genuine infrastructure development rather than speculative excess.

Skeptics point to several risks:

High capital expenditures
Uncertain long-term profitability
Rapid technological change
Competitive pressure
Infrastructure oversupply
Regulatory uncertainty

History offers examples supporting both perspectives.

Previous technology booms have included periods of excessive optimism, yet many also produced infrastructure that later enabled sustainable economic growth.

The internet provides perhaps the clearest example. During the dot-com era, many companies failed, but investments in fiber networks, servers, and digital infrastructure ultimately became essential foundations of today's digital economy.

AI may follow a similarly uneven trajectory, with some firms struggling while the broader ecosystem continues expanding.

Energy Could Become AI's Greatest Constraint

Perhaps the most important long-term challenge facing AI is electricity.

Training advanced AI models requires immense computational power, while serving billions of inference requests every day places continuous demands on energy infrastructure.

As AI deployment accelerates, electricity generation may become one of the industry's primary limiting factors.

Potential energy sources include:

Natural gas
Nuclear power
Renewable energy
Hydroelectric power
Advanced battery storage
Future nuclear fusion technologies

Each option presents trade-offs involving cost, reliability, emissions, scalability, and geographic availability.

Because data centers operate continuously, reliable baseload power remains

particularly important.

Why Data Centers Have Become Strategic Assets

Modern AI depends on hyperscale computing facilities capable of housing hundreds of thousands of specialized processors.

Today's AI data centers increasingly resemble industrial complexes rather than traditional server rooms.

Key characteristics include:

Massive power consumption
Advanced liquid cooling systems
High-bandwidth networking
Redundant electrical infrastructure
Specialized AI hardware
Extensive physical security

Governments increasingly view these facilities as strategic national infrastructure due to their importance for economic competitiveness, technological leadership, and national security.

Robotics and Physical AI

Artificial intelligence is expanding beyond digital applications into the physical world.

Robotics represents one of the most promising areas for future AI deployment.

Potential applications include:

Manufacturing automation
Warehouse logistics
Healthcare assistance
Agriculture
Construction
Inspection systems
Disaster response
Elder care

Unlike earlier industrial robots programmed for repetitive tasks, future AI-powered robots may continuously learn, adapt, and collaborate with humans.

This convergence of AI and robotics significantly increases demand for advanced processors, sensors, networking, and edge computing infrastructure.

The Rise of AI Agents

One of the most transformative concepts shaping AI's future is the emergence of autonomous software agents.

Rather than functioning merely as chatbots, future AI agents may:

Manage business operations
Coordinate supply chains
Schedule complex workflows
Monitor financial systems
Conduct research
Negotiate digital transactions
Control robotic systems
Collaborate with other AI agents

Such systems could fundamentally reshape knowledge work by automating increasingly sophisticated cognitive tasks.

However, they also introduce significant governance challenges involving transparency, accountability, cybersecurity, and ethical oversight.

Challenges That Could Slow AI Expansion

Despite rapid progress, several obstacles could influence AI's long-term growth.

Infrastructure Bottlenecks

Manufacturing advanced chips remains highly complex and concentrated among a relatively small number of companies.

Expanding fabrication capacity requires years of investment.

Power Availability

Many regions already face electricity constraints.

Large AI deployments could intensify competition for available energy resources.

Regulatory Evolution

Governments continue developing frameworks addressing AI safety, privacy, intellectual property, competition, and national security.

Regulatory uncertainty may influence investment decisions.

Talent Shortages

Highly skilled AI researchers, engineers, semiconductor specialists, and data center experts remain in limited supply.

Building sufficient human expertise may prove as challenging as expanding hardware infrastructure.

Investment Trends Beyond Software

The AI economy increasingly spans multiple interconnected industries.

Industry	Growing Strategic Importance
Semiconductor manufacturing	Core computing capability
Cloud computing	AI deployment platform
Telecommunications	High-speed connectivity
Energy	Reliable power generation
Construction	Data center development
Robotics	Physical AI deployment
Cybersecurity	Protecting AI systems
Advanced materials	Improving chip efficiency

This diversification suggests AI investment is becoming an economy-wide phenomenon rather than remaining concentrated within traditional technology companies.

Implications for Businesses

Organizations do not necessarily need to build frontier AI models to benefit from the technology.

Businesses can create competitive advantages by:

Integrating AI into operations
Improving customer experiences
Automating repetitive workflows
Enhancing data analytics
Accelerating research and development
Supporting employee productivity

Companies that successfully combine AI capabilities with strong governance and high-quality data are likely to achieve greater long-term value than those pursuing AI adoption without clear strategic objectives.

The Road to 2040

Forecasting technology over a fifteen-year horizon inevitably involves uncertainty.

Some expectations may prove overly optimistic, while others may underestimate future breakthroughs.

Nevertheless, several long-term trends appear increasingly established:

AI infrastructure spending continues expanding.
Computational demand remains strong.
Energy has become a strategic technology issue.
Robotics is moving toward broader commercialization.
AI agents are emerging as the next software paradigm.
Governments increasingly view AI as national infrastructure.

Whether annual investment ultimately reaches US$5 trillion or another figure, the direction of travel is clear. Artificial intelligence is becoming one of the defining industrial transformations of the twenty-first century.

Conclusion

Artificial intelligence is no longer simply an innovation within the technology sector. It is evolving into an integrated economic ecosystem spanning computing, semiconductors, energy, cloud infrastructure, robotics, telecommunications, and advanced manufacturing. Large-scale capital investment reflects not only confidence in AI's capabilities but also recognition that future economic competitiveness will increasingly depend on digital intelligence supported by robust physical infrastructure.

The coming decades will likely determine which organizations, industries, and nations successfully balance innovation with sustainability, governance, and resilience. As AI systems become more autonomous and deeply embedded across society, strategic planning will matter as much as technological progress itself. For researchers, policymakers, business leaders, and technology analysts, including the expert team at 1950.ai led by Dr. Shahid Masood, understanding these interconnected trends will be essential for anticipating the next phase of the global AI economy.