AI infrastructure will require a $6 trillion annual market by 2031, Bain warns

Consulting firm Bain & Company estimates that the artificial intelligence infrastructure sector will need to generate $6 trillion in annual revenue by 2031 to maintain its growth trajectory. This calculation assumes that investments in equipment and data centers will account for 25% of total sales during this period.

The 2026 technology report, released on Tuesday, October 29, starts with a projected annual investment base of $1.5 trillion for that year. The $6 trillion figure represents the annual revenue needed to sustain growth, not the total earnings accumulated by the end of the decade.

Beyond the massive cost of building the infrastructure, the challenge lies in finding customers willing to pay for it. After acquiring servers and building processing centers, companies must transform this capacity into profitable businesses that justify the investment.

Current applications are insufficient to cover the deficit

According to Bain’s projections, revenue from subscriptions and advertising in consumer products could reach between $200 billion and $400 billion by 2031. Corporate solutions would add between $1 trillion and $1.4 trillion to technology suppliers’ revenues.

Even at the upper end of these estimates, there would still be a $4.2 trillion shortfall to meet the required level. To bridge this gap, the study highlights sectors such as autonomous vehicles, robotics, drug discovery, and the development of new materials.

The difference is economic: when a company saves resources through AI, not all of those savings translate into revenue for the technology provider. If a tool reduces the time needed to complete a task, the customer retains part of the benefit. For the seller, revenue depends on the price charged and the buyer’s willingness to pay.

Who will bear the cost of the infrastructure built?

This dilemma is directly tied to the debate about how to finance AI infrastructure. Building data centers and purchasing equipment require significant upfront investments before it is known what returns the services sold on that infrastructure will generate.

The scenario outlined by Bain does not claim that the investment will fail, but identifies the conditions necessary to keep it viable. The efficiency of systems, pricing, and equipment utilization rates could alter this equation.

In today’s race, announcing computing capacity is no longer enough. The decisive factor will be the value customers are willing to pay to use it. An infrastructure capable of generating intelligence must also find demand willing to cover its costs.

Frequently Asked Questions

Will the $6 trillion be generated by 2031?

No. Bain estimates that this amount represents the annual revenue needed in 2031 to sustain the projected investments.

Is this revenue guaranteed?

No. This is an estimate based on the report’s assumptions, not a forecast of guaranteed revenue.

Why isn’t the economic benefit generated by AI enough to pay for the infrastructure?

The benefit obtained by the customer is not fully transferred to the supplier. Converting it into revenue depends on charging for the service and the customer’s willingness to pay.