Bain & Company estimates that the global AI industry will need to generate approximately **$6 trillion** in annual revenue by 2031 to justify the massive investments in data centers and computing infrastructure. This figure is significantly higher than what current AI applications and employee productivity gains can provide. According to the firm, most of this revenue will have to come from new advancements, such as autonomous systems, robotics, and medical technologies. Currently, profits are largely funded by investors rather than generated by customers, which raises concerns about the potential bursting of the AI bubble. In a recent report, economist Torsten Slok highlighted a financial imbalance within the AI industry. He pointed out that profits from chip manufacturers are primarily funded by investor capital rather than actual customer revenues. Developers of AI applications are showing negative operating margins, creating a fragile business model that relies on "the hope of future profitability." If end customers do not quickly generate concrete returns on investment, the massive funding could suddenly dry up, causing the AI bubble to burst. Some companies are already facing negative cash flow. OpenAI, often seen as the global leader in the sector, projects a cash flow loss of nearly **$280 billion** by the end of 2030; it will have exhausted investor funds by 2028. Torsten Slok is not the first economist to sound the alarm about the industry's excessive dependence on investments. In its annual report published this summer, the Bank for International Settlements (BIS) noted that the massive influx of investments in AI exceeds the benefits and available cash flows, leading these companies to issue debt instruments to raise additional funds. According to a recent report by Bain & Company, the global AI industry will need to generate approximately **$6 trillion** in annual revenue by 2031 to justify the massive investments made in data centers and computing infrastructure. The report suggests that the industry needs breakthrough innovations capable of generating sufficient revenue, approximately **$6 trillion** per year, for the current bets of investors to become profitable. However, this is far from certain. To achieve this goal, the firm estimates that the sector will have to expand well beyond traditional enterprise software. "The current debate focuses exclusively on employee productivity. The AI infrastructure economy requires thousands of billions of dollars in new revenues, beyond productivity gains," said David Crawford, president of the Global Technology division at Bain. The report indicates that existing AI services for individuals and businesses could contribute up to **$1.8 trillion** to this revenue figure, leaving a gap of approximately **$4.2 trillion** that must be filled by new AI services and applications. These additional revenues could come from areas such as autonomous machines, robotics, drug discovery, and physical AI. "What the sector needs is a wave of innovation that will far outstrip what mobile telephony and the cloud have enabled," added David Crawford. He noted that the development of AI infrastructure currently far exceeds the demand curve, adding that maintaining these investments would require AI to contribute approximately 1 percentage point to the annual growth of global GDP. Microsoft, Google, Amazon, Meta, and Oracle are investing hundreds of billions of dollars in data centers and computing infrastructure while engaging in a frenzied race to meet demand for AI-related workloads. Bain & Company predicts that "global spending on data centers will reach between **$5 trillion** and **$6.5 trillion** by the end of this decade." This will represent an increase in capacity of at least 150 gigawatts. According to the firm, this expansion will exert additional pressure on the electricity supply and other resources. The cost and scale of data centers are also increasing rapidly, doubling approximately every 12 to 16 months. Furthermore, costs could continue to rise due to memory shortages. The rise in prices for chips, particularly those provided by Nvidia and SK Hynix, as well as network equipment and other components, contributes to this increase. According to the consulting firm, annual spending on AI infrastructure, including data centers, computing capacity, and upgrades to accelerators and memory chips, could reach **$1.5 trillion** by 2031. For example, Meta's Prometheus data center in Ohio, valued at **$24 billion** for 600 megawatts in 2025, could reach a capacity of 9 gigawatts for an estimated cost of **$200 billion** by 2030. This frenzied expansion intensifies pressure on the electricity supply and will encourage major cloud players to develop their own custom chips to reduce expenses. AI is also rapidly transforming the cybersecurity landscape. According to Bain & Company's estimates, "AI has reduced the duration of a typical cyberattack from about four weeks to 18 hours." The increasing use of AI agents adds an additional level of risk, while existing approaches to software and AI vendor monitoring struggle to keep up with the frequent changes. The survey conducted by Bain among CISOs revealed that companies are increasing their remediation budgets by double-digit percentages. Some are also reallocating 20 to 25% of their cybersecurity personnel dedicated to managing alerts generated by AI-based analyses. The report identifies legacy platforms, network layers, and software-as-a-service (SaaS) providers among the main areas on which companies are focusing their compliance efforts. In its report, Bain & Company also identifies the speed of AI adoption, that is, the rapidity with which companies can deploy AI, as a new competitive factor. The main AI players are investing up to **$9.75 billion** in proactive engineering projects to help companies integrate AI into their operations. Suppliers are developing application and infrastructure layers that connect AI models to enterprise systems. Despite the drop in token prices, Bain does not expect large language models to follow the traditional path of commoditization. The report also reveals that executives expect AI to enable a 148% improvement in deployment cycle speed and a 95% increase in software developer productivity over the next one to two years. Current productivity gains range between 20% and 27%. However, testimonials suggest that AI adoption is occurring at the expense of developers' well-being. According to the firm, gains related to coding acceleration could create new bottlenecks in review, coordination, software quality, and governance, thus requiring changes at the level of the overall software development lifecycle.