HSBC analysts argue that enterprise software will remain the primary vehicle for AI adoption among major corporations, despite rapid advances in foundation models. The bank contends that 2026 marks the beginning of significant monetization within the software sector as vendors embed AI capabilities into their existing platforms.

Analyst Stephen Bersey emphasized that foundation models are "inherently flawed" for enterprise applications, lacking the reliability and error-free operation that major corporations require. While these models may excel at narrow tasks like image generation, they fall short for complex enterprise platforms that demand high fidelity and consistent performance. Major AI model developers lack experience creating enterprise-grade software, forcing them to start from scratch in highly complex environments.

HSBC believes established software vendors are best positioned to capitalize on AI's enterprise potential because they already possess critical intellectual property and run core operations for global companies. The bank argues that incumbent platforms have evolved to achieve near-perfect reliability with high throughput, making them difficult to displace even if competitors could match their code quality. As monetization accelerates, these vendors are embedding "distilled intelligent agents" into their platforms in controlled ways that overcome the limitations of foundation models, positioning them to benefit most from the AI revolution.