
Legacy Software as an Innovation Tax on Enterprise AI

Google Cloud’s latest enterprise survey, conducted with Hypothesis Group among 1,250 U.S. IT and business decision-makers, examines how existing technology environments affect AI adoption. The central finding is a paradox: the legacy software that enterprises rely on to deliver technology to employees is simultaneously functioning as an innovation tax, holding back AI transformation. The survey defines legacy technology as software running locally on computers and servers rather than natively in the cloud. Nearly half (49%) of respondents report that relying on on-premises software has actively delayed, reduced in scope, paused, or canceled AI or other technology initiatives.
The operational toll of legacy IT is concrete. Routine patch application and maintenance emerged as a primary bottleneck, especially as AI-enabled vulnerability discovery accelerates the rate at which flaws are surfaced in legacy codebases. Human security teams cannot keep pace, creating a zero-sum game between maintaining legacy systems and investing in AI modernization. IT teams spend their time and resources on upkeep rather than directing engineering cycles toward building and scaling AI systems.
The research also identifies vendor lock-in, rather than architectural preference, as a key reason organizations remain on legacy environments. Deep provider dependence introduces commercial constraints: 39% of decision-makers report that reliance on legacy software and cloud providers forces them to navigate complex licensing rules or contracts, or limits their negotiating flexibility. The survey warns of systemic risk from single-vendor concentration. When operating systems, productivity applications, cloud hosting, and AI models all come from one vendor, the result is a single point of failure. An autonomous AI agent is only as secure as the operating system and network it traverses; a vulnerable underlying stack does not mean resilient intelligence, but an expanded attack surface.
The proposed solution centers on customer choice and interoperability. 80% of leaders agree that a provider’s installed base should not determine which AI technology an organization uses, and 78% believe that the providers who led the previous era of enterprise technology may not be best positioned to lead the AI era. Organizations want freedom to choose best-of-breed AI on its own merits, not bundled defaults tied to legacy desktop software.
The article outlines several recommendations for accelerating agentic technology. First, demand architectural and provider choice: AI and cloud should not be held hostage by legacy software running local servers, email, or desktop operating systems. Second, commit to secure-by-design foundations: move past the broken cycle of releasing insecure software and patching after the fact, and engineer durable architectures where entire vulnerability classes are eliminated before code ships. Third, deploy AI-driven defense to remediate legacy debt: for vulnerable legacy software already in production, shift from manual triage to proactive, AI-assisted defense using platforms like Security Command Center and Mandiant threat analytics to surface active exploit paths, prioritize critical exposure, and automate remediation across hybrid environments.
The article concludes with a framing question for IT leaders: whether their technology foundation accelerates the future or whether they are paying a maintenance tax just to keep the past running. While the survey data is U.S.-only and focuses on decision-maker perceptions, the findings are positioned as evidence of a systemic challenge across the enterprise landscape as organizations move from integrating AI and LLMs to powering autonomous, agentic workflows. Nearly 90% of the Fortune 100 are cited as already using Gemini Enterprise to build, scale, and govern agents, underscoring the scale at which agentic AI is entering production despite the drag of legacy infrastructure.

