Google Cloud's DORA team has published an updated report, the ROI of AI-Assisted Software Development (2026.01) offering a practical framework for calculating the financial return on AI investment in software development. The report sets out a structured model for translating engineering metrics into business value. It is a followup to the 2025 DORA State of AI-Assisted Software Development report and was authored by a team from Google Cloud's DORA group and its delta innovation practice.
The report's central argument is that AI acts as an amplifier. "The greatest returns on AI investment come not from the tools themselves but from a strategic focus on the underlying organizational system: the quality of the internal platform, the clarity of workflows, and the alignment of teams," writes Nathen Harvey, the DORA team lead at Google Cloud. "Without this foundation, AI creates localized pockets of productivity that are often lost in downstream chaos." This framing directly echoes the 2025 DORA research, which found that AI magnifies the strengths of high-performing organisations and the dysfunctions of struggling ones.
A key idea in the report is the J-Curve of value realisation. The authors argue that most organisations will experience a temporary productivity dip before achieving long-term gains from AI adoption. This dip has three main causes: the learning curve as teams adapt their workflows, the verification tax imposed by reviewing AI-generated code, and the need to adapt downstream processes such as testing and change approval to handle increased code volumes. The report describes this period as "the tuition cost of transformation" and argues that leaders who misread it as failure risk pulling funding during the dip and losing the eventual return.
The report's methodology for computing ROI is built on a value model drawn from Google Cloud's Value Realisation practice. Value flows from AI adoption through a set of seven capabilities; including a quality internal platform, version control practices, and AI-accessible internal data. This flows into improved DORA delivery metrics, then into non-financial outcomes such as developer experience and user experience, and finally into financial outcomes: cost savings and revenue growth. The ROI is calculated using the standard formula: value minus investment, divided by investment. Using illustrative figures for a 500-person engineering organisation with a fully loaded salary of $176,000 per head, the report models a first-year return of approximately $11.6 million against an investment of $8.4 million, yielding a 39% ROI and a payback period of around eight months.
The report is careful not to overstate these figures. "Treat these calculations as a high-uncertainty estimate meant to spark a conversation, rather than a rigid mathematical formula," the authors write. They note that inference costs for AI models have fallen dramatically; dropping by a factor of 280 between November 2022 and October 2024 according to the Stanford Artificial Intelligence Index. This means that the true financial burden of adoption has shifted to governance: managing the verification tax, adjusting workflows, and upskilling staff.
"We don't measure AI by the code it writes but by the bottlenecks it clears."
- DORA team, Google Cloud -- DORA ROI of AI-Assisted Software Development report
The report also highlights the instability tax. Drawing on the 2025 DORA research, it notes that while AI adoption is associated with increased individual effectiveness and code quality, it is also associated with a rise in software delivery instability. More code moving faster can overwhelm existing deployment pipelines and manual review gates. The model accounts for this as a cost, and the sample calculator actually shows a negative downtime impact of $344,000 because the assumed change failure rate rises from 5% to 6% after AI adoption. The authors present this not as a reason to delay adoption, but as a reason to invest in automated testing, continuous integration, and working in small batches.
InfoQ has previously covered the evolution of DORA's AI research. In September 2025, InfoQ reported on the 2025 DORA State of AI-Assisted Software Development, which introduced the DORA AI Capabilities Model and identified the seven distinct team archetypes ranging from high-achievers to organisations stuck in a legacy bottleneck. A March 2026 InfoQ analysis of that report noted that the research drew on surveys from nearly 5,000 technology professionals and over 100 hours of qualitative interviews, and concluded that "AI will not fix broken engineering systems."
The new ROI report complements that earlier work by giving engineering leaders a concrete financial toolkit. It includes an interactive calculator that organisations can use to adjust the assumptions to their own context. The authors recommend running three scenarios: conservative, realistic, and optimistic, to build a range of outcomes and set appropriate expectations with finance teams.
Community reactions to the report have been broadly supportive of its framing. Karol Wojtaszek noted on LinkedIn that the report addresses the real question executives are asking about AI spending. Andreas Wiesmueller wrote on LinkedIn that "AI without engineering excellence just scales your problems," a view closely aligned with the report's emphasis on organisational foundations. Writing in a LinkedIn article on why AI ROI requires process redesign, Ravi Kalakota, a technology strategist based in New York and not affiliated with Google, argued that "real ROI doesn't come from the Large Language Model; it comes from the Model Redesign," adding that "deploying AI without operational redesign is just an expensive, high-speed way to stay the same."
This tension between tool adoption and organisational readiness is not new territory for DORA. The same dynamic has appeared in prior DORA research on continuous delivery and platform engineering, both of which showed initial productivity dips before long-term gains. The report draws an explicit parallel, noting that the J-Curve "consistently" appears across these technical disciplines. Research cited in the report from Stanford University's Software Engineering Productivity programme found that while AI yields a 35 to 40% productivity gain on simple, greenfield tasks, its impact on complex legacy code is often 10% or less; a finding that matters for the many organisations working primarily on existing systems.
The report also addresses what it calls the "agentic era," describing a shift from reactive AI tools to autonomous systems capable of executing multi-step workflows. In this context, the authors reframe how ROI should be understood.The report strongly discourages headcount reduction as a strategy, arguing that retaining and training existing staff is more cost-effective and preserves institutional knowledge.
"Return on investment is no longer a measure of how many developers an organization can replace. It is a measure of how much latent human creativity can be unlocked by offloading systemic toil to these autonomous agents."
-DORA team, Google Cloud -- DORA ROI of AI-Assisted Software Development report
For the longer term, the authors point to Google Cloud data showing an average 727% return on investment in Google Cloud AI over three years, and an average payback period of around eight months. They frame the first year as primarily a period of foundation building and organisational change, with compounding gains in years two and three as teams move from simple coding assistants to agentic workflows at scale. The report is available at dora.dev/ai/roi/report.