We compiled data from Gartner, McKinsey, KPMG, and the University of Oxford to examine the current state of software development costs in 2026. The data shows that 66% of enterprise software projects exceed their original budgets, with large-scale projects running 45% over budget and delivering 56% less value than projected. This analysis covers global software spending benchmarks, cost overrun rates and their primary drivers, AI’s impact on development economics, developer talent costs, and ROI benchmarks against development investment.
Global Software Development Spending: Scale and Context
Software spending has entered a period of accelerated growth in 2026, driven primarily by AI feature integration across enterprise systems. The table below presents key market-level benchmarks.
| Metric | Figure | Source |
| Worldwide software spending, 2026 | $1.44 trillion | Gartner, 2026 |
| Year-over-year growth in software spending | 15.1% | Gartner, 2026 |
| Worldwide IT spending, 2026 | $6.37 trillion | Gartner, July 2026 |
| GenAI model spending growth in enterprise software | More than doubling year-over-year | Gartner, 2026 |
| Enterprise applications featuring AI agents, end of 2026 | 40% (up from <5% in 2025) | Gartner, Aug. 2025 |
| Primary driver of software price increases, 2026 | GenAI feature integration | Gartner, Oct. 2025 |
The data shows that software is now the second-fastest growing segment of all IT spending, and most of that growth reflects cost inflation rather than new adoption. Gartner identifies GenAI feature additions as the primary driver of software price increases in 2026, meaning organisations are paying more for systems they already own. With 40% of enterprise applications projected to feature AI agents by the end of 2026, the embedded cost of software is set to continue rising regardless of whether organisations are actively pursuing new development initiatives.
Software Project Cost Overruns: The Prevailing Pattern
Cost overruns are the norm rather than the exception across enterprise software development programmes. The data below captures the breadth and magnitude of the problem.
| Metric | Figure | Source |
| Enterprise software projects exceeding original budgets | 66% | McKinsey/University of Oxford, 2012 [5,400+ projects]* |
| Large IT projects (>$15M) average cost overrun | 45% | McKinsey/University of Oxford, 2012 |
| Value delivered vs. plan for large challenged projects | 56% below projections | McKinsey/University of Oxford, 2012 |
| IT projects posing existential risk to the organisation | 17% | McKinsey/University of Oxford, 2012 |
| Budget overrun range for “black swan” projects | 200–400% | McKinsey/University of Oxford, 2012 |
| Public sector IT cost overruns vs. private sector | 3× higher | McKinsey |
*Analysis of 5,400+ IT projects conducted by McKinsey and the BT Centre for Major Programme Management, University of Oxford. See Client Note.
The data reveals that enterprise software projects carry a structural propensity for cost overrun, with two in three exceeding their original estimates. The scale of overruns is not uniform: large projects are disproportionately affected, and a subset of those programmes, approximately one in six large IT investments, become existential risks to the organisations running them. Software development is specifically identified as the highest-risk IT project category, carrying greater cost and schedule exposure than non-software IT work.
The Key Drivers of Software Development Cost Overruns
Understanding where software projects lose cost control requires looking beyond development labour, which is the most visible budget component but rarely the primary source of overruns.
| Cost Driver | Impact | Source |
| Each additional year of project duration | +15% to total cost overrun | McKinsey/University of Oxford, 2012 |
| Cloud infrastructure overspends | Enterprises spend 35% more than needed | KPMG, 2024 |
| IT budget share directed to maintenance vs. new development | 55-80% of total IT budget | Gartner |
| Schedule overruns for challenged IT projects | 46% beyond planned timeline | McKinsey |
| Engineering leaders citing AI integration as a major challenge | 77% | Gartner, May 2025 |
| Software projects vs. non-software IT categories | Carry highest cost and schedule risk | McKinsey/University of Oxford, 2012 |
The data displays that project duration and infrastructure costs are the primary multipliers of software development budgets. Each year added to a project’s timeline compounds cost overruns by 15%, and most enterprise IT budgets are already structured to favour maintenance of existing systems, with 55-80% of spending directed to sustaining what already exists rather than building new capability. Cloud infrastructure, which enterprises routinely overspend by 35%, adds a further hidden cost layer that most development scoping processes do not adequately model.
How AI Is Reshaping Software Development Cost
AI tooling is compressing some software development cost categories while shifting expenditure toward higher-value, higher-cost engineering functions.
| Metric | Figure | Source |
| Developer task completion speed with AI assistance | 55% faster | GitHub/Accenture, 2025 [4,800 developers]* |
| Time-to-market reduction with AI tools | 30% | McKinsey, 2025 |
| Routine coding tasks automated (forecast) | 70% by 2028 | Gartner, Oct. 2025 |
| Developers focused on orchestration over direct coding by end of 2026 | 75% | Gartner, Oct. 2025 |
| AI use cases purchased vs. built in-house | 76% buy vs. 24% build | a16z, June 2025 |
| GenAI model spending growth in enterprise software, 2026 | More than doubling year-over-year | Gartner, 2026 |
*Study of 4,800 software developers jointly conducted by GitHub and Accenture, 2025.
The data reveals that AI is compressing the cost of routine development work without eliminating the overall cost of building software. Labour costs tied to repetitive coding tasks are falling, but remaining development costs are concentrating in architecture, integration, and AI orchestration, functions that require senior engineering talent and carry higher per-hour costs. The 76% of AI use cases being purchased rather than built in-house reflects this dynamic commercially, as organisations redirect limited internal development capacity toward problems that vendor products cannot address.
The Developer Talent Market as a Cost Factor
Developer talent represents the largest single cost component in most software development budgets, and the market for AI-capable engineers has changed the cost structure significantly.
| Metric | Figure | Source |
| Software developers worldwide, 2025 | ~20.8 million | Stack Overflow Developer Survey, 2025 |
| Workers in roles requiring AI fluency (2025 vs. 2023) | ~7 million vs. ~1 million | McKinsey, 2025 |
| Engineering leaders citing AI integration as a major challenge | 77% | Gartner, May 2025 |
| Enterprises at risk of losing top AI talent by 2027 | 50% | Gartner, 2026 |
| Developer #1 work priority | Work-life balance (42.8%) | Stack Overflow, 2025 |
| Developer #3 work priority | Salary (32.7%) | Stack Overflow, 2025 |
The data indicates that the developer talent market has bifurcated around AI capability, creating a premium cost tier for engineers who can build and manage AI-integrated systems. The sevenfold growth in roles requiring AI fluency in two years outpaces supply, and 77% of engineering leaders already identify AI integration as a major challenge for their teams. Organisations competing for this talent face both direct cost increases and the structural risk of attrition: Gartner projects that half of enterprises without a people-centred AI strategy will lose their top AI talent to competitors by 2027.
Software Development Cost Against Return: The ROI Benchmark
The final dimension of software development cost is what it generates. The data shows a persistent gap between what projects are expected to deliver and what they actually return.
| Metric | Figure | Source |
| IT projects delivering intended value | 44% | McKinsey |
| Value shortfall for challenged projects | 39% below projections | McKinsey |
| Time-to-market improvement with AI-assisted development | 30% | McKinsey, 2025 |
| Routine coding automated, reducing delivery cost (forecast) | 70% by 2028 | Gartner, Oct. 2025 |
| Return per $1 of generative AI investment | $3.70 | IDC/Microsoft [vendor-sponsored], 2025 |
| Organisations achieving enterprise-wide financial impact from AI | ~6% | McKinsey State of AI, 2025 |
The data shows that fewer than half of software development projects deliver their intended value, meaning cost control and value delivery are distinct problems that each require active management. The strongest performance improvements come from AI-assisted development environments, where 30% faster time-to-market compresses delivery costs, though only 6% of organisations have translated that advantage into enterprise-wide financial impact. The gap between available return and realised return reflects an execution challenge more than a technology one.
To learn more or request a software development cost assessment, please reach out here.
Sources
- Gartner. “Gartner Forecasts Worldwide IT Spending to Grow 14.2% in 2026, Totaling $6.37 Trillion.” July 2026. https://www.gartner.com/en/newsroom/press-releases/2026-07-27-gartner-forecasts-worldwide-it-spending-to-grow-14-point-2-percent-in-2026-totaling-6-point-37-trillion
- Gartner. “Gartner Forecasts Worldwide IT Spending to Grow 13.5% in 2026, Totaling $6.31 Trillion.” April 2026. https://www.businesswire.com/news/home/20260422301495/en/Gartner-Forecasts-Worldwide-IT-Spending-to-Grow-13.5-in-2026-Totaling-%246.31-Trillion
- Gartner. “Gartner Unveils Top Predictions for IT Organizations and Users in 2026 and Beyond.” October 2025. https://www.gartner.com/en/newsroom/press-releases/2025-10-21-gartner-unveils-top-predictions-for-it-organizations-and-users-in-2026-and-beyond
- Gartner. “Gartner Predicts 40% of Enterprise Apps Will Feature Task-Specific AI Agents by 2026.” August 2025. https://www.gartner.com/en/newsroom/press-releases/2025-08-26-gartner-predicts-40-percent-of-enterprise-apps-will-feature-task-specific-ai-agents-by-2026-up-from-less-than-5-percent-in-2025
- Gartner. “Gartner Survey Finds 77% of Engineering Leaders Identify AI Integration in Apps as a Major Challenge.” May 2025. https://www.gartner.com/en/newsroom/press-releases/2025-05-22-gartner-survey-finds-77-percent-of-engineering-leaders-identify-ai-integration-in-apps-as-a-major-challenge
- Gartner. “Gartner Predicts by 2027, 50% of Enterprises Without a People-Centric AI Strategy Will Lose Their Top AI Talent.” May 2026. https://www.gartner.com/en/newsroom/press-releases/2026-05-13-gartner-predicts-by-2027-50-percent-of-enterprises-without-a-people-centric-ai-strategy-will-lose-their-top-ai-talent
- McKinsey & Company / University of Oxford. “Delivering Large-Scale IT Projects on Time, on Budget, and on Value.” October 2012. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/delivering-large-scale-it-projects-on-time-on-budget-and-on-value
- McKinsey & Company. “Unlocking the Potential of Public-Sector IT Projects.” https://www.mckinsey.com/industries/public-sector/our-insights/unlocking-the-potential-of-public-sector-it-projects
- KPMG. “Cloud Overspending: Cost Optimisation.” 2024. https://kpmg.com/ie/en/home/insights/2024/03/cloud-over-spending-cost-optimisation-consulting.html
- McKinsey & Company. “The State of AI: Global Survey 2025.” November 2025. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
- McKinsey & Company. “Measuring AI in Software Development.” 2025. https://www.mckinsey.com/capabilities/mckinsey-technology/our-insights/measuring-ai-in-software-development-interview-with-jellyfish-ceo-andrew-lau
- a16z. “How 100 Enterprise CIOs Are Building and Buying Gen AI in 2025.” June 2025. https://a16z.com/ai-enterprise-2025/
- Stack Overflow. “Developer Survey 2025.” 2025. https://survey.stackoverflow.co/2025/
- IDC / Microsoft. Business Opportunity of AI. 2025. [URL: to verify before publishing]
Client Note: The cost overrun statistics in Sections 2 and 3 draw primarily on McKinsey and the University of Oxford’s “Delivering Large-Scale IT Projects on Time, on Budget, and on Value” (October 2012), which analysed 5,400+ IT projects across industries and geographies. This is the most comprehensive primary study on IT project cost performance available and remains widely cited as the field’s benchmark. However, the research is from 2012, and no equivalent-scale update has been published by McKinsey or Oxford since. These figures should be presented as established benchmarks from the seminal study in this area rather than 2025–2026 primary data. The Gartner maintenance budget range (55–80% of IT budgets) is a widely attributed Gartner figure; the specific source publication should be verified before quoting publicly. The cloud overspend figure (35%) draws on KPMG’s 2024 analysis and is current.
