AI Coding Agents Push One-Third of Companies to Skip Software Purchases, McKinsey Finds

The build-vs-buy question has been effectively settled for most of the past decade: you buy. A new McKinsey survey suggests that settlement is coming unstuck.

<cite index="22-5,22-6">McKinsey published the 2026 edition of its State of AI survey on August 25, and one finding traveled faster than the rest: nearly a third of respondents report their organizations have decided against buying one or more software products or features because they could be built internally with agentic coding tools.</cite> <cite index="22-7">The survey drew 1,719 responses across 97 countries between May 4 and June 8, weighted by each country's share of global GDP.</cite>

The 32% figure is the headline, but the more interesting number sits right next to it. <cite index="23-5,23-6">In the same survey, the share of organizations reporting that AI contributed to their EBIT sat at 37%, unchanged from a year earlier. Companies are building more and buying less, and the financial results have not moved.</cite> That's the tension software vendors should be reading carefully, and also the tension their potential customers should be reading carefully before canceling a renewal.

The shift isn't uniform across the industry. <cite index="18-4,18-5">It's most pronounced among what McKinsey identifies as 'high performers,' the 6% of respondents who attribute at least 5% of their EBIT to AI. Nearly half of these high performers are skipping software purchases, compared to 31% of their peers.</cite> <cite index="19-5">Forty percent of respondents at organizations with more than $1 billion in annual revenue said they were scaling AI agents in at least one function, up from 27% a year earlier.</cite>

The risk is what happens after the sprint. <cite index="23-13">Coding agents compress the initial development phase, which is roughly 30% of a system's lifetime cost, and do almost nothing to the remaining 70% spent on maintenance, security, integration drift and eventual replacement.</cite> That math doesn't show up in a procurement comparison until well into year two. <cite index="23-14,23-15">There's also a new cost that most build models omit: AI operating and inference expense. McKinsey's survey found one in five organizations already limiting AI use because of operating costs, with the heaviest builders hitting cost constraints on coding agents about three times as often as everyone else.</cite>

In regulated industries, the calculus is harder still. <cite index="24-8,24-9">An insurer running AI in underwriting, rating, claims or fraud now operates under the NAIC Model Bulletin, adopted by 24 states plus DC as of early 2026. It requires a written AI Systems Program with board accountability, documented validation and retesting, and contractual audit rights over third-party AI, because regulators have signaled they will 'look through' vendor relationships during examinations.</cite> What that means in practice: when you build it yourself, you also own the audit trail. Buying from a vendor used to mean buying someone else's compliance posture. That's not gone, but it's now a negotiation.

The aggregate software market hasn't registered the shift yet. <cite index="27-4">Gartner's July 2026 forecast puts worldwide software spending at $1.468 trillion for the year, up 15.5% from $1.271 trillion in 2025, an acceleration on 2025's 13.9% growth, not a contraction.</cite> That matters. One-third of organizations passing on a single purchase is not the same as one-third of software revenue evaporating. What it does signal is a change in leverage at the negotiating table. Vendors who can't answer 'why can't my team build this in a sprint?' are going to have a harder time on renewals.

The missing data point in McKinsey's findings, as reported by ANI, is deployment rate. <cite index="18-10,18-11">While Forrester reported in June 2026 that roughly 75% of organizations are adopting agentic AI, only a small minority have reached meaningful production. The Gartner CIO Survey 2026 found that only 17% of organizations have actually deployed agents.</cite> Deciding to build and shipping something that runs in production are two different things. Until that gap closes, the 32% figure measures intent as much as execution.

Sources cited:
- McKinsey State of AI 2026 (via Digital Applied) (https://www.digitalapplied.com/blog/a-third-of-companies-skipped-buying-software-and-built-it)
- ANI / Tribune India (https://www.tribuneindia.com/news/business/ai-coding-agents-threaten-to-reshape-software-spending-as-companies-choose-to-build-rather-than-buy-mckinsey/)
- The D*AI*LY Brief / BERI (https://www.beri.net/article/mckinsey-state-of-ai-2026-agentic-coding-build-vs-buy-run-cost)
- Nerd Level Tech (https://nerdleveltech.com/coding-agents-saas-build-vs-buy-shift)
- AI Agent Store / AI Agents News (https://aiagentstore.ai/ai-agent-news/this-week)