SAP says its AI coding tool speeds up code production by 20%, but studies show experienced developers lose productivity. The productivity paradox may shift costs forward, not cut them.
SAP rolled out an AI-assisted coding tool in early 2025 that lets developers describe features in plain English and have the system generate the underlying code. The initial result: code was being produced 20% faster, the company said.
The tool is built on the principles of "vibe coding" – a term coined by OpenAI cofounder Andrej Karpathy. Vibe coding surged in popularity after Karpathy described it in early 2025. By 2026 Collins Dictionary had named it Word of the Year. Swedish startup Lovable hit $400 million in annual recurring revenue by early 2026, Bloomberg reported. A 2025 Stack Overflow survey found that 84% of respondents use or intend to use AI-assisted programming.
The productivity data tells a more complicated story. A study by Model Evaluation and Threat Research (METR) found that experienced developers suffered slowdowns in productivity when using AI tools in their work. Consulting firm Section reported that 67% of nonmanagers said AI had saved them less than two hours, if any time at all.
The contradiction – code is faster, developers aren't more productive – has a name: the vibe coding productivity paradox. The explanation, according to several analysts, is a "productivity tax" that offsets initial speed gains. AI-generated code often contains subtle bugs, hallucinations, or security vulnerabilities. The time saved writing code gets consumed debugging it. Developers who fail to audit the AI's output tend to create more issues than they fix, one Red Hat analysis noted.
SAP's own experience has been watched closely since it announced the tool. The company rolled out the assistant to internal teams first, then to customers through its Business Technology Platform. Code production sped up by a fifth, SAP said, without specifying whether those gains held after debugging and maintenance were factored in. The company has not disclosed user numbers or a breakdown of time savings versus quality issues.
Intuit, the maker of TurboTax and QuickBooks, estimated that efficiency gains from active AI usage by its engineers could reach 40%, the company told Fortune. Intuit's estimate was forward-looking and based on internal pilots, not full production deployment.
For SAP shareholders, the question is whether the tool translates into faster product releases and lower engineering costs – or into more bugs and higher maintenance expenses. SAP's Alpha Score sits at 55 out of 100, labeled Moderate, with the stock trading on a platform that tracks its profile here. SAP's full-year 2025 results, due in January, will include cloud revenue and margin data that could reveal whether the AI tool is improving operating leverage or adding cost.
A more disciplined approach to AI-augmented development is required for genuine long-run productivity gains, several analysts said. The role of the software developer must shift from writing code to overseeing it, with proper governance to catch errors before they compound. If SAP and its peers can make that transition, the vibe coding boom may deliver on its promise. If not, the industry may end up with faster code that isn't better – just more expensive to fix later.
The Collins Dictionary definition captured the cultural moment: vibe coding, it said, is "the act of writing computer code by describing what you want to happen in natural language, rather than using a programming language." The economics of that shift – whether it cuts costs or just shifts them forward in time – will determine whether the 20% speed gain SAP reported becomes a competitive advantage or a footnote.
Prepared with AlphaScala editorial tooling from the source reporting linked above. Indexable analysis may include a cited Alpha Score value. Publishing checks screen each story before release. Educational coverage, not personalized advice.