AI now writes nearly a third of Microsoft’s code, says CEO Satya Nadella

In a live discussion with Meta CEO Mark Zuckerberg, Satya Nadella revealed that AI now generates 20–30% of code in some Microsoft projects, with its role in software development continuing to grow rapidly

Artificial intelligence is rapidly becoming a co-author in software development across major tech firms. At Meta’s inaugural LlamaCon developer conference, Microsoft CEO Satya Nadella revealed that AI now writes up to 30% of the company’s code, marking a major milestone in enterprise adoption of generative tools.

Speaking in a live discussion with Meta CEO Mark Zuckerberg, Nadella said that AI contributions to coding within Microsoft’s repositories are growing. “In some projects, about 20 to 30 percent of the code is being generated by software. And the number is only going up,” he noted, underscoring AI’s expanding role in engineering workflows.

Meta, meanwhile, is developing AI models capable of building and improving the very tools that power its digital ecosystem. Although Zuckerberg declined to give specific figures, he estimated a dramatic shift is imminent. “We believe that in the next year, roughly half of software development could be AI-driven,” he said, predicting further acceleration thereafter.

AI reshaping tech workforce dynamics

The two tech giants—collectively employing tens of thousands of engineers—are not alone in embracing automation. Across the tech sector, companies are increasingly deploying AI not just to support developers, but to replace entire segments of the workflow.

Google CEO Sundar Pichai confirmed last year that over 25% of new code at Google is now AI-generated. Other firms are tightening hiring policies to reflect this shift. Shopify CEO Tobi Lütke instructed teams to justify any new hires by demonstrating the task cannot be done by AI. Similarly, Duolingo is reducing reliance on human contractors for roles that AI can take over, and future hiring will be contingent on the same principle.

This movement reflects broader labour market forecasts. A McKinsey Global Institute report projects that 30% of the US jobs could be fully automated by 2030, while Goldman Sachs estimates that up to half of all jobs may be automated by 2045.

While the efficiency and scalability benefits of AI are undeniable, the rapid integration also raises questions about job displacement, reskilling, and the future of work. As AI becomes an active creator—not just a tool—companies and employees alike face a need to adapt quickly.

For now, industry leaders seem clear: the AI age of software development isn’t coming—it’s already here.