This article is exclusively available to Business Insider subscribers. If you want to dive into the latest developments in technology and business, consider subscribing.

In a striking shift within the tech industry, major companies are not only developing artificial intelligence (AI) technologies but are increasingly relying on them to automate the coding process itself. Satya Nadella, the CEO of Microsoft, revealed during a discussion at Meta's LlamaCon conference that AI is responsible for generating between 20% and 30% of the code for select Microsoft projects. This substantial percentage underscores the transformative role AI is beginning to play in software development.

Nadella elaborated that the exact proportion of AI-generated code varies depending on the programming language being used. This variation indicates that different programming environments may have differing levels of integration with AI technologies. Importantly, Nadella pointed out that Microsoft's reliance on AI extends beyond mere code generation. He noted, The agents we have for reviewing code that usage has increased, suggesting that Microsoft is seamlessly incorporating AI across the entire software development lifecycle, from writing code to reviewing its quality.

Microsoft's trend is not isolated. Other tech giants are making similar strides in leveraging AI for programming tasks. Last week, Sundar Pichai, the CEO of Google, announced during Alphabet's earnings call that AI is now responsible for generating over 30% of new code at Google, a notable increase from 25% recorded in October of the previous year. This trend reflects a growing reliance on AI across the tech sector as companies look to enhance efficiency and productivity.

During his conversation with Zuckerberg at LlamaCon, Nadella inquired about the extent of AI involvement in Meta's coding efforts. Although Zuckerberg could not provide an exact percentage, he offered an optimistic prediction for the future: Our bet is that in the next year, probably, maybe half the development is going to be done by AI as opposed to people, and then that will just increase from there. This sentiment encapsulates the belief that AI will play a significantly larger role in software development in the coming years.

Currently, Meta is already utilizing AI for specific applications, such as ad ranking and conducting feed experiments, where outcomes can be precisely measured and evaluated. This strategic use of AI showcases how even more focused applications of technology can lead to significant operational improvements.

The enthusiasm for AI in coding is not confined to the largest tech firms. Marc Benioff, CEO of Salesforce, indicated in a February earnings call that the company plans to pause new engineer hiring in 2025 due to the enhanced productivity brought about by AI, which he claimed has boosted engineering efficiency by 30%. This move highlights the broader implications of AI: while it enhances productivity, it also prompts companies to rethink their workforce needs.

In a notable development earlier this year, payments company Stripe laid off approximately 300 employees, which included several positions within engineering. This decision, first reported by Business Insider, illustrates the dual-edged sword of AI integration while productivity can be boosted, it may also lead to job reductions in certain roles.

Interestingly, the trend of AI adoption is not always synonymous with decreasing staff levels. For example, Microsoft is reportedly contemplating further job cuts, targeting middle managers and non-coding roles, rather than its coding specialists. This strategic realignment aims to increase the proportion of contributors who actively participate in code writing, effectively reducing the "PM ratio" the number of product or program managers per engineer.

Looking towards the future, Microsoft CTO Kevin Scott shared ambitious expectations regarding AI's role in software development. He indicated that within the next five years, it is anticipated that a staggering 95% of all code will be generated by AI. Very little is going to be line-by-line human-written code, Scott asserted during the 20VC podcast last month. Despite this forecast, he emphasized the continued importance of human involvement in defining the overarching structure, objectives, and design of software. This balance of human creativity and AI efficiency will likely define the future landscape of software development.