More than half of the software currently used by enterprises could be replaced by AI-driven systems, according to Arthur Mensch, chief executive of Mistral AI. The comments, made in an interview with CNBC, come as investors grow increasingly concerned that artificial intelligence may erode demand for traditional enterprise software.
Mensch said AI is enabling companies to build custom applications at unprecedented speed, reducing the need for off-the-shelf software-as-a-service products. “More than half of what’s currently being bought by IT in terms of SaaS is going to shift to AI,” he said, adding that AI allows software to be developed “at the speed of light.”
The remarks follow a sell-off in major software stocks, partly triggered by growing investor anxiety over AI-native tools such as Anthropic’s Cowork product. The iShares Expanded Tech-Software Sector ETF, which includes companies such as Microsoft and Salesforce, is down more than 20% this year. Software stocks in India, including Tata Consultancy Services and Infosys, have also declined.
AI-Driven Replatforming
Mensch said enterprises with the right infrastructure can now connect internal data directly to AI systems to create applications that automate key workflows. Tasks such as procurement or supply chain management, which once required specialized vertical SaaS products, can now be handled through custom AI applications built in days rather than years.
He described the shift as a broader “replatforming” of enterprise IT, with businesses reassessing software purchased decades ago that has become costly to maintain. According to Mensch, AI offers a path to rebuild those systems in a more efficient and less expensive way.
Mistral AI is seeing growing interest from large organizations pursuing this transition. Mensch said the company now works with more than 100 enterprise customers that are actively considering replacing parts of their existing software stack with AI-driven alternatives.
Not all software categories face the same level of disruption. Mensch said systems-of-record software, which manages core enterprise data, is unlikely to be replaced and will instead work alongside AI systems. That view aligns with comments from Bipul Sinha, chief executive of Rubrik, who told CNBC that workflow software could be significantly disrupted, while data infrastructure software enabling AI would benefit.
Expansion in India
Mensch also said Mistral plans to open its first office in India this year, expanding into a market where global technology companies are racing to establish a presence. The company already works with multinational customers that operate in India and is now targeting local public and private sector organizations.
While Mistral is building its own data centers in Europe, Mensch said it plans to partner with local infrastructure providers in India. The strategy reflects government pressure for AI models that can run locally and keep data stored domestically.
India’s linguistic diversity is also a factor. Mensch said Mistral’s large language models can support languages such as Hindi and Punjabi, which he said will become increasingly important for consumer-facing AI applications in the country.
The comments highlight how advances in generative AI are reshaping expectations for enterprise software and intensifying scrutiny of long-established SaaS business models.
