Microsoft is actively searching for artificial intelligence startups to acquire as it seeks to develop a proprietary, cutting-edge AI model independently of its long-term partner, OpenAI. According to five people familiar with the matter, the company aims to bolster its internal AI talent through these potential acquisitions. Microsoft is currently in discussions with Inception, a startup founded in mid-2024 by a team from Stanford University that focuses on a method of developing large language models using diffusion—a technique typically applied to image and video generation. The company reportedly previously considered acquiring code-generation startup Cursor earlier this spring but backed away due to concerns regarding regulatory scrutiny.

The market for AI talent remains highly competitive, with researchers commanding compensation packages of tens of millions of dollars. Inception is reportedly seeking a valuation exceeding $1 billion and has engaged a bank to assist in deal negotiations. Microsoft’s venture fund, M12, previously invested in Inception’s $50 million seed round in late 2025. Competition for such deals is intense, with Elon Musk’s SpaceX reportedly courting Inception and having already secured a deal with Cursor shortly after Microsoft withdrew its interest.

Building a frontier-level model presents significant technical hurdles. Industry experts note that leading AI labs are currently developing models reaching roughly 10 trillion parameters—a metric of model sophistication that has surged from approximately 1 trillion parameters three years ago. While diffusion-based methods may theoretically increase processing speed by generating and refining multiple tokens simultaneously, researchers question the technology's unpredictability and its ability to scale to such massive model sizes.

Microsoft’s strategic pivot follows years of an evolving relationship with OpenAI, which began in 2019 with a $1 billion investment. By an April 29 securities filing, Microsoft had contributed $11.8 billion of its pledged $13 billion to the lab. While earlier contracts provided Microsoft exclusive access to OpenAI's technology, tensions over restrictive clauses have persisted. An amendment in late 2025 granted Microsoft the freedom to develop its own artificial general intelligence. More recently, an April agreement provided OpenAI with greater flexibility to build products with competitors such as Amazon.

Microsoft is actively searching for artificial intelligence startups to acquire as it seeks to develop a proprietary, cutting-edge AI model independently of its long-term partner, OpenAI. According to five people familiar with the matter, the company aims to bolster its internal AI talent through these potential acquisitions. Microsoft is currently in discussions with Inception, a startup founded in mid-2024 by a team from Stanford University that focuses on a method of developing large language models using diffusion—a technique typically applied to image and video generation. The company reportedly previously considered acquiring code-generation startup Cursor earlier this spring but backed away due to concerns regarding regulatory scrutiny.

The market for AI talent remains highly competitive, with researchers commanding compensation packages of tens of millions of dollars. Inception is reportedly seeking a valuation exceeding $1 billion and has engaged a bank to assist in deal negotiations. Microsoft’s venture fund, M12, previously invested in Inception’s $50 million seed round in late 2025. Competition for such deals is intense, with Elon Musk’s SpaceX reportedly courting Inception and having already secured a deal with Cursor shortly after Microsoft withdrew its interest.

Building a frontier-level model presents significant technical hurdles. Industry experts note that leading AI labs are currently developing models reaching roughly 10 trillion parameters—a metric of model sophistication that has surged from approximately 1 trillion parameters three years ago. While diffusion-based methods may theoretically increase processing speed by generating and refining multiple tokens simultaneously, researchers question the technology's unpredictability and its ability to scale to such massive model sizes.

Microsoft’s strategic pivot follows years of an evolving relationship with OpenAI, which began in 2019 with a $1 billion investment. By an April 29 securities filing, Microsoft had contributed $11.8 billion of its pledged $13 billion to the lab. While earlier contracts provided Microsoft exclusive access to OpenAI's technology, tensions over restrictive clauses have persisted. An amendment in late 2025 granted Microsoft the freedom to develop its own artificial general intelligence. More recently, an April agreement provided OpenAI with greater flexibility to build products with competitors such as Amazon.