Microsoft is intensifying its search for artificial-intelligence startups as the technology giant moves to reduce its dependence on longtime partner OpenAI and position itself to compete more directly at the frontier of advanced AI model development, Reuters reports, citing people familiar with the matter.
The acquisition discussions reveal a major shift inside Microsoft, which for years relied heavily on OpenAI’s technology to power its AI ambitions across products, including Azure, Office, and GitHub Copilot.
Now, after investing more than $13 billion into OpenAI and helping transform the startup into one of the most valuable companies in the world, Microsoft is increasingly preparing for a future in which it may need to stand on its own in the race to build cutting-edge artificial intelligence systems.
Register for Tekedia Mini-MBA edition 20 (June 8 – Sept 5, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
According to people familiar with the matter, Microsoft is exploring deals that could help it secure elite AI talent, gain access to new model architectures, and accelerate efforts to develop a world-class proprietary foundation model by next year.
The company’s interest comes as competition for AI startups and researchers reaches unprecedented levels across Silicon Valley, with major technology firms spending billions of dollars to secure scarce expertise and computing advantages.
Among the companies Microsoft considered acquiring this year was Cursor, the rapidly growing code-generation startup whose AI-assisted programming tools have become popular among software developers. People familiar with the discussions said Microsoft ultimately backed away from pursuing a deal because of internal concerns that regulators would likely scrutinize the acquisition heavily, given Microsoft’s ownership of GitHub and its AI coding platform GitHub Copilot.
Regulators in the United States and Europe have intensified scrutiny of large technology companies seeking to consolidate power in artificial intelligence, particularly in areas involving developer tools, cloud infrastructure, and foundation models.
Microsoft’s dominance in enterprise software and cloud computing already places the company under close regulatory watch, making major AI acquisitions more politically sensitive.
Still, Microsoft continues to pursue other opportunities. The company is currently in discussions with Inception, a small AI startup founded by researchers from Stanford University that is developing an alternative approach to large language models, according to people familiar with the talks.
Microsoft’s venture arm, M12, previously participated in Inception’s $50 million seed funding round in late 2025. The discussions remain ongoing and may not result in a transaction. Inception has drawn interest because of its work on diffusion-based language models, an emerging AI architecture that differs significantly from the autoregressive systems used by most leading AI developers today.
Traditional large language models generate text sequentially, producing one token at a time. Diffusion models instead generate and refine multiple tokens simultaneously, a technique more commonly associated with AI image and video generation.
Researchers believe the approach could dramatically improve inference speed and computational efficiency, two of the biggest constraints facing the AI industry as models become larger and more expensive to operate.
However, the technology remains experimental.
AI researchers caution that diffusion-based language systems can behave unpredictably, and it remains unclear whether they can scale effectively to frontier-level models containing trillions of parameters.
Microsoft’s Post-OpenAI Strategy Emerges
For years, the company’s partnership with OpenAI defined its position in the generative AI race. Microsoft first invested $1 billion into OpenAI in 2019 when the research lab remained relatively obscure. That partnership later became one of the most consequential alliances in modern technology after OpenAI launched ChatGPT in late 2022, triggering the global AI boom.
The relationship gave Microsoft privileged access to OpenAI’s models while helping drive enormous growth for its Azure cloud business. According to a securities filing released in April, Microsoft has already provided $11.8 billion of its promised $13 billion investment into OpenAI.
The company’s broader AI spending has become even larger.
Michael Wetter, who oversees Microsoft’s corporate development operations, testified in court Wednesday that Microsoft has spent more than $100 billion on OpenAI investments as well as the infrastructure and hosting systems required to support AI development. Those expenditures underscore the staggering financial scale of the AI arms race now unfolding across the technology sector.
Yet the Microsoft-OpenAI relationship has also become increasingly complicated. The original partnership granted Microsoft exclusive access to OpenAI technology while giving OpenAI guaranteed access to enormous computing resources for research and model training.
Over time, however, tensions reportedly emerged as OpenAI’s ambitions expanded. OpenAI executives concluded that their infrastructure needs were growing faster than Microsoft could supply, according to people familiar with the relationship. At the same time, Microsoft faced contractual restrictions that limited its ability to build foundation models competing directly with OpenAI’s offerings.
The companies have repeatedly renegotiated their agreements over the years as the balance of power shifted. One major change came in late 2025 when the partnership was amended to allow Microsoft to pursue development of artificial general intelligence, or AGI, a still-theoretical form of AI capable of outperforming humans across a broad range of complex cognitive tasks.
Another revision arrived in late April, when Microsoft and OpenAI reached an agreement allowing OpenAI greater flexibility to build products with rival cloud providers, including Amazon.
Those changes signaled a gradual loosening of what had once been one of Silicon Valley’s most tightly integrated AI partnerships.
The New AI Talent War
Microsoft’s acquisition search also underpins the escalating battle for elite AI talent. Compensation packages for top researchers now routinely reach tens of millions of dollars, while startup valuations are soaring as investors rush to secure exposure to promising AI technologies.
The market has become so competitive that even Microsoft is facing mounting pressure from rivals.
People familiar with the matter said Elon Musk’s SpaceX has emerged as one of Microsoft’s biggest competitors for AI startup deals following SpaceX’s acquisition of Musk’s AI company xAI earlier this year.
Shortly after Microsoft abandoned talks involving Cursor, SpaceX announced a partnership with the startup. SpaceX also reportedly courted Inception, according to people.
One source said Inception recently hired an investment bank to assist with negotiations and is seeking a valuation exceeding $1 billion.
The bidding competition shows that technology companies are no longer merely buying products or intellectual property. Increasingly, they are attempting to secure scarce clusters of researchers capable of designing next-generation AI architectures before competitors do.
Building an Independent AI Stack
Microsoft’s acquisition push is also believed to be an indication of a growing realization across the industry that long-term AI leadership may require tighter vertical integration. Companies increasingly want control over models, infrastructure, cloud services, and developer ecosystems simultaneously.
That helps explain why Microsoft is investing so aggressively in internal AI development alongside its OpenAI partnership.
Part of that effort is being led by teams under Mustafa Suleyman, the co-founder of DeepMind, who joined Microsoft last year to strengthen its consumer AI operations and long-term research capabilities.
The company’s broader strategy appears aimed at reducing reliance on any single external AI provider while preserving flexibility in an industry evolving at extraordinary speed. That transition is becoming increasingly important as frontier AI systems grow more expensive and strategically important.
Some leading AI laboratories are now building models containing roughly 10 trillion parameters, according to researchers, up from about 1 trillion parameters only three years ago. Those enormous systems require unprecedented levels of computing infrastructure, engineering talent, and capital investment.
For Microsoft, the stakes extend beyond competition with OpenAI. Artificial intelligence is rapidly becoming central to cloud computing, enterprise software, search, cybersecurity, and productivity applications, all areas critical to Microsoft’s long-term growth.
The company, therefore, appears determined to ensure it remains one of the industry’s dominant AI platforms even if its relationship with OpenAI eventually becomes less central to that ambition.