Microsoft aims for AI self-sufficiency: its own models instead of relying on OpenAI

Microsoft aims to become fully independent in artificial intelligence

The head of the company's AI department—Mustafa Suleyman—told the *Financial Times* that Microsoft intends to move toward “true self-sufficiency” in developing its own AI models. This decision emerged after restructuring relations with OpenAI in October last year, when the giant announced a desire to create cutting‑edge solutions independently without relying on a partner.

Key points of the strategy
What Microsoft is doing | Why it matters
Developing its own models | Reducing dependence on OpenAI and controlling the technology stack.
Investments in infrastructure | Providing gigawatt-scale computing power for training and operating AI systems.
Collecting large datasets | Essential material for training “first‑class” models.

> “We must develop our own foundational cutting‑edge models, equipped with gigawatt-scale computing power and some of the best AI training teams in the world,” Suleyman emphasized.

Relations with OpenAI
* Microsoft previously used OpenAI models for its Copilot services and other products.
* Last year the company agreed to a startup restructuring, retaining a stake valued at $135 billion and access to models until 2032.
* As a result, OpenAI gained greater freedom in choosing infrastructure partners and became a direct competitor to Microsoft.

Investments in third‑party developers
Microsoft has invested in Anthropic and Mistral, accelerating its own work on AI models slated for release this year. This strengthens competition with OpenAI and Google, which are actively promoting enterprise solutions.

Impact on workflows
Suleyman predicts that within the next 12–18 months AI will fully automate most office tasks—from lawyers to marketers. In two or three years, AI agents will be able to coordinate processes in large organizations and self‑improve.

> “Creating a new model will be about as simple as recording a podcast or posting on a blog,” he noted, emphasizing the ease of developing adaptive solutions for any organization.

Financial aspects
* Microsoft’s capital expenditures for the current fiscal year (ending June) will total $140 billion.
* Investors remain skeptical: the company’s shares fell 13 % in the last month.

> “We are all confident that these results will accumulate and increase revenue with profit,” Suleyman assured.

Other AI projects
* Medical superintelligence—support for healthcare systems.
* Humanistic superintelligence—technologies that will not resist developers’ will and remain under human control.

> “These tools are intended to enhance human well‑being and serve humanity, not surpass it,” the Microsoft AI head concluded.