APREX · EST. 2026
For years, Microsoft defined its AI ambitions through its partnership with OpenAI. That is no longer the case. At this week's annual Build conference, Microsoft presented a range of internally developed AI initiatives that send one clear message: the company is becoming a fully independent AI player in its own right – and is ready to compete.
From dependency to independence
Microsoft has invested more than $13 billion in OpenAI since 2019, according to research sources used in this article. The technology has been the backbone of products such as Copilot and Bing. Over time, however, the relationship has grown more strained, and in October 2025 the agreement between the two companies was significantly restructured.
Perhaps the most important change came in April 2026, when the exclusivity agreement formally expired. OpenAI can now distribute its products through any cloud provider – not just Azure. Microsoft's right to be OpenAI's preferred supplier of computing capacity has been removed, and the company's intellectual property rights to OpenAI's models are no longer exclusive. The model licensing agreement runs until 2032, but without exclusivity.
"We have to prove that we can do everything we need from scratch" — Microsoft AI chief Mustafa Suleiman
Proprietary models and a new super-app
At the Build conference, Microsoft put forward what can be described as a comprehensive response to OpenAI's dominance. The company presented internally developed reasoning models, AI agents with similarities to competitors' solutions, a new cybersecurity tool, and what is being called a "super-app" – a unified platform for AI-driven workflows, according to The Verge.
The proprietary models include MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2, among others. Microsoft has also diversified its AI investments by taking stakes in companies such as Anthropic and Mistral AI – direct competitors to OpenAI.
OpenAI responds with an enterprise offensive
OpenAI, for its part, is far from passive in the new landscape. The company has launched a clear enterprise strategy for 2026, led by CEO Sam Altman. According to the research material, Altman has declared that AI is now "an application problem, not a training problem" – and that the enterprise segment is the primary value driver going forward.
In June 2026, GPT-5.6 was launched in limited preview, featuring a three-tier model structure: Sol (flagship), Terra (half the cost, GPT-5.5 performance), and Luna (speed and affordability). ChatGPT Enterprise has, according to sources, seen nearly a tenfold year-on-year increase in seats, and the company now serves more than 7 million workplace users.
What does this mean for Norwegian businesses?
Microsoft is one of the largest technology suppliers to Norwegian business and the public sector. Thousands of Norwegian organisations use Microsoft 365 and Copilot as their everyday workplace AI. The shift in Microsoft's strategy – from OpenAI-dependent to self-sufficient – could affect which models power these tools going forward.
At the same time, the new competitive landscape opens up the possibility for Norwegian businesses and developers to choose from a broader range of suppliers. The fact that OpenAI can now serve customers on cloud platforms other than Azure gives Norwegian organisations more flexibility – but also more complexity in their AI infrastructure decisions.
The capital expenditure Microsoft is planning for 2026 – estimated at around $190 billion – suggests the company is going all-in to cement its position as one of the world's leading AI players. Norwegian businesses should watch developments closely, as the products they use today could look very different in one to two years' time.