realizing

In April 2019, a small group of Openai engineers flew to Seattle to make a demo of a GPT-2 Supervitaminated Version A Bill Gates. Microsoft’s co -founder was impressed, and made clear his intention to invest in the company led by an almost absolutely unknown Sam Altman. In Redmond, Gates’ opinion took into account, and shortly after the events precipitated. The company’s cto, Kevin Scott, wrote an email Shortly after, in June, Nadella already Gates to warn of the danger of being behind. He was “very, very worried.” Nadella agreed. A month later, on July 22, 2019, Microsoft advertisement a Investment of 1,000 million dollars In OpenAi. That seemed love at first glance. That money promoted Openai’s efforts, which continued to develop its models and that in August 2020 launched an amazing GPT-3 But without opening it to the general public. Two years later, the bombing: OpenAi announced chatgpt without knowing that this would detonate the current AI fever. This divorce can be very expensive to Microsoft The initial impact was amazing, and Chatgpt became On the fastest platform he had grown up in history of the Internet. In two months he had already attracted 100 million users, and Microsoft, wanting to take another step in that idyllic relationship with Openai, redoubled its commitment to the company. Or multiplied it, because in January 2023 advertisement a multimillionaire investment that is estimated – although the figure was never detailed— at 10,000 million dollars. That investment It seemed to be perfect For both companies. Among other things, Openai obtained access to Microsoft’s cloud infrastructure to train their models and offer them to the general public (inference). Meanwhile, Microsoft gained priority access to OpenAi models, which it could sell as if they were his. That is just what they did, first with GPT-3 as a Github Copilot baseand then Freaking of other “co -drivers” that were nothing more than a chatgpt rehash. And there began the problems for Microsoft. Above all, because little by little that theoretically idyllic relationship He began teaching his seams. The convenience marriage was no longer so satisfactory for both parties. On the one hand, Openai kept asking for more and more money and better conditions when using the Microsoft computing cloud. On the other, Microsoft, OPENAI’s totally slave For his AI options, he began to look for alternatives. Both decided to look for alternative plans. Openai searched Change girlfriend and of Alliesand After the rumors In April 2025 he announced the greatest financing round in history, which amounted to 40,000 million dollars. For his part, in Microsoft, knowing their Openai absolute dependencethey began to move in March 2024, when they created their own division of AI and put in front of it To Mustafa Suleymanco -founder of Inflection AI and before Deepmind. The objective: to develop their own foundational models to avoid being chained to OpenAI. How was the thing? For now, OpenAi very well. To Microsoft, not so much. The company led by Sam Altman has not stopped growing in users and In income. At the moment they are still insufficient to make it profitable, but there is a clear thing: Nowadays Chatgpt is what Google went to searches. There are options, yes, but for the vast majority of users, (almost) do not count. I may launch of GPT-5 has been disappointingbut still Altman’s strategy to sell promises and Hype -with the Stargate megaproject in front – it works. But for Nadella and yours are much more complicated. Despite Copilot’s absolute integration in all business areas, Microsoft is almost an “coupled” in the world of AI. A pay. Your own modelsPhi-3 and Phi-4 are interesting for their “Edge AI” (artificial intelligence that runs at home, as in our mobiles), but their performance and capacity has made them almost a laboratory experiment. The gigantic infrastructure of Azure is his great asset to make them indirect leaders of AI, but Suleyman’s leadership It is committed to the simple reason that the results of the Microsoft strategy are not especially visible. It doesn’t matter if Microsoft has a powerful Trojan horse in Windows to infiltrate AI solutions in the company: who are winning that battle They are Anthropic and, of course, OpenAi. Meanwhile, Suleyman himself published on his blog An article warning of the dangers of treating AI as a person and ensuring that we are close to seeing a “apparently conscious” that can aggravate that problem. It is a valid and important argument, but it leaves Microsoft just like it was: without its own models and without changes in a strategy of doubtful success. In spite of everything, Microsoft also has its badges For Microsoft the only consolation is that its gigantic investment in OpenAi gives you the right to participate in the benefits of that company. In fact, the tense current relationship between the two adds to Openai’s intentions of becoming a profit company. That has implications for the future of the relationship, and In The Information indicated that Altman has proposed yield to Microsoft 33% of the company But renouncing future benefits. Participation is enormous and very juicy for Redmond’s, especially now that Openai is valued in 300,000 million dollarsbut there is much more at stake. In fact, in that potential divorce the one that seems to be winning is Openai, which has some of the best foundational and popular models in the world (GPT-4O, GPT-5), and who is the beautiful girl in the market: everyone wants to stick to it. Meanwhile, Microsoft is staying absolutely behind in the market, at least compared to its rivals. Let’s see: Google: It has invested a lot and well in own models (Gemini, Deepmind) and has a gigantic infrastructure, software and data. Not to mention Android and his search engine, AI entrance door for billions of users. XAI: Although less remarkable, the startup created by Elon Musk has managed to stand up with a Grok with a very different approach –zero censorship– which has managed to integrate with some … Read more