Here’s what’s on our plate today:

• 🧪 Microsoft is hedging its OpenAI bets with a multi-model AI strategy.

• 🧠 Google sued over AI, China hits Nvidia, Big Tech builds a translator.

• 🧰 Brainsnack for Builders: A no-code tool to generate app screenshots.

• 📊 Poll: How do you feel about Microsoft’s Anthropic move?

Let’s dive in. No floaties needed…

The Laboratory

Inside Microsoft’s multi-model AI strategy

When examining the current smartphone market, one finds numerous hardware options. However, when it comes to the software, the market is dominated by two platforms, Apple’s iOS and Google’s Android operating systems. While there may be many iterations of Android, with manufacturers developing their skins on top of the base Android, a third platform is sorely missing from the picture. This duopoly in the smartphone OS space was not created overnight; it is the result of decades of competition between companies that once included names like Nokia’s Symbian and Microsoft’s Windows.

Whenever there is a lack of competition in a market, users are left with little choice but to settle for whoever dominates. In this context, when one looks at the market for artificial intelligence, it is still the early days. Several companies are working on frontier models, and all of them are competing to ensure not just dominance but also survival. OpenAI may have the upper hand now, but there’s no guarantee it will keep building the biggest and most powerful AI models.

A recent decision by Microsoft to use some AI from Anthropic, thereby reducing its reliance on OpenAI, reflects not only the need for competition but also how competition in business influences company relationships. The shift to Anthropic is a strategic move and reflective of how big tech operates to ensure its continued viability. Microsoft and OpenAI have a history dating back to 2019, when OpenAI was relatively unknown to the larger public.

Microsoft’s early bet on OpenAI

OpenAI was launched in 2015 as a nonprofit research lab, explicitly promising “to pursue AI most likely to benefit humanity”, unconstrained by investor returns. As the company worked on improving its capabilities, it created OpenAI LP, a ‘capped‑profit’ subsidiary controlled by the nonprofit to raise capital for its frontier AI models.

In 2019, Microsoft invested $1 billion and became OpenAI’s exclusive cloud partner to build Azure supercomputers for OpenAI’s models. This set the technical and commercial foundation for everything that followed.

In 2023, the partnership deepened when Microsoft announced a further multibillion-dollar investment in OpenAI. The deal allowed OpenAI to continue scaling, while Microsoft leveraged the tech to revamp its products, including search engine Bing. At the time, Microsoft hoped to counter Google in the search business, while also ensuring it could rival its upcoming AI products.

On the surface, things seemed to be going great for both companies. OpenAI was scaling faster than rivals, and Microsoft was able to leverage the tech to introduce features in its own products. However, things hit a speed bump in November 2023 when OpenAI’s board abruptly ousted CEO Sam Altman and named Mira Murati interim CEO.

What followed was a shift in the power dynamics between the two companies.

Power struggle after Altman’s ouster

After Sam Altman’s ouster, OpenAI's staff threatened to quit the AI startup and join former boss Sam Altman at Microsoft's new division unless the board resigns. The decision came after Microsoft announced it would hire Altman and Greg Brockman to lead a new advanced AI effort in a show of support that also safeguarded Microsoft’s product roadmap.

OpenAI soon reinstated Altman as the CEO of OpenAI, and he, in his first official message after taking back the reins of the company, said that Microsoft will take a non-voting observer position on the board.

Microsoft, however, was unable to hold on to this position and had to give up its board observer seat at OpenAI in a move aimed at easing U.S. and UK antitrust regulators' concerns about the extent of its control over the AI startup.

Since then, the two companies have been at loggerheads over OpenAI’s restructuring plans and Microsoft's role and access to the new entity.

Governance and restructuring tensions

Between 2024 and 2025, OpenAI began pursuing a conversion of its for‑profit arm into a Public Benefit Corporation (PBC) while keeping nonprofit control. However, this conversion required hammering out new terms with Microsoft (and other stakeholders): equity splits, IP access, hosting/exclusivity, and a long‑debated ‘AGI clause’ that could affect Microsoft’s access to future models.

While the details were being discussed, OpenAI began diversifying its cloud infrastructure, signing deals with CoreWeave, Oracle, and rival Google to reduce reliance on Microsoft.

However, OpenAI was not the only one looking to diversify.

Why Microsoft needs a multi-model strategy

The restructuring of OpenAI was not just about shifting to a for-profit. It was also an attempt to ensure continued investments in the company and eventual IPO. However, since Microsoft was an early backer, it was concerned about access to the startup’s technology, intellectual property, and revenue.

In September 2025, a report from the Financial Times said the companies had signed a non-binding memorandum of understanding, marking a significant step forward in the start-up’s effort to convert to a more investor-friendly, for-profit structure.

Since then, Microsoft has been taking steps to reduce dependence on OpenAI. The deal with Anthropic marks a significant step in this strategy.

Anthropic’s role in Microsoft 365

While Microsoft is looking to reduce its dependence on OpenAI, it will have to pay for the use of Anthropic’s models in Microsoft 365 features.

For Microsoft, the additional cost is justified as Anthropic’s Claude Sonnet 4 performed better than OpenAI’s models in automating tasks such as financial functions in Excel or generating PowerPoint presentations based on instructions, the report said, citing one of the two people involved in the effort.

According to a Reuters report, along with the performance gains, Microsoft is also developing its own artificial intelligence models, while integrating DeepSeek's models into the Azure cloud as it seeks to maximize AI output to meet growing demand. This allows Microsoft to have a multi-model strategy, which is important for sustainable growth.

As for OpenAI, the startup will hope to use the revenue share it pays to Microsoft and other partners to build up avenues to secure more funding as it works towards an IPO.

Business strategies, not emotions, dictate moves

The Microsoft-OpenAI collaboration is neither ending nor headed for divorce. It is about Microsoft executing a portfolio strategy: keep OpenAI, add Anthropic where it wins, and build in‑house.

That strategy became inevitable once OpenAI’s governance and restructuring talks made access and control a moving target. The immediate aftereffect of the latest development will bring Claude‑powered features inside Microsoft 365, even as OpenAI models remain embedded across Azure and Copilot. Either way, users benefit.

The story also highlights the importance of multiple competitors to drive progress and give users more choice. The current AI market is a competitive one, and deals like the one signed by Microsoft and Anthropic are part of shifts. If we continue to see such deals between rivals, we may escape a future where AI, not unlike smartphone OS, is diminished to a duopoly where new entrants either fizzle out or are gobbled up by big tech companies.

Wednesday Poll

Quick Bits, No Fluff

• Google’s AI summaries face legal pushback: Penske Media sued Google, claiming its AI-generated summaries copy and repurpose full articles without permission.

• China targets Nvidia in antitrust probe: Chinese regulators found Nvidia violated anti-monopoly rules during a preliminary investigation, with penalties likely to follow.

• Big Tech races to universal translation: Apple, Google, and Meta are all pushing toward real-time AI-powered translators — but each is taking a different path.

Brain Snack (for Builders)

Meme Of The Day

Rate This Edition