The restructured OpenAI - Microsoft relationship creates as many questions as it answers

The restructured OpenAI - Microsoft relationship creates as many questions as it answers

While uncertainty about OpenAI’s governance and Microsoft’s stake is clarified, the deal introduces confusion about the definition of AGI and the entangled nature of competition in AI.

 

OpenAI’s multibillion-dollar restructuring produced a seemingly amicable arrangement for how the two companies will share OpenAI’s research in the pursuit of AGI. As we cover in-depth here, OpenAI is now freed from the shackles of the previous capped-profit model, and Microsoft is freed from the infamous AGI clause that barred it from independently developing AGI using OpenAI’s research.

 

With the new deal, the responsibility of defining and determining what constitutes “AGI” will now be left up to an “independent panel,” the composition of which is unknown, as is the criteria by which it will determine if and when AGI has been achieved. At the point it is achieved, Microsoft’s deal to receive 20% of OpenAI’s revenue will end. Regardless of if and when that happens, Microsoft will retain intellectual property rights to OpenAI’s models and products, including post-AGI models, until 2032. That’s a change from the original deal, where achieving AGI would have terminated Microsoft’s access to new OpenAI models. If it all sounds unusual, it is, but then again, so is most of what’s happening in the AI world.

 

In other news, OpenAI partner Nvidia continued its spending spree in Europe. Not long after the commitment to spend billions on U.K. data centers, Nvidia set its sights on Germany, with a Deutsche Telekom co-financing deal for a $1.2 billion data center. German software leader SAP is already set as a primary customer, signaling a strong enterprise demand for AI compute and integration within Europe’s manufacturing sector. Both Deutsche and Nvidia see promise for a broader initiative to expand AI infrastructure across Europe, and this DC project will accelerate Germany’s industrial and digital modernization, aligning with Germany’s goal of securing 100,000 GPUs, and the EU’s goal to spend $20 billion in the race to catch up with the U.S. and China in AI infrastructure.

 

Because telecom is one of the sectors leading in AI infrastructure investment, RCR Wireless’ Kelly Hill spoke to a company that has recently signed deals with telecom leaders like Telefonica Deutschland, Swisscom, and Lumen. In this new “RCR Talks” video interview with Blue Planet VP and GM Kailem Anderson, Kelly talks about what AI is and what it is not in the realm of telecom, and how high-quality data is the heart of real-time, context-aware use cases that will require AI intent.

Susana Schwartz
Technology Editor
RCRTech

AI Infrastructure Top Stories

Another bewildering AI deal: OpenAI restructuring answers some questions, but raises others about the definition of AGI and the entanglement of AI innovators.

Boosting Germany’s digital sovereignty: Nvidia and Deutsche Telekom to build $1.2 billion data center in Germany, with SAP signed as primary customer.

AI infra in telecom: In RCR Talks video Q&A, Blue Planet VP & GM Kailem Anderson discusses how telcos can make their data context aware, with intent.

In partnership with

AI-Powered Telecom Infrastructure
Supermicro, in collaboration with NVIDIA, delivers AI-powered infrastructure tailored for telcos, enhancing operational efficiency, network management, and customer experiences. Explore now 

AI Today: What You Need to Know

Making history: Meta and Blue Owl’s $27 billion deal is not only the largest private credit deal in history, but it might be a blueprint for financing AI data centers in the future.

AFB-turned-DC: FirstLight will provide fiber infrastructure for Green 4 Maine, a former Air Force Base-turned data center park in Maine. It also operates at the former Loring AFB.

Call for clarity in AI policy: ITI outlined how the Trump Administration could boost AI innovation if it cleans up regulatory misalignment, and ensures frameworks keep pace with technological change and changing business environments.

Smarter computing: Startup Counterintuitive has set out to build “reasoning-native computing,” enabling machines to understand rather than mimic. The company says it will help shift AI from pattern recognition to genuine comprehension.

Neoclouds to the rescue: Neocloud providers like Nebius, Lambda Labs and Nscale are charging onto the generative AI landscape, claiming to fulfill the nearly insatiable appetite for generative AI compute power.

Inconspicuous winners: Power constraint is creating winners in unglamorous businesses, such as thermal management, electrical equipment, and specialty contractors. Motley Fool takes a closer look.

Upcoming Events

This one-day virtual event will discuss the critical issues and challenges impacting the AI infrastructure ecosystem, examining the growth and evolution of the AI ecosystem as it scales and the need for flexible, sustainable solutions. 

Industry Resources

Related posts

Chips, data centers and engineers: All are evolving rapidly in the world of AI infrastructure

Rural data centers need more fiber in their diet

Building the AI superhighway through partnerships and collaborations

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Read More