AI’s “looping” makes me dizzy

On Tuesday, Microsoft, Nvidia, and Anthropic announced a strategic partnership, summed up by Microsoft CEO Satya Nadella: “We will increasingly become each other’s customers.”

Incredibly “Now, let’s go around the mulberry bushes.” right? Microsoft purchases Anthropic’s model. Anthropic runs Claude on Microsoft’s Azure cloud. Anthropic acquires Nvidia chips. Both Microsoft and Nvidia have invested in Anthropic. If that sounds like a big circle going around and back…that’s because it is. To be honest, it makes me dizzy.

But this trio loop-de-loop is not an anomaly. Recently, it has become a dominant business model in the AI ​​industry. Hyperscalers, model labs, and infrastructure companies are increasingly forming closed-loop partnerships that serve as a type of AI mutual assurance arrangement. In other words, everyone is a partner, vendor, and customer at the same time.

This kind of circle game has been the key to success over the past three years for Nvidia, which smashed revenue targets and posted a 62% jump in its third-quarter results yesterday. as luck‘s Shawn Tully recently detailed how the company is building its own circular ecosystem by investing in, and in some cases funding, its own customers, from OpenAI to CoreWeave. The goal is to design a perpetual motion machine that consumes GPUs. This is a way to guarantee demand in a world where hyperscalers are looking to build their own chips. For example, Nvidia pledged to invest up to $100 billion in OpenAI in September. As part of the agreement, OpenAI “will buy at least 10 gigawatts worth of capacity on Nvidia AI chips. It’s very uncertain,” Seaport analyst Jay Goldberg said recently. [for its chips]? ”

There are many other non-Nvidia loops out there. For example, Anthropic has long had a similar arrangement with Amazon. Amazon is a major investor in Anthropic and immediately provided Anthropic with access to AWS infrastructure, Amazon’s custom Tranium chips, and key partners to train and run AI models. As a result, Amazon will see increased revenue from its cloud business and AI chip business. Most recently, OpenAI announced a multi-year partnership with AMD in October. OpenAI will acquire 6 gigawatts of AMD GPUs, while AMD has given OpenAI the option to acquire up to 10% of the company’s stock. AMD gets guaranteed demand. OpenAI acquires second chip supplier. Another loop.

There is also a sovereign loop. Just yesterday, AMD, Cisco, and Saudi-backed HUMAIN formed a joint venture to build up to 1 gigawatt of AI infrastructure in the Kingdom. Both companies are both investors and exclusive suppliers. AMD and Cisco have committed capital to the joint venture, which is contractually designed to purchase AMD’s GPUs and Cisco’s networking equipment within HUMAIN’s Saudi data centers. This is the same circular logic. Investors will fund suppliers, suppliers will buy from investors, and everyone will be able to advertise significant growth.

And NVIDIA isn’t immune to this either. Yesterday, alongside Elon Musk’s xAI, we also announced a partnership with HUMAIN to build a massive new AI data center in Saudi Arabia.

Is “looping” a bad thing? Well, Nvidia has long benefited from it, and startups like OpenAI and Anthropic probably wouldn’t be where they are today without it. However, there are also inherent risks. It is the concentration of power within a small group of players. Huge debts between companies that have not yet proven sustainable business models. Actual market signals are unclear, making it difficult to determine whether there is real demand. What happens if one of the players in the circle stumbles? And what happens to the player removed from the circle?

After all, circle games are fun. But we all know what happens at the end of “Ring Around the Rosey.” They all fall down. Nvidia’s strong quarterly results have allayed fears of an AI bubble for now, but how long will this loop-de-loop business model last? We don’t know the answer. But at this point you may need some Dramamine.

So, here’s more AI news for you.

sharon goldman
sharon.goldman@fortune.com
@SharonGoldman

The fate of AI

Nvidia says it is “on track to generate $5 trillion in revenue” by 2026. That would make it one of the largest companies in America — Matthew Heimer

Nvidia CEO Jensen Huang name-checked Saudi AI company Humane three times during an earnings call. Here’s why – Jeremy Kahn

Stock markets are surging toward AI’s “Show me the money” moment, with the possibility of a global crash — Jim Edwards

AI power and water consumption is troubling the agricultural sector: “Remember, we also need AI to grow food” — Angelica Ang

Nvidia exceeds revenue targets, predicts AI infrastructure spending to reach trillions of dollars by end of decade – Sharon Goldman

AI in news

President Trump’s draft executive order targets state AI laws. President Donald Trump is considering challenging state AI regulations, a leaked draft report says. Reuters. The executive order aims to override state AI laws through litigation and by making conditions on access to federal funds. The draft specifically criticizes California’s SB 53, calling it “complicated and burdensome.” The order creates an AI Litigation Task Force to sue states and could withhold broadband funding from states with strict AI rules. The proposal follows President Trump’s push to include similar preemptive measures in the upcoming National Defense Authorization Act. It will likely face backlash at the state level, and has already sparked a MAGA backlash. read more Reuters.

Meta partners with Yann LeCun’s new startup. The longtime head of AI at Meta is leaving the company to start his own startup. In a post on LinkedIn, LeCun said he is building a startup to advance the Advanced Machine Intelligence (AMI) research he has been working on at Meta’s FAIR and NYU. Last week, the Financial Times reported that Mr LeCun was planning to launch his own startup and was in early talks to raise funding. He called the creation of FAIR (Meta’s AI Research Lab) his proudest non-technical accomplishment. Read the full post here.

Google DeepMind expands its robotics efforts with new hires. Earlier this month, Google DeepMind hired former Boston Dynamics chief technology officer Aaron Sanders as the company’s new vice president of hardware engineering, Google DeepMind reported. Wired. The move is part of CEO Demis Hassabis’ DeepMind ambition to turn Gemini into an operating system for physical robots. Hassabis previously said DeepMind is building an AI system that can work “out-of-the-box and across all body configurations.” Boston Dynamics, known for its advanced legged robots, was actually owned at one time by Google, which bought it in 2013 and sold it four years later. Sanders worked on developing a six-legged amphibious prototype during his tenure and became CTO in 2021. wired.

Europe rolls back privacy and AI laws. The EU is proposing a series of changes to its landmark privacy regulation, the GDPR, and a postponement of key provisions of the AI ​​Act. The proposal would ease restrictions on data sharing, allow personal data to be used for AI training under certain conditions, extend compliance deadlines for high-risk AI systems, and reduce cookie pop-ups seen across Europe. The European Commission has branded the changes as a simplification to boost innovation, but the leaked draft has already sparked a backlash from civil rights groups and lawmakers, who say the EU is bowing to pressure from Big Tech companies. read more The Verge.

AI calendar

November 26th-27th: World AI Conference, London.

December 2nd – 7th: NeurIPS, San Diego.

December 8th-9th: fortune brainstorming A.I. San Francisco. Click here to apply for participation.

January 7th to 10th: Consumer Electronics Show, Las Vegas.

From March 12th to 18th: Austin, SW.

March 16th-19th: Nvidia GTC, San Jose.

April 6th to 9th: HumanX, San Francisco.

Pay attention to AI numbers

60%

That’s the percentage of U.S. adults who have tried generative AI since the launch of OpenAI’s ChatGPT, according to a new report from the Computer and Communications Industry Association. It is also the fastest adopted technology in history.

According to the 2025 SPICE AI Report, approximately three in five U.S. adults have used GenAI within three years, outpacing the adoption curves of both smartphones and the internet. Daily GenAI usage among U.S. adults has jumped from 12% to 17% in just eight months, and workplace integration is accelerating even further. Approximately 40% of employees are currently using AI tools at work, reporting an average increase in productivity of 15%. Daily employee usage of AI jumped from 21% to 31% between March and July 2025. The report found that most AI users (77%) have a positive impression of the technology, and that sentiment tends to become more positive over time.