How does the map of connections look like in the AI sector

In the high-stakes race for AI supremacy, the traditional rules of venture capital are being rewritten. We aren’t just seeing simple investments; we are witnessing the rise of a complex, interconnected "circular financing" ecosystem.

What is "circular financing"?

In simple terms, circular financing is a strategic loop where a supplier funds its own customers so they have the money to buy the supplier's products.

In the AI world, this typically happens when a "Big Tech" company (like Microsoft or Nvidia) invests billions of dollars into an AI startup (like OpenAI or Mistral). The startup then immediately uses that cash to pay for the investor’s cloud services or high-end chips.

Imagine a lemonade stand owner giving a neighbor $10, with the unspoken agreement that the neighbor will use that $10 to buy 10 cups of lemonade. The owner gets their $10 back as "revenue," and the neighbor gets the lemonade they couldn't otherwise afford.

"The AI boom is a financial ouroboros, swallowing itself while calling it progress." - Cory Doctorow, Tech Journalist and author of Enshittification

Mapping the AI Circular Financing Ecosystem

The AI industry is currently powered by a massive financial feedback loop that connects the world’s most powerful tech giants, chipmakers, and AI labs. In this "circular" ecosystem, companies aren't just investing in one another; they are actively financing their own future customers to ensure a guaranteed stream of revenue.

From hardware giants like Nvidia and AMD to cloud providers and frontier model developers like OpenAI and xAI, this web of multi-billion dollar contracts and equity stakes functions as the industry’s secret engine. The map below illustrates these complex interdependencies, revealing how capital flows between these players, why it’s accelerating the build-out of global AI infrastructure, and where the risks of this "financial ouroboros" lie.

Map of connections in the AI sector

As you can see, the complexity of these relationships is staggering. What initially looks like a simple investment is actually a web of strategic dependencies. To make sense of the map above, it helps to categorize these connections into four distinct pillars that keep the AI ecosystem moving:

• Hardware & Software: These connections represent the physical backbone (like NVIDIA’s H100s);

• Direct investments: These are the capital injections, where tech giants buy an equity stake in a startup to secure their place at the table;

• Venture capital: These connections represent the "patient capital" from institutional funds that scale startups;

• Services & API dependencies: This is the most active layer of the loop. It represents the ongoing "operational" flow, where startups pay tech giants for cloud computing, API access, or data storage.

1. The NVIDIA / CoreWeave / Microsoft triangle

This is the most famous example of the "compute-for-equity" cycle.

• The investment: Nvidia and Microsoft have both invested heavily in CoreWeave (a specialized cloud provider).

• The purchase: CoreWeave uses that capital to buy billions of dollars worth of Nvidia H100s.

• The service: Microsoft then actually leases back Nvidia chips from CoreWeave to run its own AI workloads when its own data centers are at capacity.

The result: Nvidia gets a guaranteed buyer for its chips, CoreWeave gets the hardware others can't buy, and Microsoft gets the capacity to power OpenAI.4

2. The Microsoft / OpenAI / Oracle loop

• The investment: Microsoft has invested roughly $13 billion in OpenAI;

• The spending: Much of that "investment" is actually in the form of Azure cloud credits;

• The expansion: Because Microsoft cannot build data centers fast enough for OpenAI’s training needs, they recently partnered with Oracle.

The circle: OpenAI now uses Microsoft’s money to pay Microsoft, which in turn pays Oracle to provide the infrastructure, which is filled with Nvidia chips.

3. The "GPU Cloud" specialized players: Nebius & Nscale

These companies represent the new "AI infrastructure" tier.

• Nebius (born from Yandex) and Nscale are positioning themselves as pure-play AI clouds. They receive backing or priority allocations from Nvidia to build clusters.

Startups like Mistral (the French AI leader) receive investment from the likes of Microsoft and Nvidia. Mistral then takes that capital and spends it on the infrastructure provided by these GPU clouds to train their models.

4. Vertical AI: Ambience, Harvey, and Anysphere

These are "Applied AI" companies. They are often the end-users in this cycle.

• Anysphere (Cursor), Harvey AI (Legal), and Ambience Healthcare (Medical) have received funding from major venture arms (including OpenAI’s Startup Fund or Nvidia).

The flow: The cash flows from the chipmaker/platform holder -> to the startup -> and immediately back to the platform holder in the form of API fees (to OpenAI) or Cloud fees (to Microsoft/Oracle/Nvidia).