BIS finds 55.2% of AI funding comes from other AI firms

The AI industry has no shortage of investors. Increasingly, those investors look a lot like the companies they are backing.

A new study from the Bank for International Settlements found that 55.2% of incoming investment into AI firms between 2021 and 2025 came from other AI firms. Put another way, more than half the money fueling the sector is the sector funding itself.

The research, published as BIS Bulletin 137 on October 1, 2026, carries the title “Circular relationships among AI firms.”

Inside the loop

The BIS team examined 1,246 AI companies spread across different layers of the AI supply chain. Within that group, they identified 972 investment relationships where one AI firm put money into another.

On the outgoing side, 28.7% of the deal value AI firms deployed went into other AI firms.

Deals tied to commercial supply-chain relationships made up 46.4% of intra-AI deal value. Yet those same deals represented only 16.1% of intra-AI deals by count.

Translated: the deals where a supplier invests in a customer, or trades with a firm it funds, are relatively rare. When they happen, though, they tend to be enormous.

The report notes that larger investment deals are a significant driver of these circular structures.

Why the BIS sees both benefits and risks

The bulletin does not treat circular financing as automatically bad. The authors point to genuine advantages.

An investor that also supplies your critical inputs probably understands your business better than a generalist fund. That closeness can reduce information asymmetries, the classic problem where one side of a deal knows far more than the other.

These ties can also help companies lock in access to key inputs. In a sector where hardware and computing capacity are bottlenecks, guaranteed supply is a real strategic asset.

The BIS flags risks of macroeconomic instability, greater opacity, and contagion during periods of financial stress.

The report also places hyperscalers, the large tech companies running massive cloud and computing operations, at the center of this picture. These firms are pouring money into capital expenditures while simultaneously investing in AI companies and securing supply commitments for hardware and services.

The bulletin was authored by Jon Frost, Rudraksh Kansal, Kumar Rishabh, Vatsala Shreeti, and Leanne Si Ying Zhang. It belongs to a broader BIS series studying what the AI investment boom means for financial stability.

Media coverage on October 2, including reports from MarketScreener and Yonhap Infomax, zeroed in on how unusual the closed-loop structure is. Most sectors do not raise the majority of their outside capital from their own peers.

What this means for investors and regulators

For investors, the most immediate issue is concentration. When over half of a sector’s incoming capital comes from inside that sector, the funding base is far less diversified than it appears from the outside.

There is also a revenue-quality question worth asking. If a supplier invests in a customer, and that customer then spends heavily with the supplier, analysts may want to separate organic demand from demand the supplier effectively financed itself.

The 46.4% figure makes this more than a theoretical concern. Nearly half of intra-AI deal value sits in exactly these supplier-customer arrangements, concentrated in a small share of very large deals.

Earlier BIS research has already raised warning signs about possible overvaluation and potential pullbacks in AI investment, and this bulletin adds a structural reason for caution on top of the pricing concerns.

For regulators, circular relationships are not illegal, and the BIS itself acknowledges their practical benefits. But clearer reporting on supplier investments, supply commitments, and cross-holdings would let markets price these relationships more accurately.

The BIS data covers 2021 through 2025, a period of rapid AI expansion.