Fed’s Schmid Questions Whether the AI Ecosystem Is Becoming “Too Big to Fail”
The artificial intelligence boom is increasingly attracting attention not only from technology investors but also from central bankers concerned about financial stability.
Kansas City Federal Reserve President Jeff Schmid said Friday that the Federal Reserve needs to determine whether the expanding network of AI companies, data centers, financing arrangements and commercial contracts could eventually become so interconnected that it creates a systemic risk.
Schmid specifically questioned whether the industry is moving toward a “too-big-to-fail AI ecosystem”, drawing a comparison with the interconnected financial system that required extraordinary government support during the 2007–09 financial crisis.
His comments do not mean the Fed believes the AI industry is currently in a financial crisis or that government support will be required. Instead, they highlight a question policymakers are beginning to examine: what happens to the wider economy if an AI-related company, financing structure or infrastructure project fails?
That question is becoming more relevant as AI investment expands into semiconductors, cloud computing, electricity, data centers, construction and corporate finance.
What Does “Too Big to Fail” Mean for AI?
The term “too big to fail” became closely associated with the global financial crisis.
Large banks and financial institutions had extensive connections through lending, securities, derivatives and other contracts. When losses spread through the financial system, policymakers feared that allowing major institutions to fail could cause damage far beyond those individual companies.
Schmid's concern is that a different form of interconnectedness could emerge around AI.
The AI ecosystem includes chip designers and manufacturers, cloud providers, model developers, data-center operators, electricity providers, construction companies, software businesses and financial institutions.
These companies are increasingly connected through investments, supply agreements, computing contracts and infrastructure spending.
Schmid said the Fed needs to understand “what’s inside” this expanding network and whether there is anything systemic within it.
In other words, the issue is not simply whether an individual AI company could lose money. The question is whether the failure of one important part of the ecosystem could trigger problems elsewhere.
Why Data Centers Are Central to the Concern
The rapid construction of data centers is one of the clearest physical manifestations of the AI boom.
Training and running advanced AI models requires enormous amounts of computing power. That has pushed technology companies to invest heavily in specialized chips, servers and data-center capacity.
The build-out also creates demand for electricity, land, construction equipment, cooling systems and transmission infrastructure.
The Kansas City Fed has previously highlighted how data-center expansion and AI adoption are contributing to rising U.S. electricity demand after years of relatively limited growth.
That means an AI slowdown could theoretically affect a much wider range of industries than software and semiconductor companies.
A company postponing a data-center project, for example, could affect equipment suppliers, construction contractors, utilities and financing arrangements connected to that project.
The economic impact would depend on the scale and nature of the disruption, but the growing network is precisely what policymakers want to understand.
The Financial Side of the AI Boom Matters
AI infrastructure is also becoming increasingly capital-intensive.
Companies need to spend billions on computing capacity before all the expected future revenue from AI services is necessarily visible.
That creates a familiar financial question: how much investment is supported by current cash flows, and how much depends on expectations of future growth?
If AI revenues grow rapidly, large infrastructure investments could generate strong returns.
If demand grows more slowly than expected, companies could face excess capacity, weaker returns on capital and pressure to reduce spending.
For investors, this distinction is critical.
A slowdown in AI capital expenditure would not necessarily mean that AI technology itself is failing. It could simply mean that companies are becoming more disciplined about the amount they spend on infrastructure.
But because the sector is interconnected, a sharp reduction in spending could affect multiple suppliers simultaneously.
This Is Not the Same as the 2008 Financial Crisis
Schmid's comparison should not be interpreted as saying the AI industry currently resembles the financial system of 2008 in every respect.
The underlying mechanisms are different.
The 2008 crisis involved highly leveraged financial institutions, mortgage-related assets, complex securities, derivatives and short-term funding markets.
The AI ecosystem is primarily connected through technology supply chains, capital expenditure, long-term contracts and infrastructure financing.
That difference matters.
A decline in AI valuations or a failed AI project would not automatically create the same type of liquidity crisis that occurred when banks became uncertain about each other's financial exposures.
Still, interconnectedness can create economic spillovers even without a banking-style crisis.
Why Central Banks Are Watching AI
The Federal Reserve has another reason to monitor AI: its impact on economic growth and inflation.
Schmid has previously said AI could eventually contribute to a more productive, supply-driven growth cycle. The Kansas City Fed has noted that productivity-enhancing AI could increase output without necessarily creating the same inflationary pressure associated with demand-driven growth.
But the transition is not necessarily smooth.
A large increase in AI-related investment can itself boost aggregate demand.
Companies building data centers need construction workers, electrical equipment, power generation, chips and other infrastructure. That spending can increase demand for scarce resources.
If supply cannot keep pace, the investment boom could contribute to higher prices.
This creates an unusual situation for monetary policymakers: AI could simultaneously increase the economy's productive capacity over time while adding demand and inflation pressure during the infrastructure-building phase.
What Happens If AI Investment Slows?
One of the most important risks for investors is a potential mismatch between infrastructure spending and eventual AI revenue.
Suppose companies collectively build more computing capacity than customers ultimately need. The result could be lower utilization rates and weaker returns on data-center investments.
That could affect:
Semiconductor demand
Cloud infrastructure spending
Data-center construction
Utility investment
Equipment suppliers
Commercial real estate linked to data centers
Debt financing for infrastructure projects
The impact would not necessarily be uniform.
Companies with diversified revenue streams and strong balance sheets could have more flexibility than businesses whose growth depends heavily on a small number of AI customers.
That makes company-level analysis particularly important for investors.
AI's Circular Business Relationships Are Worth Watching
Another emerging issue is the increasingly complex commercial relationship between major technology companies.
A large AI company may purchase computing capacity from a cloud provider. That cloud provider may purchase chips and networking equipment from other technology companies. The same AI company can also receive investment from major technology firms that benefit from its growth.
Such arrangements can accelerate the development of the ecosystem because capital and demand reinforce each other.
But they can also make the underlying economics harder to evaluate.
Recent analysis has highlighted concerns about circular financial and commercial relationships across the AI supply chain, although the existence of such relationships does not by itself establish a systemic financial risk.
For investors, the important question is whether transactions represent genuine end-user demand or whether spending is increasingly dependent on a relatively small group of companies financing and purchasing from one another.
What Could Make the AI Ecosystem Systemically Important?
Several developments would make Schmid's concern more significant.
Heavy leverage
If data-center and AI infrastructure projects increasingly depend on large amounts of debt, falling revenues could create greater financial stress.
High concentration
If a small number of companies control critical chips, cloud infrastructure or AI models, problems at one firm could have broader consequences.
Long-term contractual commitments
Large take-or-pay contracts and other fixed commitments can create financial pressure if demand assumptions change.
Infrastructure dependence
AI relies on electricity, networks, semiconductor manufacturing and data centers. Problems in one area can affect the rest of the chain.
Broader economic dependence
The risk would become more important if AI investment represented a sufficiently large share of corporate capital expenditure, economic growth or financial-market valuations.
None of these factors alone means an AI crisis is inevitable. They are indicators policymakers and investors can use to assess how interconnected the ecosystem is becoming.
What Investors Should Watch
Schmid's comments make several metrics particularly important.
AI capital expenditure: Track whether major technology companies continue increasing data-center and computing investment.
AI revenue growth: Rising infrastructure spending needs to eventually translate into sustainable revenue.
Data-center financing: Watch the amount and structure of debt supporting new infrastructure.
Semiconductor demand: Chip orders can provide an early indication of whether AI infrastructure spending is accelerating or slowing.
Cloud utilization: Strong demand for computing capacity can provide evidence that infrastructure is being absorbed.
Electricity demand: Rapid growth in data-center power consumption shows how deeply AI investment is spreading into the physical economy.
Corporate concentration: Increasing dependence on a handful of major AI customers or suppliers could amplify risks.
Credit-market conditions: Higher borrowing costs can make capital-intensive AI projects harder to finance.
The Fed's Question Is About Systemic Risk, Not Whether AI Is Real
AI's economic potential is not in question simply because a central banker is examining its financial risks.
The Kansas City Fed itself has previously discussed the possibility that AI could contribute to productivity and supply-driven growth.
The issue Schmid raised is narrower and more consequential for financial stability: could the scale of AI investment and the connections between companies eventually make a major disruption difficult for policymakers to contain?
That is an open question, not a prediction.
The answer will depend on how the AI ecosystem develops, how companies finance infrastructure and whether AI-generated revenues ultimately justify the enormous capital being committed to the sector.
Bottom Line
Kansas City Fed President Jeff Schmid has raised a new policy question around the AI boom: whether the growing network of companies, contracts and data-center investments could eventually become sufficiently interconnected to create systemic risks.
The comparison with the 2008 financial crisis should be treated carefully. Today's AI ecosystem is structurally different from the pre-crisis financial system, and there is no indication from Schmid's comments that the Fed believes an AI bailout is currently necessary.
For investors, however, the warning highlights an important shift. AI is no longer just a story about software and chip valuations. It increasingly involves debt, infrastructure, electricity, construction, cloud contracts and the broader economy.
The key indicators to watch are whether AI revenues keep pace with infrastructure spending, whether leverage rises, and whether commercial relationships become concentrated enough to transmit problems from one part of the ecosystem to another.
Follow the blog for more updates on AI, technology, Federal Reserve policy, markets and the global economy.
This article is for informational and educational purposes only and should not be considered investment advice.