Imagine trying to map the electrical grid of a continent while the power lines are being laid in real time. That’s the challenge facing investors trying to understand the artificial intelligence boom, where billions flow through a tangled web of deals connecting tech giants, startups, and financial players. A new Morgan Stanley analysis reveals what many suspected but few could quantify: the AI industry has become a complex ecosystem of circular financing where suppliers fund customers and customers fund suppliers, creating unprecedented interdependencies.
The Financing Maze Behind AI’s Infrastructure Boom
Morgan Stanley’s “Mapping the AI Ecosystem” study, highlighted in recent Financial Times coverage, shows that a significant portion of AI funding remains circular and tied to long-dated compute purchase agreements. These arrangements – including supplier financing, revenue-sharing deals, take-or-pay contracts, and vendor repurchase agreements – function as financing mechanisms enabling participants to scale infrastructure beyond what their standalone cash flows would support. The analysts warn the industry is reaching “unprecedented levels of capital intensity” with a mismatch between near-term capital needs and when AI revenues will actually materialize.
What makes this ecosystem particularly opaque? Current disclosures are inadequate to fully understand the interrelated nature of these transactions. “The sophistication has outpaced current accounting standards,” the Morgan Stanley team notes, making it difficult for investors to assess the true economics of the system or identify who ultimately bears the economic risk. This complexity creates what one analyst calls “the finest Pepe Silvia tradition” of interconnected deals – a reference to the convoluted conspiracy board from the TV show “It’s Always Sunny in Philadelphia.”
Global Bets and Calculated Risks
While the Morgan Stanley analysis focuses on the structural complexity, individual players are making massive, concentrated bets that reveal both the scale of opportunity and the potential for systemic risk. SoftBank’s recent $30 billion commitment to OpenAI provides a striking example. The Japanese conglomerate may temporarily exceed its self-imposed 25% loan-to-value ratio limit with this investment, raising investor concerns that have already contributed to a 45% share price decline since last October. S&P has revised SoftBank’s outlook to negative, with analyst David Gibson warning: “There’s [an estimated] $50bn of funding, between OpenAI, investments and refinancing, that they have got to put in place in the course of 2026. The loan to value will hit 25 per cent or more. So to me that’s the story as I’m not sure the market is prepared for it.”
Meanwhile, smaller players are finding niches in this high-stakes environment. Air Street Capital, Europe’s largest one-person venture capital firm, recently raised $232 million for a new AI-focused fund, bringing its total assets under management to about $400 million. Founder Nathan Benaich explains the strategy shift: “One of the reasons to go bigger now is the opportunity set has accelerated dramatically. Companies want to raise faster and raise larger rounds, so you need to adapt the model for the game that’s being played.” His firm focuses on what he calls “the deployment market and vertical applications and selected infrastructure tools” – essentially, the practical applications of AI rather than the massive foundational models.
The Infrastructure Builders and Their Growing Pains
The AI boom has created unexpected winners in infrastructure, with companies like Nscale transitioning from cryptocurrency mining to AI-focused data center development. Originally Arkon Energy, Nscale faced significant financial challenges including a $102 million loss in 2024 and loan defaults to Sandton Capital Partners. Yet the company now boasts a $14.6 billion valuation and is pursuing ambitious data center projects in the UK, Europe, and US with support from Nvidia. Sandton co-founder Rael Nurick downplays the defaults as “an accounting technicality,” noting that paperwork delays occurred because “everyone was focused on closing the Series A” funding round.
This infrastructure race has global implications, particularly in Europe where Siemens CEO Roland Busch warns that the continent’s focus on building sovereign AI infrastructure could be a “disaster” by slowing innovation and economic growth. Speaking to the Financial Times, Busch argues that the EU should prioritize deploying existing AI tools rather than waiting for domestic infrastructure, citing concerns that excessive regulation and security measures are causing Europe to fall behind the US and China. “You should not throttle your innovation speed for the sake of creating sovereignty,” he states bluntly.
Who Benefits – and Who Gets Left Behind?
The concentration of capital and infrastructure raises fundamental questions about who will benefit from AI’s economic transformation. BlackRock CEO Larry Fink warns in his annual shareholder letter that AI risks intensifying wealth inequality by concentrating gains among a small group of businesses and investors who finance AI growth. “The massive wealth created over the past several generations flowed mostly to people who already owned financial assets,” Fink notes. “AI threatens to repeat that pattern at an even larger scale.”
Fink highlights that companies with data, infrastructure, and capital to deploy AI at scale will benefit disproportionately, creating what he calls “the broader question of who participates in the gains.” His concern echoes through the Morgan Stanley analysis, which shows how complex financing arrangements create interdependencies that could amplify both gains and losses across the ecosystem.
The Transparency Challenge
As the AI investment cycle accelerates, the lack of transparency in these complex arrangements becomes increasingly problematic. Morgan Stanley analysts emphasize that “investors require greater disclosure transparency to evaluate the durability of demand, the source of revenues, and the ultimate bearers of economic risk.” Without enhanced disclosures, they argue, it’s challenging for investors to fully assess the risks and rewards of the AI investment cycle.
The question facing businesses, investors, and policymakers is whether this complex web of financing represents a sophisticated adaptation to AI’s unique capital requirements or a house of cards built on circular dependencies. As one industry observer puts it, we’re witnessing the creation of an entirely new financial architecture for technology investment – one that may determine not just which companies succeed, but how broadly the benefits of AI are distributed across the global economy.