Does The AI Industry Now Need Financial Engineering To Continue To Grow?

Up until quite recently, the AI industry has largely been funded by means of a strategy we’re all quite familiar with. That is, venture capital-backed startups, growth stories that are rewarded by public markets and technology giants that spent heavily on research and development. But now, as AI is moving from software into physical infrastructure, the scale of capital required is beginning to look quite different.

Recent reports around NVIDIA’s efforts to help unlock as much as a whopping $500 billion in AI infrastructure spending indicate that the industry may just be entering a new phase. That is, rather than just building models and applications, companies are now financing data centres, energy networks and computing capacity on a scale more commonly associated with airports, railways and power stations.

Is This a Logical Next Step for the AI Industry?

The question is whether this is simply the next stage of growth or a sign that AI now needs increasingly sophisticated financial structures to keep expanding.

Sherif Higazy, Founder and CEO of Megaton AI, believes there are clear historical parallels: “Similar to the railroad buildout of the 19th century that pushed capital markets to grow and evolve to finance them, we’re seeing a similar dynamic play out where businesses are growing to unprecedented sizes, building mega infrastructure projects, usually reserved for nation states and regulated utilities.”

And if that comparison is accurate, then traditional venture funding alone may no longer be enough. Indeed, if there’s one thing we can all agree on, it’s that AI infrastructure requires enormous upfront investment, often with long timelines before returns are finally realised.

But that doesn’t necessarily mean that the industry is becoming dependent on financial engineering for its survival. Olga Kokhan, CEO of Tinkogroup, argues that finance should be viewed as an enabler rather than the main growth driver. According to Kokhan, “The increasing role of financial engineering in AI reflects a broader shift: AI is no longer viewed as an emerging technology category but as critical infrastructure.”

She believes that the real challenge remains execution. Capital may help companies build infrastructure faster, but it can’t guarantee sustainable businesses or customer demand.

Removing Obstacles?

Others suggest that NVIDIA’s latest moves are part of a broader effort to remove obstacles that could slow AI adoption. Lior Prosor, Partner at Deep33, points out that financing may simply be the next bottleneck being addressed: “If GPUs and data centers can be financed more like aircraft, power plants or traditional infrastructure, you suddenly open up enormous pools of capital from banks, private credit, infrastructure funds, insurers and eventually pension funds.”

And if that happens, access to capital could become significantly easier. But Prosor also notes that “easier credit is still easier credit”, raising the possibility that excessive leverage could contribute to future market distortions.

That concern reflects a wider debate about whether the economics of AI have been fully proven. Charles Radclyffe, CEO and founder of EA Global AI, warns us that there’s a difference between merely financing growth and validating it, on the other hand.

“The danger comes when the financing becomes the story.” Quite poignant, I must admit.

As Radclyffe notes, AI infrastructure ultimately has to generate enough value to justify the capital invested in it. If increasingly complex funding arrangements become necessary simply to maintain momentum, investors may eventually begin questioning the assumptions underpinning the sector.

For now, financial innovation appears less like a flashing red warning sign and more like a reflection of AI’s growing maturity. Industries like energy, telecoms and transport all evolved new financing models as they scaled.

Whether AI follows the same path successfully will, most liekly, depend on one key issue. That is, can the industry turn hundreds of billions of dollars of infrastructure spending into productivity gains, profits and lasting economic value?

Because if it can, financial engineering may become a permanent feature of the AI economy. But if it can’t, investors may eventually discover that financing growth is far easier than creating it.

The Experts:

• Sherif Higazy: Founder and CEO at Megaton AI
• Olga Kokhan: CEO at Tinkogroup
• Lior Prosor: Partner at Deep33
• Charles Radclyffe: CEO and founder of EA Global AI
• Jeremy Hand: Chair of Sabio
• Frances Doherty: Corporate Partner at Simmons & Simmons

Sherif Higazy, Founder and CEO at Megaton AI

“Similar to the railroad buildout of the 19th century that pushed capital markets to grow and evolve to finance them, we’re seeing a similar dynamic play out where businesses are growing to unprecedented sizes, building mega infrastructure projects, usually reserved for nation states and regulated utilities.

“Financing this requires NVIDIA to help its customers tap into a more diverse mix of financing channels and institutional investors rather than rely on a single channel like traditional risk capital or bond markets.”

Olga Kokhan, CEO at Tinkogroup 

“The increasing role of financial engineering in AI reflects a broader shift: AI is no longer viewed as an emerging technology category but as critical infrastructure. You can’t just have traditional venture funding anymore when you’re talking about hundreds of billions of dollars for data centers, chips, energy capacity, and computing resources to support the scale of investment that’s needed.

“That said, I don’t think exotic financing structures will be the driver of the next phase of AI growth. They are an enabler. As a result, the companies that can efficiently deploy capital and convert infrastructure into real-world results for customers will remain the ones that create lasting value.

“Having founded Tinkogroup, a data services company that supports AI and data-driven initiatives, I’ve learned that one of the biggest challenges isn’t just access to funding, it’s the ability to build reliable data pipelines, manage operational complexity, and deploy resources effectively. Capital can speed growth, but it can’t replace effective execution.

“What we are seeing today is the coming of age of AI. As the industry grows more infrastructure-heavy, it will resemble other capital-intensive industries that rely on different financing models. The real question is not whether financial engineering is needed but whether companies can translate that capital into sustainable business results.”

Lior Prosor, Partner at Deep33

“NVIDIA has been systematically going after the major bottlenecks that could slow down the AI scaleout.

“They’ve invested across the physical supply chain, from materials and fiber with Corning, to optics and equipment with Coherent and Lumentum, to semis with Intel. Now they’re going after another major bottleneck: financing.

“If GPUs and data centers can be financed more like aircraft, power plants or traditional infrastructure, you suddenly open up enormous pools of capital from banks, private credit, infrastructure funds, insurers and eventually pension funds. NVIDIA’s 25% backstop makes that even more powerful.

“The regulatory environment is moving in the same direction. Recent SEC guidance makes it easier to securitize certain data-center financings and distribute that credit to third-party investors. That should make capital more available and ultimately cheaper.

“If the move is successful, and I assume it will be, one bottleneck will ease, but the physical world ones are still here for the foreseeable future- power, cooling, chips, memory construction and land build approvals.

“The flip side is that easier credit is still easier credit. More leverage could contribute to the formation of a bubble, and I also assume CDS spreads will widen across the hyperscalers.”

Charles Radclyffe, CEO and founder of EA Global AI

“The short answer is probably yes, but we should be careful what we mean by “financial engineering”.

“AI is moving from a software business into an infrastructure business. Chips, data centres, electricity, cooling systems and networks are expensive physical assets with long payback periods. That naturally brings in project finance, structured debt, long-term purchasing agreements, public capital and institutional investors.

“In that sense, AI is beginning to look less like a conventional startup sector and more like energy, telecoms or railways. Nvidia’s reported partnership with major Wall Street firms to raise more than $500 billion for AI infrastructure is a sign of that transition. The technology industry is increasingly looking for ways to mobilise the kind of capital traditionally used to build national infrastructure.

“The danger comes when the financing becomes the story.

“AI is, at its core, about channelling capital from investors into machines that replicate human capabilities. That capital only makes sense if those machines create enough value to repay it. If the model depends on increasingly elaborate partnerships, guarantees, vendor financing and long-dated contracts to keep the money moving, we may be looking at a financial structure supporting an economic assumption that has yet to be proven.

“There is a wider political issue too. Once an industry becomes dependent on vast pools of capital, ownership tends to concentrate. That is how we arrive at the version of “Technofeudalism” I have spoken about: the public pays for access, the technology companies own the infrastructure, and the risk gets socialised when the numbers stop working. The important questions are who owns the infrastructure, who receives the benefits and who carries the downside.

“So yes, sophisticated finance may be needed to build AI at scale. It should come with sovereign capability, public scrutiny and a clear account of where the productivity gains go. The UK’s own AI ambitions show the difficulty. A few hundred million pounds sounds substantial until it is compared with the cost of national infrastructure. If we want British or Welsh companies to have a meaningful place in this economy, we need capital that builds local capability and reduces dependence on American platforms.

“My concern is that we are becoming very good at financing the transition while remaining rather poor at discussing its purpose. AI can release people from a great deal of repetitive work. The test is whether we use that productivity to distribute opportunity more fairly, protect work involving creativity, strategic judgement, accountability and relationships, and preserve the human touch.

“Financial engineering can help build the machines. It cannot decide what kind of society they are built for.”

Jeremy Hand, Chair of Sabio

“AI is increasingly becoming an infrastructure investment story, which requires funding to be structured around differing time horizons, risk profiles and expected returns. This is the normal evolution of financing as an industry scales. Capital is coming and will continue to come from a broad ecosystem – public markets, venture capital, private credit, infrastructure funds, strategic corporates and, where appropriate, public and sovereign sources.

“The UK’s opportunity is to create the conditions in which that capital can be deployed confidently and at scale: a stable policy and regulatory framework, appropriate planning and energy infrastructure, and mechanisms that encourage long-term institutional investment. The Government’s objective should be to ensure that commercially viable AI businesses and infrastructure are financed and built in the UK, rather than elsewhere.”

Frances Doherty, Corporate Partner at Simmons & Simmons

“AI is critical to nations, not only to the defence of countries (through next-generation autonomous weapons), but also to their political functions as well as access to prosperity and value generation. Funding this technology revolution requires never-before-seen requirements of capital as infrastructure is built to support the need for ever-increasing levels of computing power that AI requires. The traditional routes to financing for high-growth companies was venture and growth capital, and private equity.

“The amounts now required cannot only be sustained by those routes alone and we have seen many novel approaches to fundraising, from financial institutions which may not have been historically interested in tech investing, and combinations of different investors, uniting to back a critical technology.

“Given its importance to national security and defence, we are also seeing a rise in the support for these technologies from governmental institutions. This is outlined in the UK Government’s AI Action Plan and seen in the increasing activity of the British Business Bank. However, it is to be seen if there will be a change of approach following the change of Prime Minister in the UK.”