• AI IPO tsunami strains investor capacity, liquidity buffers
• Hyperscalers burn cash while executives’ AI ROI doubts grow
• Funding race risks broader market churn amid bubble fears

Global equity capital markets are at a historic inflexion point, with a wave of AI issuance hitting stock exchanges all at once, putting new stresses on investor capacity amid colossal demand for direct financing.

This is happening at the same time as some of the largest corporate users of AI are beginning to question the cost benefits of the vastly expensive technology which is, in many cases, not producing the business deliverables required to justify such huge corporate outlay.

On both the supply and demand side of the AI industry, cash burn is the prevailing narrative.

For hyperscalers, the need to continue to grow data centre capacity has led to a rush to tap the equity capital markets and what is being asked of investors is unprecedented for a single industry over such a short period of time.

The IPO of SpaceX, very much an AI play with an intergalactic twist, is being sized at USD 75bn, not including any greenshoe or possible upsize; last week, Nasdaq-listed Google parent Alphabet unveiled a USD 85bn raise capital raise to help fund its AI-spend.

There are also the potential listings of Claude developer Anthropic, and OpenAI, heavily speculated to be undertaken this year as well. If both deals are priced in 2026, issuance statistics would skyrocket.

Global ECM issuance this year is now the second highest in a decade, just after the post-pandemic catch-up-fuelled run seen in 2021.

While global public equities perhaps represent the deepest and most liquid investor base in the world, the equity capital markets remain a highly concentrated buyer pool, with top global mutual funds and shareholders commanding the lion’s share of allocations in most deals deal.

“The capacity required for all this is crazy,” said an ECM banker. “We have tried to model it a bit this year before Alphabet, and we thought there was going to be significantly over USD 200bn of AI issuance this year before that deal was announced.”

The combined AUM of the world’s largest top 20 largest asset managers is roughly USD 65tn, according to calculations by ECM Pulse using publicly available data. Assuming an average cash reserve of roughly 3%, this would represent some USD 2.1tn in available liquidity.

If we further assume 60% of the USD 300bn AI issuance wave is allocated to the world’s top twenty global shareholders, a fairly low number in a typical large-scale IPO given issuer preference for long-term fundamental investors over hedge funds, they’d be buying around USD 180bn in stock.

Of course, there may be far more equity capital markets paper to come from SpaceX alone, given a staggered lock-up period across 2026.

It’s conceivable that participation in these offerings could slash cash reserves at these largest asset managers by roughly 10%.

Several investment firms would be unwilling to take such a hit to their war chests and might instead choose to sell other holdings to take part.

While there has been some speculation of a churn in the Magnificent 7, the impact of this equity rotation could be far more widely spread.

“We have had several European CEOs in completely unrelated sectors worried about investors selling their stocks just to create some liquidity to part in these huge AI raises,” the banker added.

AI cash bonfire

The rush by AI hyperscalers to raise equity is a change from a market where the perception was that cash rich mega-caps – like Alphabet – were set to fund their AI expansion through their own cash reserves. Now it seems equity markets are required to stump up a significant part of the bill.

Alphabet is the second-listed leviathan to raise capital for AI expansion this year, following on from Oracle in February.

Not only is Alphabet’s equity raise the largest ever, but it is also a highly unusual move for a business which has not raised any equity capital in over 20 years, according to Dealogic data.

“Alphabet has exhausted free cash flow as a source of funding its AI buildout and it is reluctant to tap debt markets further, at least for now. It clearly sees an opportunity to raise cheap capital,” said a fund manager. “This equity raise is coming together while its stock price is near a record high.

“It is dilutive now, but it can buy stock back later during a market correction, even using cash flow from its AI ROI if it comes to fruition. The key question is whether that AI spend ultimately translates into future cash flow.”

The fund manager noted that across the Mag 7 and the hyperscalers, strong free cash flow generation and share buybacks have been central to their valuation expansion. He added while he thought Mag 7 businesses would see a return on investments, he predicted that might not be for next 12 to 18 months.

“In the interim, there could be a pullback in their valuations or a flatter market performance until their ROIC becomes clearer,” the fund manager added.

This might indeed lead to more AI issuance in the near future as Mag 7 hyperscalers take advantage of high equity valuations before any market pull-back.

Following the Alphabet deal, it was reported that Meta was also now exploring a mega equity raise of its own to keep up in the AI arms race.

That same impetus is driving the wave of mega-IPOs.

As this column noted two weeks ago when analysing SpaceX’s S-1, the huge capex required to fuel expansion for artificial intelligence is likely to require multiple rounds of equity raising. Anthropic and OpenAI are both rushing to market because they need further supplies of cash, rather than simply a desire to become public companies.

AI’s enormous capital demands are “changing the debate about companies wanting to stay private for longer,” said the banker. The IPOs of these businesses would likely be just the opening salvo of equity raises for their gargantuan technology spend.

For now, while the market is in its infancy there is still considerable excitement around these investments with different investors seeking to back different horses in the AI race.

“The AI field is still in relatively early stages; the winners aren’t clear yet,” noted an ECM advisor. “As long as it remains open, when serious players go out to raise, people are interested.

“There’s still enormous capital out there, and people have their own theories on who’s going to win.”

But there is a question about how long investors will happily stump up cash to fund almost unimaginable levels of spending.

Goldman Sachs estimates that around USD 7.6tn of capital expenditure is set to be deployed over 2026-2031 across compute, data centers, and power.

“Investors are starting to get concerned that this is the biggest capital bonfire that they have ever seen and every bit of money going to be burned and companies will come again for more,” the banker added.

Client concern prompts bubble fears

For such huge AI spend to make sense for investors, hyperscalers need to follow-through on the world-transforming promises they are laying out to investors.

The SpaceX IPO is a wonderful microcosm for this, particularly when looking at the IPO research of Goldman Sachs, the lead-left bank on the IPO, a claim which has caused some consternation among its Wall Street rivals, who point to alphabetical bias.

Goldman analysts tell investors not to worry about SpaceX’s mind-boggling numbers now because its AI revenue is set is set to grow to USD 322bn by 2030, from USD 3.2bn in 2025 – 100 times in just five years.

No business in the history of mankind has grown revenue that fast. But assuming SpaceX becomes the monopolistic provider of AI technology by 2030, and adoption remains ubiquitous across all industries, the number might look plausible, with a bit of imagination.

However, there are fears from customers that the technology is becoming too expensive and yielding too little.

As ECM Pulse wrote last week, Uber – one of the most enthusiastic adopters of artificial intelligence in corporate America – had burnt through its entire 2026 AI budget in just four months.

The company also noted it was failing to draw a clear line between AI investment and improvement in final customer offerings.

As ECM Pulse’s M&A equivalent Continental Drift has argued, CFOs managing inflation-related margin impacts will demand clear ROI from AI token usage, potentially trimming pioneers’ revenue growth.

OpenAI CEO Sam Altman noted in a summit last week that token usage was accelerating at an exponential rate and that its clients were now beginning to express real concerns over the costs, something he said had not been the case before.

If AI adoption slows or just plateaus, then the exponential revenue growth that these businesses need to justify their bullish projections will come under pressure at the same time as this huge spike in capex.

“If at some point companies think there is no value in AI over a human doing the same task, given the expense and final result, then this could be a huge problem for some of these AI businesses,” noted the banker.

Whether AI does accelerate at the rate its advocates promise or whether it is a bubble that finally bursts, global equity capital markets are in for a wild ride either way.