Since Oracle kicked off the hyper-scaler borrowing spree in September, one of the hottest debates in market circles was whether and to what extent AI capex plans would impact the IG corporate credit market.
Over nine months following the $18 billion Oracle offering that started it all, Alphabet, Amazon, Meta and Oracle have sold more than $200 billion of debt between them, and not just in the US.
In February, Alphabet shattered UK and Swiss corporate credit records with huge sterling and franc offerings. A month later, Amazon completed the largest-ever corporate bond sale in the euro market, raising almost $17 billion in an eight-part deal.
In May, Alphabet broke the record for a CAD sale with a four-part, $6.2 billion offering that was more than three times oversubscribed. Then promptly set a new record for a foreign corporate issuer in Japan with a $3.6 billion yen debut.
The figure above gives you an idea of the cadence, size and scope.
Although it’s fair to ask if the overseas borrowing and modest price concessions (most tranches in Meta’s six-part April offering priced wide of the company’s October deal, for example) suggest fatigue’s setting in, all this supply was readily absorbed.
Yes, tech spreads widened and Oracle’s ballooning CDS made headlines intermittently for the better part of four months, but by and large, capital markets proved more than willing to finance a big chunk of hyper-scaler capex. And thank God. Because those outlays would otherwise consume almost every dollar of free cash flow.
The chart above, from Nomura, shows you the impact of the AI buildout on the ratio of cash flow to capex at the aggregate, index level. Were it not for debt sales, the figure for 2026 would be much higher than 46%.
Now, there’s a new wrinkle: Alphabet’s raising equity, and quite a bit of it. $80 billion to be precise about things, a staggering sum that surely counts among the biggest such packages on record.
$40 billion of that’s an ATM offering, with the remainder split between $30 billion of underwritten convertible preferred and common, and a $10 billion private placement with the peanut brittle people.
In a press release announcing the follow-ons, Alphabet said it’ll use the money to fund “capital expenditures to scale AI infrastructure and global compute.” The company spun the offerings — not disingenuously — as a prudent way to diversify the funding mix “at an expansionary moment.”
“The company is experiencing strong demand for its AI solutions and services at levels that are exceeding the company’s available supply,” Alphabet went on to explain, describing the equity offering as “part of [a] plan to fund its investments in a balanced way while retaining a healthy balance sheet.”
For context, Alphabet’s operating cash flow was $174 billion in the 12 months to end-March and its capex plan for the full-year 2026 is $185 billion at the mid-point. The company plans to spend “significantly” more than that in 2027.
The chart above never ceases to amaze. Depending on how much faith you have in Jensen Huang’s “new industrial revolution” narrative, we’re really out on a limb here.
“[Alphabet’s] combined debt and equity raise [in 2026] accounts for 71% of the capital expenditures Google announced for this year,” JonesTrading’s Mike O’Rourke remarked, tallying up all the borrowing mentioned above with the newly-announced equity program. “The company had already increased its debt load by 87% this year and is now set to issue equity equivalent to nearly double the value of $45.7 billion of last year’s share repurchases,” he added.
The implication’s clear: This is getting pretty excessive. “Escalatory” may be an understatement.
The figure, from Goldman, is one of my favorites so far in 2026. It speaks for itself, but I added a prominent, red annotation. You know, just in case.
In the same note quoted above, O’Rourke conceded the obvious: When your stock’s up 150% in the space of 12 months, as Alphabet’s is, and people are chasing the rally in names associated with a revolution you’re well-positioned to lead, it’s a good time to sell equity. Still, he cautioned, “when a free cash flow machine like Google raises this much money, it should be at least mildly concerning.”
Another, broader, concern wonders about the ramifications for equity supply and demand in the event this becomes a trend. Imagine, for example, that other big players in the AI race follow Alphabet’s lead at the same time Anthropic, OpenAI and SpaceX go public. That’s a lot of stock, and not all of it’s cheaply (or even fairly) valued.
As Reuters put it early Tuesday, “The question is where [this] leaves the leaderboard of top companies, relative index weightings and the concentration of AI in overall equity benchmarks.” Investors, the same linked piece went on, would do well “to remember that huge IPO waves in the past have sometimes marked the [peak] of speculative frenzies.”






I remember when Viacom raised. Something happened thereafter. Epic.
For GOOGL it’s a completely different story and it’s smart from their side. If I would be a shareholder, I would consider it a better option than issuing debt as well. This gives them a free option (minor dilution) to chase their plan. Will be more interesting to see what happens when more questionable / cyclical businesses caught up in the AI trade issue a significant amount of equity.
I’d be curious to see earnings growth scenarios using different depreciation timelines. I know Burry has been doing his finding the next bubble routine and maybe he already projected it out, but I do continue to wonder what happens to these massive investments as chips and LLMs get better. Given my penchant for being wrong, it probably means scaling even greater heights.
I also wonder how this has changed the economics of crypto mining. That still wins my award for dumbest use of compute power and energy in history.
So, you think the equity prices are correlated to earnings? 🙂