The AI boom has a funding problem that doesn’t show up in the usual headlines. Everyone talks about model releases, chips, and data centers. Very few talk about the fact that a growing share of this entire buildout is sitting on corporate credit – and that credit is starting to flash yellow.
Across the AI supply chain, companies are borrowing at a pace that looks less like normal capex and more like a race against time. Tech giants from Alphabet to Oracle have nearly doubled their investment-grade bond issuance this year to roughly the high-hundreds of billions, a record haul driven heavily by AI and cloud infrastructure. At the same time, private credit funds have stepped in alongside public markets, lending well over a hundred billion to finance data centers that are then leased back to the same tech groups.
It isn’t just one name. Meta recently sold on the order of tens of billions in unsecured notes, the largest single high-grade tech bond sale of the year, to support AI and infrastructure spend. Amazon returned to the bond market for the first time in years to raise well over ten billion for its own AI-driven cloud and logistics investments. Oracle, meanwhile, has become the poster child for the more aggressive end of the spectrum, linked to a pending $38 billion loan package and a series of bond deals tied directly to AI data centers serving OpenAI.
The common thread is simple: the AI buildout is no longer being funded primarily by cash flow. It’s being funded by leverage.
$57 Billion in NVDA Revenue, 62% YoY Growth. And stocks still fell… What now?
Nvidia just posted a record-breaking quarter… yet the markets dropped. Why?
Experts say that even the top AI earnings couldn’t calm the fear of a potential bubble.
After soaring at the open, the S&P reversed sharply, wiping out over $2T of value in hours.
The “Great Bitcoin Crash of 2025” only wiped out ~$1T by comparison.
Wall Street’s finally asking: What if AI isn’t enough?
So, where can investors diversify when public markets stop making sense?
Now, for members-only → blue-chip art.
It’s not just for billionaires to tie the room together. It’s poised to rebound.
With Masterworks, +70k are investing in shares of multimillion dollar artworks featuring legends like Basquiat and Banksy.
And they’re not just buying. They’re selling too. Masterworks has exited 25 investments so far, including two this month, yielding net annualized returns like 14.6%, 17.6%, and 17.8%.*
My subscribers skip the waitlist:
*Past performance is not indicative of future returns. Investing involves risk. Reg A disclosures: masterworks.com/cd
That would be less worrying if credit markets were relaxed. They’re not. Investment-grade and high-yield corporate spreads have been widening in recent months, even as Treasury yields fall and markets price in future rate cuts. Some long-dated corporate bonds now trade at notable discounts, and credit strategists are flagging a structural shift: investors are demanding more compensation to hold corporate risk, even in the parts of the market that are supposed to be safest.
The AI complex sits right in the middle of that tension. On one side, you have enormous projected demand: trillions of dollars in infrastructure spending over the next decade and a belief that AI will reshape productivity, revenue, and margins. On the other side, you have a financing structure where much of that spending is being pushed onto the balance sheets of partners, data-center developers, and cloud providers, while the revenue tied to those commitments is still ramping.
Oracle has become the cleanest expression of that risk. Its debt load has grown sharply as it positions itself as a key infrastructure provider to OpenAI, and banks are now assembling record-sized loan packages to fund sites dedicated to that relationship. Bond investors have pushed its yields higher, and the cost of insuring its debt has become a widely watched way to bet against the more leveraged side of the AI story. In effect, Oracle’s credit profile is turning into a proxy for how much strain the AI borrowing binge can take.
Zoom out, and the pattern is bigger than any one company. Globally, firms tied to AI – from chip makers to data-center operators to cloud platforms – have raised hundreds of billions this year across bonds and private credit. Analysts are beginning to worry that this wave of issuance is overwhelming demand, pushing down prices on new bonds and quietly tightening financial conditions for everyone else. If the cost of new financing keeps rising, the marginal AI project becomes harder to justify on economics alone.
Equities can ignore this for a while. Narratives are sticky. As long as growth stories are compelling and headline numbers look strong, stocks can trade at elevated multiples. Credit is less forgiving. Bond investors care about getting paid back on time and in full. When they start to demand higher spreads, shy away from long maturities, or use credit default swaps to hedge exposure to specific names, it’s usually because something fundamental has shifted in the risk–reward.
The uncomfortable question hanging over all of this is straightforward: what happens if the revenue and profit from AI don’t arrive fast enough to service the debt that’s being stacked up today?
There are two paths forward. In the optimistic version, AI adoption scales smoothly, companies convert infrastructure into cash flow, and the current wave of borrowing looks like a rational investment in a transformational technology cycle. In the other version, demand grows more slowly than planned, some projects underperform, refinancing costs stay higher than expected, and the weakest balance sheets in the AI world become pressure points for the broader credit market.
Either way, the signal right now is clear. The AI story is no longer just about innovation, chips, and models. It’s about funding. The buildout is running on borrowed money, and that money is starting to ask harder questions.
—
Education, not investment advice.
Sources:
