GM. This is Milk Road Macro, the newsletter that watches Big Tech build data centers the way kids stack LEGO, except the pieces cost $471B and someone’s gonna finance the last brick with debt.
Here’s what we’ve got for you today:
- ✍️ Taking a look at the cost of the AI boom
- 🎙️ The Milk Road Macro Show: Why 2026 Will Be a Volatile, Politically Driven Market w/ Mel Mattison
- 🍪 Trump teased his preferred pick for the next Fed Chair
Decentralized Token Folios (DTFs) on Reserve are indices that bundle multiple crypto assets into a single token. Read more about DTFs here.

Prices as of 10:00 AM ET.

TAKING A LOOK AT THE COST OF THE AI BOOM
The “AI revolution” effectively started when OpenAI’s ChatGPT (version 3.5) was first launched for public use on November 30, 2022.
This single event sparked a frenzied stampede of mega-cap tech companies falling over each other to join the race for AI supremacy.
It’s been a wild ride since then.
And it’s about to get even more wild.
Spending on data centers and AI infrastructure is exploding as the AI build-out scales up.
And more and more debt is now coming into the funding picture.
Let’s take a look at the ridiculous numbers involved in the AI boom.
And ask:
Is there a debt problem?
Is it all worth it?
And are the Magnificent 7 still a good bet?
Big spenders
As the AI boom has intensified, the costs associated with the next big technological advance have soared.
Gargantuan data centers and AI infrastructure are expensive.
So capex trends among hyperscaler tech firms are growing rapidly.
In 2026, five big spenders (Amazon, Microsoft, Google, Meta, Oracle) are expected to splurge approximately $471bn on capex.

These tech giants are cash-printing behemoths - and up until very recently, much of this AI spending has been funded through these big cash flows.
But these huge capex numbers mean more and more of the spare cash is being used for AI-related spending.
And so the world's largest companies have had to start taking on more debt - at a scale much bigger than previous years.
And this has ramped up in recent weeks.

Concerns have been mounting about the growing amount of debt.
This is particularly true of Oracle.
Credit Default Swaps* for Oracle have been rising.
*A Credit Default Swap is a financial instrument that is essentially “insurance” against the default of a borrower.

These charts have been all over social media and are very exciting for AI skeptics.
But here’s what this Oracle Credit Default Swap chart actually means: a 1.93% probability of default per year and a 9% 5-year cumulative probability of default.
So, it’s not nothing - but it’s also not “let’s all panic and run around screaming” either.
Three years after ChatGPT made its first appearance, people are now asking:
- How profitable will AI ever really be?
- Will it ever cover the costs to keep it afloat?
- At what point will capex spending soak up all the cash flow from operations?
- How much debt can the world's largest companies take on before the market rebels?
- Is direct government assistance needed?
CRYPTO INVESTING AS A SMOOTHIE
Buy all the ingredients. Chop it all up. Blend it. Add sugar. That’s how long it takes to make a smoothie.
Or… you just grab a ready-made one.
Well, Reserve is basically building the ready-made smoothie of crypto.
Decentralized Token Folios (DTFs) on Reserve are indices that bundle multiple crypto assets into a single token.
Here’s why we’re impressed by DTFs:
- Backed by Sam Altman & Peter Thiel.
- Partnered with Bloomberg, CoinDesk and CoinMarketCap.
- Just launched CMC20, the crypto version of the S&P 500.
No more managing complex portfolios, just buy DTF & chill.
Read more about the DTFs here.

TAKING A LOOK AT THE COST OF THE AI BOOM (P2)
Is there a debt problem?
According to an analysis by JP Morgan, $5.3 trillion is likely to be spent on data center and AI infrastructure, and related power supplies, globally over the next five years.
So where will all the money to pay for this come from?
JP Morgan estimates annual data center funding needs in 2026 will be roughly $700 billion - which could be entirely financed by hyperscaler cash flows and a little help from High Grade bond markets.
But the annual funding needs looking further ahead just get bigger and bigger.
And could reach in excess of $1.4 trillion per year by 2030, which will likely require funding contributions from across all capital providing markets.

JP Morgan calculates that, as of this moment, tech hyperscalers are generating over $700bn of operating cash flow per year, and reinvesting ~$500bn of that back into capex.
If we also take into account $250bn of research and development spending per year, the Magnificent 7 complex (Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta, Tesla) could be fast approaching a cash flow breakeven, as a whole.
So if free cash flow is now potentially nearing a tipping point, what does that mean?
Simple: enter the debt!
Now, of course, debt isn’t necessarily “bad” - but growing debt loads change the shape of the AI conversation entirely.
JP Morgan predicts that, after cleaning out all available debt markets, there could still be a roughly $1.4 trillion “funding hole” left.
So who will plug the funding gap?
JP Morgan thinks it might have to be “assistance from Governments”:
"Government involvement has run the gamut from aggressively and publicly supportive in the U.S. to a safety-focused regime in the EU. More aggressive financial support by governments is possible, particularly if/when national defense concerns around AI grow."

Is it worth it?
There are now massive amounts of money on the line.
Big bets are being taken, and those bets may have to be backstopped by governments.
But will they pay off?
JP Morgan estimates: "To drive a 10% return on modeled AI investments through 2030 would require ~$650 billion of annual revenue into perpetuity, which is an astonishingly large number."
For context, that equates to $34.72/month from every current iPhone user, or $180/month from every current Netflix subscriber.
AI will undoubtedly reshape the world - but a lot of revenue needs to come from somewhere.
Some people have made comparisons between the AI build-out and the disastrous “over-build-out” of telecom and fiber networks in the late 1990s.
But this is far from the truth.
Back then, big networks were being built without the demand - and it took years to “use up” the infrastructure built during the dot-com boom.
But today, demand for AI compute vastly outstrips supply.
It’s very unlikely that there will be any “over-building” in the AI sphere for many years to come.
Nevertheless, JPM concludes that even if everything works out perfectly, there will be (continued) spectacular winners, and probably some equally spectacular losers as well, given the amount of capital involved and “winner takes all” nature of portions of the AI ecosystem.
So, is the "Magnificent 7” still a good bet?
According to legendary investor and Wall Street research veteran Ed Yardeni: maybe not.
Yardeni recently recommended that investors effectively go underweight the Magnificent 7 megacap technology stocks versus the rest of the S&P 500.
While the Magnificent 7 are building and funding the AI tools - it is the rest of the S&P 500 that may ultimately benefit more from those tools, according to Yardeni.
“We see more competitors coming for the juicy profit margins of the Magnificent 7 and expect that the productivity and profit margins of the rest of the S&P 500 will be boosted by tech”, Yardeni said.
He added that in effect, “every company is evolving into a technology company”.
For 15 years, Yardeni has been a prominent cheerleader for the tech sector and the Magnificent 7.
He’s been recommending investors adjust exposure to be overweight the Information Technology and Communication Services sectors.
And he’s been right.

But now, for the first time since 2010, he no longer thinks this makes sense.
Instead, he recommends dialing down exposure to those sectors, and instead going overweight financials, industrials and healthcare - sectors he believes will benefit most from the AI boom.
The Magnificent 7 in particular merits special caution as “they’re competing more aggressively against each other and they’ve got more competition coming out of nowhere”, Yardeni said.
Wrapping up
It’s pretty clear - the amount of dollars involved in fueling the AI boom is enormous.
And it’s likely that these numbers will just get bigger and bigger.
This means more and more debt will be needed to finance the revolution (not necessarily a bad thing) - and maybe even assistance from governments.
While the jury is out on whether or not the huge AI capex spending will eventually be recouped - it is probably worth pondering whether mega-cap tech companies are the best way to position for the AI age from an investment perspective.
They’ve dominated the stock market for more than a decade - but that dominance might start to fracture in the coming years.
That’s it for this edition.
Before you go, since this is our last one of the year, we’d love your quick take on Milk Road Macro.
Tap a rating below, then leave a short note in the comments with any feedback. We read every note, and it helps us make the newsletter better for you!
What'd you think of today's edition?

2026: VOLATILE, POLITICAL, OPPORTUNISTIC 🚨
In today’s episode, we sat down with Mel Mattison, writer, investor, and fintech executive, to talk about why 2026 is shaping up to be a politically charged, high-volatility market with major macro shifts on deck.
Take a look at what surprised us:
- Why the Fed’s new T-bill purchase program isn’t just housekeeping, and what it signals for liquidity in 2026.
- How housing policy and fiscal stress will shape market incentives ahead of the US midterms.
- Why volatility is the new normal, and where stock-picking will matter most.
- What comes next for crypto, AI investments, and tokenization after a wild 2025.
It’s a banger of an episode, don’t miss it 👇
YouTube | Spotify | Apple Podcasts

Bridge helps businesses send, store, accept, and launch stablecoins instantly. Learn how Bridge is powering stablecoin-backed money movement.
Chainlink is the industry-standard oracle platform bringing the capital markets onchain, and the market leader powering the majority of decentralized finance. Learn more about Chainlink.
Fuse is a $300M ARR energy company launching The Energy Network on Solana. Discover the future of energy now.

BITE-SIZED COOKIES FOR THE ROAD 🍪
Silver has been the success story of the past few weeks - with a ferocious rally seeing the price of the metal nearly double since late October. Here’s everything you need to know about the silver frenzy.
Meta has agreed to buy Manus, a popular Singapore-based AI agent with Chinese roots. The deal values Manus at more than $2 billion.
President Donald Trump teased that he has a preferred candidate to be the next Federal Reserve Chair. Trump said: “I’ll announce him at the right time - there’s plenty of time”.

MILKY MEME 🤣


ROADIE REVIEW OF THE DAY 🥛













