I’ve been a trader and investor for 44 years. I left Wall Street long ago—-once I understood that their obsolete advice is designed to profit them, not you.
Today, my firm manages around $4 billion in ETFs, and I don’t answer to anybody. I tell the truth because trying to fool investors doesn’t help them, or me.
In Daily H.E.A.T. , I show you how to Hedge against disaster, find your Edge, exploit Asymmetric opportunities, and ride major Themes before Wall Street catches on.
Table of Contents
H.E.A.T.
THE CONSENSUS GOSPEL
Wall Street is still arguing about the wrong thing. Is AI real? Is Nvidia too expensive? Are we in a bubble? Those questions are entertaining. They are also secondary.
The real question — the one that decides who wins and loses over the next 24 months — is deceptively simple: Who gets paid while the concrete is still wet?
Because AI is no longer just a software story. It is the largest infrastructure buildout since the telecom boom of the late 1990s — except this time the customers are the richest companies on Earth, and the spending isn't optional.
McKinsey's analysis puts global data-center capital spending at roughly $7 trillion by 2030, with AI driving the bulk of it. Add the hyperscalers' own declared capex plans and the all-in ecosystem figure — power, grid, chips, land, cooling — approaches figures not seen since the great infrastructure booms of the 20th century.
Meta. Google. Microsoft. Amazon. Oracle. Five companies. Trillions in planned capital expenditure over five years. The formerly debt-light Google borrowed $32 billion from the bond market. Meta issued $30 billion in bonds in late 2025 alone. These are not speculative bets by underfunded startups. These are the most profitable corporations in the history of capitalism — and they are going all-in.
| "The biggest spenders may not make their money back first — but the companies that sell them the concrete, the copper, and the cooling will almost certainly get paid either way." |
|
THE CRACK IN THE GOSPEL
Here is the problem. Someone has to earn that capital back.
Work through the math. A $7 trillion investment, at a reasonable 10% cost of capital, requires $700 billion in annual profit — just to break even on a risk-adjusted basis. To generate that profit at a one-third margin implies the need for more than $2 trillion in new annual revenue. That is a significant fraction of the entire U.S. software market in 2025. It has to appear, largely from scratch, within the next several years.
For context: Wells Fargo analysts estimate that AI accounts for approximately 10 percentage points of Meta's 25% advertising growth — roughly $20 billion annually. Annualize that across every hyperscaler's most optimistic scenario, and you still fall well short of the required revenue math. The rest must come from new products, new markets, and new behaviors that don't fully exist yet.
Timing matters enormously when you have debt to service. A landmark MIT study, cited extensively by the Harvard Business Review, found that roughly 95% of enterprise GenAI programs currently fail to deliver bottom-line returns. The telecoms bubble of the late 1990s collapsed not because the internet was fake, but because traffic was doubling annually — not every 90 days as the models assumed. That single miscalculation destroyed hundreds of billions in capital. A similar mistiming here doesn't destroy Google. But it could severely punish any investor who treated hyperscalers as risk-free compounders.
| "Around 95% of enterprise GenAI programs currently fail to deliver bottom-line returns. The telecom bubble didn't collapse because the internet was fake — it collapsed because the timing was wrong. History doesn't repeat. But it rhymes loudly." |
|
THE MECHANISM — WHAT ACTUALLY DRIVES RETURNS
The critical insight most analysts miss: the AI infrastructure boom has two very different types of beneficiaries, operating on completely different risk profiles.
The construction clock runs now. Contracts get signed today. Equipment gets ordered today. Power gets reserved today. This is where the suppliers collect.
The monetization clock runs later. The profits show up after the buildout is complete. This is where the spenders get judged — and where the market hands out trophies or punishment based on how closely the cash flow matched the capex.
The first group — the hyperscalers themselves — faces genuine re-rating risk. If AI monetization accelerates as promised, their massive capex looks prescient. If it stalls, they have locked themselves into multi-decade fixed-asset commitments funded partly by debt, with shrinking free cash flow to return to shareholders in the interim. Meta's fixed-asset turnover has already cratered to just $1 of revenue per $1 of property and equipment, versus over $8 at Apple. That spread is about to get wider.
The second group — power providers, chip manufacturers, cooling infrastructure, and copper suppliers — faces an entirely different equation. They are not betting on which AI model wins. They are not betting on whether Meta's ad CPMs rise 10% or 25%. They are simply selling essential inputs into a construction boom that is already underway, contracted, and funded. Their customers are the most creditworthy companies on Earth. Their product — power, compute, copper, cooling — is consumed regardless of whether the final AI application ever turns a profit.
This is the picks-and-shovels thesis, updated for the third decade of the 21st century. Levi Strauss made more durable money selling denim to Gold Rush miners than most miners ever extracted from the California foothills. The question is not whether trillions will be spent. Most of it is already committed. The question is who gets paid on the way in — before we find out whether the gold is actually there.
~$7T Data-center capex by 2030 McKinsey | $900B Annual profit to break even FT / Wharton | 125 GW New capacity needed by 2030 McKinsey | 95% GenAI programs that miss ROI MIT/HBR |
🔎 SPOTLIGHT: THE ANTI-TRADE Apple (AAPL) — The Balance Sheet Nobody Is Talking About Apple is the most interesting company in technology right now precisely because of what it is not doing. While Alphabet earmarks roughly $185 billion for capital expenditure this year alone, Apple has maintained one of the leanest balance sheets in megacap tech. Its fixed-asset turnover — over $8 of revenue per dollar of property and equipment — towers over Meta's $1 and Amazon's $2. The thesis is clean: if AI migrates to the edge — models running locally on iPhones and Macs, barely touching a data center — Apple's device ecosystem becomes the most valuable AI distribution network on Earth. If AI remains cloud-native and centralized, Apple risks being locked out of the platform entirely. Carlyle's head of strategy Jason Thomas described it as 'a binary option.' The balance sheet optionality is real and measurable: the gap between what Apple is not spending and what its peers are deploying represents one of the most unusual strategic positions in the history of American technology. |
INVESTMENT IMPLICATIONS
TIER | TICKER | COMPANY | WHY IT MATTERS |
▲ WINNERS — Structural Beneficiaries (Paid on the Way In) | |||
★★★ | NVDA | Nvidia | Every dollar of hyperscaler capex buys chips first. H100/B200 GPUs are the irreplaceable compute unit — backlog secured regardless of which AI model ultimately wins. |
★★★ | VST | Vistra Energy | 125 GW of new data center capacity = structural baseload power demand. Vistra's nuclear + gas fleet is already contracted to hyperscalers — demand is baked in, not speculative. |
★★★ | CEG | Constellation Energy | Co-locating nuclear plants next to hyperscaler campuses. Clean, always-on power is the scarcest single input in the AI buildout. Microsoft's 20-year deal validates the model. |
★★ | EQIX | Equinix | Colocation REIT with 260 data centers globally. Benefits from hyperscaler overflow and enterprise hybrid-cloud migration — protected by long-term leases. |
★★ | FCX | Freeport-McMoRan | Each GW of data center capacity consumes ~40 tonnes of copper. FCX is the purest large-cap exposure to the physical buildout constraint. |
★★ | CRWV | CoreWeave | The largest independent GPU cloud — OpenAI's preferred data center landlord. Captures overflow demand when hyperscalers can't build fast enough. |
★ | AAPL | Apple | The only Big Tech holding a lean balance sheet while peers drown in capex. Edge AI tailwind + significant optionality without touching a server farm. |
▼ PRESSURE POINTS — Margin & Timing Risk (Paid Later, Judged Harder) | |||
⚠⚠⚠ | MSFT | Microsoft | Capex ballooning; stock fell 10% despite beating estimates. Revenue backlog is real but the cost-of-capital clock is ticking on $185B earmarked for AI infrastructure. |
⚠⚠⚠ | META | Meta Platforms | Zuckerberg admits spending for 'the most optimistic cases.' $620B capex over 4 years. No cloud revenue to offset — every dollar must return through ad CPMs or Llama licensing. |
⚠⚠ | GOOGL | Alphabet | Borrowed $32B to fund data centers. Fixed-asset yield cratering. If AI lifts ad revenue by only 10 points (Wells Fargo est.), the math runs thin against $185B+ capex. |
⚠⚠ | AMZN | Amazon | Revenue backlog doubled, but so did the capital required to service it. AWS margin expansion now competes with a multi-hundred-billion construction bill. |
⚠ | OpenAI (pvt) | OpenAI | Stargate spend moderated from $1.4T to $600B — still existentially leveraged. No real business to fall back on if AI monetization stalls. Highest-risk entity in the stack. |
★★★ High Conviction ★★ Moderate Conviction ★ Speculative / Optionality | ⚠⚠⚠ High Pressure ⚠⚠ Moderate Pressure ⚠ Watch
⚠ WHAT WOULD BREAK THIS THESIS 1. Demand proves slower than expected — MIT/HBR research: ~95% of GenAI programs currently fail to deliver bottom-line returns. The payback cycle is longer than the sales pitch. 2. DeepSeek / efficiency shock — If compute-per-query keeps falling 10x every 18 months, the capacity being built today becomes massive oversupply before contracts roll off. 3. Interest rate re-acceleration — Google and Meta are now meaningfully leveraged. A 200bps rise in long-end rates raises the ROI hurdle materially — and shrinks free cash flow available to shareholders. 4. Regulatory intervention — Data center energy consumption is approaching national-grid levels in several EU jurisdictions. Permitting freezes and carbon mandates are credible tail risks. 5. Chip obsolescence cycle — Nobody knows how long H100/B200 silicon stays competitive before the next architecture renders it worthless. Capital tied up in depreciating silicon is not recoverable. |
5 KEY TAKEAWAYS | |
1 | The picks-and-shovels play is cleaner than the model play. Nvidia, power utilities, and copper producers get paid on every dollar of data center capex — regardless of which hyperscaler wins the AI arms race or whether monetization materializes on schedule. |
2 | Power is the binding constraint. 125 gigawatts of new data center capacity by 2030 requires roughly adding three United Kingdoms of electricity demand. Nuclear-adjacent utilities (Vistra, Constellation) are the single most under-discussed beneficiary of the entire AI buildout. |
3 | The hyperscalers are not risk-free. Google, Meta, and Microsoft are now meaningfully leveraged to AI monetization timelines they do not fully control. Free cash flow compression is real — any demand slowdown pressures their ability to reward shareholders in the near term. |
4 | Apple's balance sheet is a call option on being right. If edge AI is the future — models on devices rather than in data centers — Apple wins the platform war without spending a dollar on server farms. The gap versus Alphabet's capex budget could become its greatest competitive advantage. |
5 | OpenAI is the wildcard that could reprice the entire sector. With no real business to fall back on and a self-sacrifice clause in its charter, OpenAI is the highest-risk entity in the stack. A stumble there — financial or strategic — would send shockwaves through every company holding OpenAI supply agreements. |
News vs. Noise: What’s Moving Markets Today
Iran’s President came out yesterday and said his country has the necessary will to end the war. That, plus quarter end, was enough for an almost 3% rally. So far this morning more green in stocks and red in oil.
If you look at a chart of SPY you have a couple of undercut and rally points, the 10 day moving average and the November 21st low…..
My preffered way to play this is still the gold miners. If the war is really over I think you get a rally in gold, if it isn’t, I think you can still get a rally in gold……
Just remember though, the best time to buy gold miners is when nobody wants them. I was pounding the table on this trade a week or so ago, now I am seeing it show up in Wall Street research reports.
There were actionable long entry moves all over the place yesterday in the areas we have been watching—-optics, memory, semi’s, software, etc. The question is how much of that was just the machines buying the headlines and end of quarter re jiggering, and how much is real? Remember, the AI trade was wobbly before the war, if anything it should be worse now. I’d be very cautious.
ETF News
MEMY Holdings Update:
Sold $STX ( ▲ 8.01% ) $CEG ( ▲ 0.08% ) $LLY ( ▲ 3.78% ) bought $ARM ( ▲ 2.51% ) $AAPL ( ▲ 0.73% ) $PPTA ( ▲ 4.98% ) All 5% positions.
For a full list of MEMY holdings, visit:
https://incomeblastetfs.com/etf/memy
Distributor: Foreside Fund Services, LLC
A Stock I’m Watching
Today’s stock is Apple $AAPL ( ▲ 0.73% ) ……
It’s interesting to me on a number of levels, first is what I talk about above. Unlike some of it’s Mag 7 friends, it’s not spending tons of money on AI. This could end up good or bad, but I think they have earned the right to have me believe they know what they are doing. Undercut and rally at the 200 day moving average yesterday also.
In Case You Missed It
I had the pleasure of talking to Dividend Degenerates on why I like put spreads better than covered calls for income…..
The H.E.A.T. (Hedge, Edge, Asymmetry and Theme) Formula is designed to empower investors to spot opportunities, think independently, make smarter (often contrarian) moves, and build real wealth.
The views and opinions expressed herein are those of the Chief Executive Officer and Portfolio Manager for Tuttle Capital Management (TCM) and are subject to change without notice. The data and information provided is derived from sources deemed to be reliable but we cannot guarantee its accuracy. Investing in securities is subject to risk including the possible loss of principal. Trade notifications are for informational purposes only. TCM offers fully transparent ETFs and provides trade information for all actively managed ETFs. TCM's statements are not an endorsement of any company or a recommendation to buy, sell or hold any security. Trade notification files are not provided until full trade execution at the end of a trading day. The time stamp of the email is the time of file upload and not necessarily the exact time of the trades. TCM is not a commodity trading advisor and content provided regarding commodity interests is for informational purposes only and should not be construed as a recommendation. Investment recommendations for any securities or product may be made only after a comprehensive suitability review of the investor’s financial situation.© 2026 Tuttle Capital Management, LLC (TCM). TCM is a SEC-Registered Investment Adviser. All rights reserved.