Mike Lee, founder of Michael Lee Strategy, returns to Miles Franklin Media to revisit his early-2026 silver pullback call -- which proved accurate at roughly 35% from the highs -- and lay out his forward outlook for precious metals and equities. He targets silver at approximately $100 and gold in the mid-$5,000 range within a year, describing a structural reset rather than a collapse. On equities, Lee argues that artificial intelligence infrastructure represents the most widely misunderstood investment theme on Wall Street, with token demand growing on multiple hockey-stick curves simultaneously, and that the AI buildout is nowhere near bubble territory by historical capex-to-cash-flow metrics. His highest-conviction names are Palanteer, Nvidia, Coreweave, and Oracle, each framed around undervalued earnings expectations rather than speculative sentiment. He also acknowledges a K-shaped economy where AI investors thrive while the broader population faces stagflation pressures from the Iran conflict and persistent fiscal deterioration.
Who is Mike Lee? Mike Lee is the founder of Michael Lee Strategy, a former Morgan Stanley strategist who now manages an equity model portfolio and provides market commentary. He has made a public track record of contrarian calls on both precious metals and AI infrastructure, most notably predicting the silver pullback from parabolic highs in January 2026 and maintaining an aggressive bullish stance on AI capex beneficiaries throughout the same period. He is reachable at michaelstrategy.com and @MikeLeeStrat on X.
Key Points
Silver completed a structural reset from its parabolic highs, dropping roughly 35% to the $75 range, and Lee now expects a grind back toward $100 over the next year rather than another parabolic move.
The silver spike was driven by a physical-versus-paper supply demand imbalance, not by industrial fundamentals. That imbalance has corrected. The same dynamic -- artificial rather than fundamental -- is why Lee was cautious at $119.
Gold is expected to recover to the mid-$5,000 range. The recent weakness was partly explained by petrol-state nations selling liquid assets including gold to fund operations during the Iran conflict, creating a temporary artificial suppression.
AI token demand is growing on three simultaneous hockey-stick curves: volume of data in large language models, model complexity, and the transition from frontier models to agentic AI. Token prices doubled in 2026 even after falling 60x from early 2023 to early 2026, signaling that infrastructure cannot keep pace with demand.
Current AI capex as a ratio to S&P 500 cash flow is less than one. At the peak of the dot-com boom it was six. Lee uses this as his primary evidence that markets are nowhere near bubble territory.
Nvidia is described as a "steal" at roughly 14 to 15 times forward earnings versus a market multiple of 22, with 100% year-over-year revenue growth on a $40 billion base when China sales are excluded.
Palanteer is Lee's single highest-conviction name, describing it as "one of one" -- an enterprise AI operating system with 156% net dollar retention, no direct competitor at scale, and a projected revenue path from $8 billion in 2026 to over $60 billion by 2030.
Coreweave is positioned as a high-controversy, high-upside data center play using debt rather than equity to finance GPU capacity. Lee expects a 5x revenue increase from Q4 2025 to Q4 2026, with GPU rental rates actually increasing for older hardware due to token scarcity.
Oracle matters because it controls approximately 35% of the world's enterprise data, making it a structural gatekeeper for any company that wants to run AI on its own data through a large language model.
The Iran conflict is real but characterized as a sideshow for US equity investors, particularly those holding AI infrastructure names. Its main effect has been an oil-driven price shock that is inflationary in the short term but mean-reverting once resolved.
A K-shaped economy is now fully visible. Investors in AI and equities are thriving. Those reliant on wages, driving-intensive lifestyles, or non-asset income are absorbing the cost of higher energy and structurally elevated inflation.
The biggest market risk Lee identifies is an exogenous event -- something not currently anticipated. The internal earnings trajectory is strongly positive, with this past quarter coming in at 28% year-over-year growth against a 13% expectation.
Quotable Moments
Quotable moments are auto-generated from the transcript. Speaker attribution and quote accuracy should be verified against the original source before republishing or sharing.
Mike Lee -- on where we are in the AI cycle
"We are in one of the largest capex cycles probably like in the history of our country and we're nowhere near the end... the way that how we use computers and how we use technology is fundamentally changing in perpetuity. It's happening in front of us and Wall Street does not have their arms around it."
Why it works: Lee anchors his bull case in a concrete historical comparison -- dot-com capex was 6x cash flow; today it is below 1x. The statement lands because it is falsifiable, not aspirational.
Mike Lee -- on trailing multiples in a growth inflection
"Trailing multiples are as useless as you know what on a nun... you don't drive down the highway looking through your rearview mirror the whole time. You may check it occasionally, but if you're going 100 miles an hour, you're looking through the windshield."
Why it works: It cuts straight through the "but it looks expensive" objection that most retail investors hold. The analogy is blunt enough to be memorable and accurate enough to be instructive.
Mike Lee -- on Palanteer's uniqueness
"There's no other company like it in the world... they're going to do 8 billion in revenue this year. That number is likely to be north of 60 in 2030 and it's going to continue to grow. It makes AI actually work for these large enterprises and there's literally only one company that can do it."
Why it works: Lee collapses a complex product thesis into a single claim -- "one of one" -- supported by a specific revenue projection that can be tracked over time. Conviction without vagueness.
Mike Lee -- on the K-shaped economy
"The K-shaped economy is like AI and everything else... if that provides 3 or 4% of GDP growth, we may only grow at 2% because everything else is contracting. But if you're an investor in the S&P 500 or the Q's, you're going to be doing great."
Why it works: Honest and uncomfortable. Lee does not pretend the macro picture is clean. He separates the investor experience from the broader economic one, which gives his bullishness more credibility precisely because he acknowledges what it does not apply to.
Concepts and Ideas
Precious Metals Framework
The Parabolic Reversion Principle
When an asset with no earnings -- like silver or gold -- moves parabolic, Lee argues that the move is almost always driven by a temporary supply-demand imbalance rather than a fundamental re-rating. The correct response is to identify the structural level the asset should trade at on fundamentals, and expect reversion to that level. For silver, that structural level was roughly $75.
Physical vs. Paper Supply Imbalance
The silver spike to $119 was driven by a mismatch between the physical silver market and the paper (futures/derivatives) market. When paper demand outpaces physical delivery capacity, prices can move far beyond fundamental value. Once that imbalance clears, reversion is sharp and fast. This is distinct from an organic demand-driven move and requires a different analytical framework.
Petrol State Liquidity Selling
Lee attributes part of gold's recent weakness to oil-producing nations selling liquid assets -- including gold and US Treasuries -- to fund operations during the Iran conflict. This is a temporary, event-driven suppression of price rather than a structural signal. As conflict resolution approaches and revenues recover, the selling pressure reverses and demand resumes.
AI Infrastructure Investment Framework
Token Demand as the Core Metric
A token is roughly four units of code or text processed by a large language model. Token demand is the foundational unit of measure for AI infrastructure need. Lee tracks it across three simultaneous growth curves: the volume of data fed into models, the complexity of the models themselves, and the transition from large language models to agentic AI. Each curve is independently exponential.
Capex-to-Cash-Flow Ratio as Bubble Detector
Lee's primary anti-bubble argument is a simple ratio: S&P 500 capex divided by S&P 500 cash flow. At dot-com peak, this ratio was six. Today it is below one. This means the market as a whole is spending less than one dollar of capital investment for every dollar of cash flow generated, even as AI spending accelerates. By this measure, the cycle has enormous room to run before bubble conditions emerge.
Forward vs. Trailing Multiples
In a period of earnings inflection, trailing PE ratios are actively misleading. A company growing earnings 100% per year will always look "expensive" on trailing numbers. Lee argues that forward earnings -- which capture the actual trajectory -- show companies like Nvidia trading at steep discounts to the broader market, even after massive stock appreciation.
Net Dollar Retention as SaaS Quality Signal
Net dollar retention (NDR) measures how much revenue a software company collects from its existing customer base in a given period compared to the same period a year prior, including expansions, contractions, and churn. A score above 100% means the company grows revenue from existing customers alone. Palanteer's NDR of 156% at its current revenue scale is described as historically unprecedented for any SaaS company.
Debt-Financed Growth in Infrastructure
Coreweave and others in the "neocloud" category are using debt -- not equity or operating cash flow -- to finance GPU data center buildout. Lee frames this as a rational capital structure decision: when gross cap rates on a data center are 50 to 60% and cost of debt is mid-single-digits, leverage dramatically amplifies returns on invested equity. The market penalizes the debt load without fully pricing the income it generates.
Macro and Political Framework
The K-Shaped Economy
A K-shaped recovery or expansion describes an economy where one segment (typically asset owners, investors, high-skill workers) sees strong upward trajectories while another segment (lower-income workers, high fuel cost households, non-asset owners) experiences stagnation or decline. Lee applies this lens to 2026, arguing that AI and equity investors are thriving while broader wage and cost-of-living pressures intensify for those outside the investor class.
Fiscal Dominance as a Long-Term Floor for Metals
Regardless of which party controls US government, Lee expects spending to remain elevated because fiscal constraints require cross-party dealmaking that always results in higher expenditure. This chronic deficit spending is a long-run tailwind for hard assets including gold and silver, even if short-term price action is influenced by event-driven selling from petrol states or market dislocations.
Exogenous Risk as the Primary Unknown
Lee's candid answer to "what could go wrong" is that the risk is something not yet anticipated. The internal logic of the AI trade is intact. Earnings are dramatically outperforming. The risk is not a known catalyst -- it is an unknown shock to a capital structure that has been extending itself rapidly. This is the honest answer, and it implies that position sizing and hard-asset diversification remain rational insurance even for committed equity bulls.
Implementation
Implementation steps are auto-generated from the transcript content and are provided for informational purposes only. They do not constitute professional advice of any kind. Always consult a qualified professional before acting on any information presented here.
1
Audit Your Portfolio for Trailing Multiple Bias
Lee's "rearview mirror" analogy is a practical diagnostic. Go through any current equity holdings and check whether your buy or sell decisions have been driven primarily by trailing PE ratios. If you are using last year's earnings as the valuation anchor for a company in a growth inflection, you are likely pricing the wrong thing. Identify the forward earnings expectations and stress-test them against a realistic scenario for what AI or other structural tailwinds deliver over the next two to three years.
2
Learn to Track Token Demand as a Leading Indicator
Token demand is the unit of measure Lee uses to assess whether AI infrastructure spending is justified. Track quarterly commentary from major cloud providers (Microsoft, Google, AWS, Oracle) on token pricing, token volume growth, and backlog. When token prices rise despite falling per-unit costs, it signals demand outstripping supply -- which is a leading indicator for more infrastructure spend, which flows into the companies Lee names. This gives you a data-grounded view rather than sentiment-driven positioning.
3
Understand the Capex-to-Cash-Flow Ratio for Your Sector Exposure
Pull up the aggregate capex and free cash flow figures for the S&P 500 or your specific tech holdings and calculate a simple ratio. Compare to historical peaks. Lee's core argument is that this ratio at dot-com peak was 6x and today is below 1x. If you can verify and track this yourself, you have an objective anchor for assessing whether the AI cycle has reached excess. When this ratio starts approaching 2 or 3x, it warrants reassessment regardless of earnings momentum.
4
Evaluate Palanteer Using Net Dollar Retention as the Primary Lens
Lee's bull case for Palanteer is built almost entirely on net dollar retention of 156%. Before taking a position in any enterprise software company -- not just Palanteer -- pull this metric from the most recent earnings. NDR above 120% at scale is exceptional. Above 140% is rare. 156% at a company doing billions in revenue with no direct competitor is what Lee calls "one of one." Use NDR as your primary quality filter before looking at any other valuation metric.
5
Stress-Test Your Precious Metals Allocation Against Event-Driven Selling
Lee's framework distinguishes between structural demand for gold and silver versus temporary suppression caused by liquidity-driven selling from petrol states or paper market dislocations. Before increasing or decreasing metals positions, identify whether the current price action is structural or event-driven. His recommendation is a 5 to 10% allocation to physical gold as a portfolio insurance position, not as a primary growth vehicle, and he explicitly reduced silver weighting in his personal model due to the higher idiosyncratic risk relative to gold at this stage.
6
Understand Debt-Financed Infrastructure Before Dismissing It
Coreweave's controversy is its debt load. Lee's counter is that when a data center has 50 to 60% gross cap rates and the cost of that debt is in the mid-single digits, the math strongly favors borrowing as much as the market will allow. If you encounter a company using debt to finance infrastructure in a constrained-supply environment with strong pricing power, do not dismiss it on leverage alone. Model the net operating income and the debt service coverage ratio before reaching a conclusion. Lee is explicit that Coreweave's numbers "really, really work."
7
Recognize the K-Shape in Your Own Income and Spending
Lee's K-shaped economy observation is not just a portfolio observation -- it is a personal financial diagnostic. If your income is primarily wage-based and you carry high fuel or transportation costs, you are likely on the lower arm of the K. If your assets are heavily weighted toward equities and you hold AI-adjacent positions, you are on the upper arm. Understanding which arm you occupy changes how you should think about inflation exposure, emergency reserves, and the difference between portfolio performance and actual purchasing power.
8
Follow Adam Parker at Trivariate Research for Earnings Analysis
Lee specifically names Adam Parker -- former chief strategist at Morgan Stanley, now running Trivariate Research -- as a source he trusts for AI's impact on S&P 500 earnings. Parker's estimate that AI infrastructure touches 263 companies in the S&P 500 is the kind of quantified framing that goes beyond sector labels. If you want to understand how wide the AI earnings wave actually is across the index, Trivariate Research is worth exploring as a professional research source.
Full Transcript
This transcript was auto-generated and may contain errors in speaker attribution, transcription accuracy, or formatting. Long transcripts may be truncated due to processing limits. Confirm accuracy and completeness against the original source before referencing or republishing.
Opening segment -- host teaser and news context read by Michelle Makori.
[00:00]
Michelle Makori
We are nowhere near bubble territory. There is so much more to go. We are in one of the largest capex cycles probably in the history of our country. How we use computers and how we use technology is fundamentally changing in perpetuity and Wall Street does not have their arms around it. This stuff in Iran is a total mess, but it's a sideshow to what's going on in artificial intelligence. You were right on the money ahead of silver plunging. It was around $120 earlier this year and you said we would see an over 35% pullback. You were correct. People didn't like it, but you were correct. Mike, what are you forecasting for silver now?
Mike Lee
It's on like Donkey Kong.
News segment -- Michelle reads economic data headlines. S&P 500, NASDAQ, and Dow closed at record highs on the same day PCE rose to 3.8% year-over-year, Q1 GDP was revised to 1.6%, jobless claims reached 215,000, and the personal savings rate fell to 2.6%.
Michelle Makori
Mike, welcome back to the real story. Good to see you again. Before we get into the broader market and the broader macro picture, I do want to start off with silver because the last time that you were on the show, you made a call that generated a lot of reaction from our audience -- not necessarily positive reaction -- but I do give credit where credit is due. At the time silver was around $119 in late January and you said that we would see quite a meaningful pullback around 30%. Sentiment at the time was extremely bullish. We had many analysts talking about $150 silver, $200 silver, even higher perhaps. But you took a much more cautious view, saying that when something moves in a parabolic way, the way silver had, that it generally reverts to where it should be on a fundamental basis. And then by February 2nd, silver had fallen roughly 29%. It's down 35% since those January highs. So let's play the clip and remind the viewers what you said in late January.
Clip replay -- Mike Lee from prior appearance in late January 2026.
Mike Lee (clip)
For the price of silver you could get another 20 to 50 bucks on it but kind of when something moves parabolic like this it generally reverts to where it should be on a fundamental basis and I don't know -- maybe the case for that is $75 bucks.
Michelle Makori
You elaborated on that idea, but basically you called that pullback and you were correct because we did see silver drop around 35% in just a few days after that clip. So what's your outlook on silver now?
Mike Lee
Well, first, I'm wearing the exact same shirt, so I'm going to be as exactly as accurate. Anyway, just as a recap -- when something like, so you see Dell stock today, Dell stock was up like over $100 because of a fundamental earnings change in the company that's far beyond expectations. When the price of silver was moving much higher that quickly, it was because of a supply demand imbalance between the physical and the paper. And so we have now structurally reset in silver into this $75 range and I think we continue to grind higher. I've been buying physical silver since it was 50 bucks to add -- I don't think I bought any over $100 -- buying coins for my kids to put away in their collections with their gold coins. I think silver continues to grind higher. And if you look at where gold has come during this war, how it's pulled back and silver's hung in there, I think it's a very bullish statement. I think that both the gold and the silver -- I don't know that we're going to get in these parabolic moves higher -- but I think it continues to grind higher over the next year.
Michelle Makori
Do you have a price outlook you're comfortable sharing considering you were very accurate last time wearing the same shirt and all?
Mike Lee
Yeah. I'd imagine silver grinds up to around $100 bucks. So if we're looking at this a year from now, probably at $100 here. And then gold -- we get back into that mid-$5,000 range. I would like to see it higher. You would have to think that the bull case for these assets, given the fiscal situation in the US and not only worldwide, that you get there -- and ideally a closing of the Straits of Hormuz... a lot of these oil nations like really load back up on the cash and start resuming their purchases of these assets.
Michelle Makori
So just focusing on silver -- you were saying that the main issue was that it needed to have a fundamental reset. Are you comfortable saying that that reset has now happened? Do you still think that there's significant downside risk ahead for silver?
Mike Lee
No, not at all. Silver was in this $20 to $30 range for years, right? You pull on a rubber band. I think this $75 to $80 range -- maybe we set closer to $100 -- kind of on a go-forward basis. The $75 call from where we were was based on the fact that that's when the chart went straight up. Now commodities don't move in a parabolic fashion unless something happens. Silver doesn't earn anything like gold doesn't earn anything -- it has a value relative to other assets. And I think that value increases over time, but it's just a different scenario than, say, a stock.
[05:00]
Michelle Makori
So $100 silver by end of 2026? What's your timeline?
Mike Lee
Hopefully a year from now we're at our new home at $100 bucks. Because again, there's a lot of good things happening in the US, but unfortunately the fiscal situation is not one of them.
Channel promotion segment -- Michelle briefly thanks viewers and promotes Miles Franklin at info@mfranklin.com and milesfranklin.com.
Michelle Makori
Focusing a little bit on silver before we get onto your broader AI outlook -- does the industrial play that could come from all of these AI infrastructure stocks, electrification, satellites with SpaceX, do you see an industrial angle for silver perhaps as a bigger trigger for a more significant move?
Mike Lee
Not especially. Given that the commodity price has doubled, I think it has more to do with somewhat structurally higher inflation and the fact that we're unlikely to get any rate hikes. This inflation we're seeing now is a price shock from oil from the conflict in Iran. Once that resolves itself, inflation -- I do think the next couple prints are probably worse than this month -- but then you kind of see a pretty big mean reversion. You're also seeing any effects from tariffs, which I think are far less than what the tariff naysayers prescribe. You kind of start to see all these things resolve themselves. Inflation comes down. But you've got deficit problems. The Democrats are going to win the house. Republicans aren't going to hold the House unless there are insurmountable changes. So the spending is not going to change. It's not going to get better.
[10:00]
Michelle Makori
Last time you were on you said that you had never been more bullish about equity markets. What's your outlook now on the equity markets and the broader economy at large?
Mike Lee
The infrastructure needed to build out for artificial intelligence is the most widely misunderstood concept on Wall Street I have ever seen. The amount of data that goes into these large language models is growing on a hockey-stick exponential growth. The models themselves are getting more complex on another hockey stick. And then when you go from LLMs and frontier models to agentic AI it's another hockey stick. So a token is four units of code on an LLM. Token prices doubled this year even though they had fallen by a factor of 60 from January 2023 until the beginning of this year -- because the token demand is far outstripping that of the infrastructure buildout. There's an estimate for a trillion dollars of AI infrastructure spend next year. It's too low. It will most likely be more and that's expected to grow to three trillion on an annual basis in 2030. AI touches 263 companies in the S&P 500 according to Adam Parker, who was chief strategist at Morgan Stanley when I was there and now runs Trivariate Research -- I think he's brilliant. It's on like Donkey Kong. If people can look at trailing multiples -- which are as useless as you know what on a nun -- you don't drive down the highway looking through your rearview mirror the whole time. What are the forward multiples? Nvidia is trading at like 14 or 15 times forward earnings. The whole market trades at 22. So the stock could go up 50% without earning a dollar more, just from getting a market multiple. The lack of appreciation for what's going on with these companies is staggering. Broadcom, AMD, Marvell -- they're trading at roughly the same valuation they were six months ago even though the stocks have gone essentially parabolic, because their future earnings expectations have gone through the roof. The stuff in Iran is a total mess. But it's a sideshow to what's going on in artificial intelligence.
Michelle Makori
You're saying the broader equity market is just going to be carried up by this AI uptrend?
Mike Lee
Yes. The history of analyst estimates is that you enter six months out from a quarter and estimates slowly trickle down so that by the time the quarter shows up you're expecting 10% growth when you started at 15%. This quarter we were expecting 13% year-over-year growth and we've come in at 28. I've never seen anything like that. And estimates are going up for later this year and then they'll be higher the year after. It all has to do with the underestimation of the demand for AI. We are in one of the largest capex cycles probably in the history of our country and we're nowhere near the end. On the S&P 500 as a whole, the capex is less than one times cash flow. In the dot-com boom it was at six. We are nowhere near bubble territory. There is so much more to go. The way that we use computers and how we use technology is fundamentally changing in perpetuity. It's happening in front of us and Wall Street does not have their arms around it.
[15:00]
Michelle Makori
Last time you were very bullish and correctly so on AI -- you had Palanteer, Nvidia as some of your picks. They've performed well. Are there any particular stocks in the AI sector that you're extremely bullish on now? Is the opportunity in photonics? In memory like we saw SanDisk have a phenomenal run as well this year?
Mike Lee
If you look at Micron, SanDisk, Western Digital -- they haven't sold additional units. Their sales growth comes almost entirely from a supply demand imbalance -- pricing of memory has gone parabolic. Whereas Nvidia -- off a $40 billion base, they did $80 billion this quarter. That's 100% year-over-year growth when you take out China, for one of the most cutting-edge leading companies in one of the biggest revolutions ever. So Nvidia I would say right now is an absolute steal. Palanteer -- for those of you that don't quite understand what Palanteer does, it's the operating system. They tie all the software of an enterprise together plus their structured and unstructured data, feed it into the large language model, and when it comes out, feed it into the business -- all in a secure environment so your data isn't comingled with someone else on the cloud. The Department of Defense had seven different tech silos between the army, navy, air force, space force, coast guard, JCO, and CENTCOM. They all had their own tech stacks and now they're on one operating system called Maven -- that's Palanteer's Maven system, soon to be Lynchpin. And with that you see things like where we can go into a country and extract a president in the middle of the night or hit 13,000 targets in Iran in 38 days. There's never been anything like it. And if they can do it on the scale they've done it in life-and-death circumstances where the stakes couldn't possibly be bigger, they can certainly do it for an enterprise. On a net dollar retained basis -- essentially the same-store sales of software -- they were at 156% this quarter. There's never been a SaaS company with that high a number at this much revenue. They are one of one. And the stock has gotten beat up because people like Michael Burry who have something against this company on a personal basis are writing short theses based on Reddit and Glassdoor and salty ex-employees -- which is creating an unbelievable opportunity to enter.
Mike Lee
The next most misunderstood company is Coreweave. Coreweave is a data center play. If you can stand up a data center and rent out all sorts of Nvidia GPUs simultaneously to the largest players -- or by the hour to the smallest -- you're essentially selling water in the desert. The gross cap rate for a data center is somewhere between 50 and 60%, with net operating income closer to 25%. So if you set up a billion dollar data center, you have between $500 and $600 million of rent coming in annually. With those numbers you can borrow as much as anybody will give you at any rate below 20%. Coreweave has gotten their cost of capital down from the mid-teens to the mid-single digits over the last year. Because they're using debt rather than cash flow or equity to finance growth, the street doesn't like them. But their competitor Nebius in the neocloud space has gone absolutely parabolic and Nebius is nowhere near the size of Coreweave. They have nowhere near the backlog. Coreweave is going to 5x their revenue from Q4 last year to Q4 next year. That stock is one that -- it's around $100 to $110 -- it's controversial because of their debt load, but the numbers really work. And those Nvidia GPUs at the end of three years are described as worthless when in reality the companies renting ones that are five, six, and seven years old are getting increasing rental rates because of the scarcity for token availability that's going to exist for the next half decade.
[20:00]
Michelle Makori
Photonics -- for viewers that don't know, photonics uses light rather than electricity to move data between processors and servers. Many experts think this could become critical as AI workloads grow and copper-based systems hit physical limits. Is this the next structural play with AI?
Mike Lee
Look, this is not a rabbit hole I've gone down deeply enough to give you a quantifiably good answer. It could be. The theme here for AI infrastructure is that at every single turn there is a logjam. There's not enough land to build the data centers. There's not enough nuts and bolts and material to stand them up. There's not enough copper. There's certainly not enough energy and there aren't enough GPUs. And so what that is going to do is preserve pricing power across all of these constraints. I would look into those photonics names, but it's kind of like quantum in that a lot of these names barely do any revenue. And if you look at a name like Oklo -- which makes small nuclear reactors -- they just got approval to use spent fuel to make a 1.2 gigawatt power structure for Meta's data center buildout in Ohio. Zero revenue, $12 billion market cap. I would tell investors that there are companies making money hand over fist where you can see many multiples on your invested capital without going into something experimental. So Nvidia is going to print money and the stock should be double -- if not 250% -- from where it's trading just by getting its historical multiple back. Nvidia's also going to beat earnings and is the leader in the space developing all sorts of other businesses to keep you in their ecosystem.
Mike Lee
Another name starting to bounce back but way off its highs is Oracle. Oracle works because they control 35% of the world's data -- they are the world's largest database manager. And so what does ChatGPT, Claude, Grok run on? It runs on data. And if you're an enterprise, you want it to run off your data. So you need to go through Oracle in all likelihood to get your data through the LLM to make it work for your enterprise. And Larry Ellison isn't going to let you have that for free. They're taking on a massive amount of debt and again it spooks investors because two or three years ago Meta, Google, Microsoft, Oracle were all cash flow machines. Now they're spending in excess of their cash flow on capex, plus taking on a ton of debt. It requires a leap of faith. But to me, the underlying fundamentals are that the revenue and earnings upside is far greater than the street gives them credit for. You saw that today with Dell.
[25:00]
Michelle Makori
Let's bring it back to the macro picture. You're calling the Iran conflict a sideshow. Markets have become a little desensitized to it now. I mean, the latest is that there could be a deal this weekend, but we've heard that song before. We still have higher energy costs feeding into the broader inflation picture, supply chain disruptions, shipping costs, commodity inflation beyond oil, a potential impact on interest rates and monetary policy. You know, this puts the Fed in an even harder corner than it usually is. How can you call this a sideshow? Many people are saying that even if there is some kind of resolution now, the effects have been baked in and the slowdown on the global economy is going to hit because of the backlog.
Mike Lee
Those are all things I basically agree with -- this is a net negative. But for equity investors, particularly US equity investors watching their stock portfolio, if they're in this AI infrastructure trade, it is a sideshow. When you start looking at what's going on globally -- and for people that own physical gold and wondering why it's down -- I think as this war happened and started in all these petrol states, they couldn't you know -- you're losing a tremendous amount of daily cash flow and you have to fund your operations as a nation. And to fund your operations you sell what you can, not what you want. Gold at or near all-time highs -- you sold a ton. US Treasuries -- things that are very liquid. So it put pressure on all of these assets. It's becoming clear that the US is driving the world economy. And now we're spending all this money domestically and building out the US. These price shocks from oil have nowhere near the effect on us that they would have say 10 or 15 years ago.
[30:00]
Michelle Makori
I still have this macro data that is just not painting a good picture. PCE rising to 3.8% year-over-year in April. First quarter GDP revised downward to 1.6%. Jobless claims climbing to 215,000. Personal savings rates falling to 2.6%, the lowest since June 2020. My concern is that there's a disconnect and at some point that disconnect catches up.
Mike Lee
Yeah. You're making the argument for a K-shaped economy. If you own assets and you're particularly involved in these AI trades, you have not a care in the world. If you drive a lot and this is harmful... the cost is real. This PCE number -- we were only at 0.2% on a core month-over-month basis -- but those numbers are likely to get worse the next couple months even if we get a deal this weekend because there's a lag effect. That's bad. But from an investor perspective in the S&P 500 -- of which 30-something percent is tech and another 28% is communications, which is basically Google and Facebook -- you have a different picture entirely.
Michelle Makori
So what do you see as the biggest market risk? What could derail your thesis here?
Mike Lee
It's an exogenous event. Something we haven't thought of. Something we don't see. Because you go into this earnings quarter expecting 13%, okay, maybe you get 15 or 16% -- that's awesome. We came in at 28. I've never seen anything like that and it's all driven by this AI trade. You might even say the K-shaped economy is like AI and everything else. But it's so big and it's happening truly once in a lifetime, once in a generation, in terms of the buildout going on in front of us. And it's happening primarily domestically in the US because the Europeans don't want to embrace it. This is a US phenomenon with US beneficiaries. So if that provides 3 or 4% of GDP growth, we may only grow at 2% because everything else is contracting. But if you're an investor in the S&P 500 or the Q's, you're going to be doing great -- not good, great. Whereas those not part of this investor class are going to be living very differently than those watching their portfolios hit new highs on a daily and weekly basis.
Michelle Makori
If you had $100,000 today, how are you dividing that? How much is going into AI stocks? How much is going into physical gold, if any?
Mike Lee
So I'd buy some Palanteer, some Coreweave, some Nvidia, some Oracle. I'd definitely buy some physical gold. My portfolio has 8.5% of my equity model portfolio for my clients in BAR, which is a gold ETF. But if I'm putting a dollar to work today, between 5 and 10% is going to gold. Less so silver because I think gold is the loser from an idiosyncratic event -- all these petrol states and nations around the world needed liquidity and they sold gold. That's an artificial suppression, the same way that six months ago there was an artificial bump due to the paper and physical supply demand imbalance. Now gold's going to recover. Once it recovers to new highs, how much momentum gets behind it is the real question. So yes, I would be a gold investor here because the market could be fickle. If this war drags out another six months and capital markets tighten up a little bit and some IPOs don't go well, that could weigh on stocks, where gold is going to recover in my opinion. And then you're always adding physical gold on a periodic basis as far as I'm concerned.
Michelle Makori
Final question. Point of highest conviction as we wrap up.
[35:00]
Mike Lee
Palanteer. It's one of one. There's no other company like it in the world. They're going to do $8 billion in revenue this year. That number is likely to be north of $60 billion in 2030 and it's going to continue to grow. It makes AI actually work for these large enterprises and there's literally only one company that can do it.
Michelle Makori
Mike, for people that want to learn more about you and your work, where can they find you?
Mike Lee
MichaelStrategy.com. You can Google Michael Lee Strategy. I'm @MikeLeeStrat on X, formerly known as Twitter.
Closing -- Michelle thanks Mike, asks him to wear the same shirt next time. Standard channel outro from Michelle follows with Miles Franklin contact information.
AI Master Prompt
The AI prompt on this page is auto-generated from the transcript content and is intended to support further exploration of the topics, concepts, and conclusions discussed. It is provided for informational purposes only. The user is solely responsible for all outcomes resulting from its use.
Master Prompt -- Mike Lee / Miles Franklin Media -- Silver Reset and AI Supercycle
You are a knowledgeable discussion partner helping me explore the investment framework and market thesis presented by Mike Lee (founder, Michael Lee Strategy) in a June 2026 interview with Michelle Makori on Miles Franklin Media.
The core of Lee's framework rests on two parallel theses operating simultaneously.
The first thesis is on precious metals. Lee called silver's parabolic move from $50 to $119 as unsustainable in late January 2026 and targeted a reversion to roughly $75 on the basis that parabolic moves in non-yielding commodities are driven by supply-demand imbalances in the paper market rather than fundamental re-ratings. That call proved accurate at roughly 35% down. He now expects silver to grind to approximately $100 over the next year without another parabolic event, and gold to recover into the mid-$5,000 range from a temporary suppression caused by petrol states selling liquid assets during the Iran conflict. His precious metals allocation in client portfolios is 8.5% in a gold ETF (BAR). He allocates 5 to 10% of new dollars to gold and less to silver, given gold's stronger risk-adjusted case at this stage.
The second thesis is on AI infrastructure equities. Lee argues that AI token demand is growing on three simultaneous exponential curves: the volume of data in large language models, model complexity, and the shift from frontier LLMs to agentic AI. Token prices doubled in 2026 even after falling 60x from early 2023 to early 2026, signaling that infrastructure supply cannot keep up with demand. He estimates AI infrastructure spend will exceed $1 trillion in 2027 and grow to $3 trillion annually by 2030. His primary bubble-detection metric is the capex-to-cash-flow ratio for the S&P 500, which sits below 1x today versus 6x at the dot-com peak. His highest-conviction names are Palanteer (net dollar retention of 156%, "one of one" enterprise AI OS, $8B revenue growing to $60B+ by 2030), Nvidia (100% YOY revenue growth on a $40B base, trading at 14-15x forward versus market's 22x), Coreweave (data center play using debt financing with 50-60% gross cap rates and a 5x revenue growth trajectory), and Oracle (controls 35% of global enterprise data, a structural gatekeeper for enterprise AI). He acknowledges a K-shaped economy where AI investors thrive and the broader population absorbs stagflation from the Iran conflict and chronic fiscal deficits.
Key principles from this framework:
- Parabolic moves in non-yielding commodities revert to fundamental levels; identify where the chart "went straight up" to find the reversion target
- Token demand is the leading indicator for AI infrastructure investment need -- track it quarterly from cloud provider earnings commentary
- Forward multiples are the only useful valuation tool in a growth inflection; trailing PE ratios are actively misleading
- Capex-to-cash-flow ratio below 1x on the S&P 500 as a whole means bubble conditions are not present by historical standards
- Net dollar retention above 140% at revenue scale is a rare quality signal in enterprise software; use it as a primary filter
- Debt-financed infrastructure is not inherently dangerous when gross operating yields far exceed cost of debt; model net operating income before reacting to leverage headlines
- Physical gold is portfolio insurance against exogenous risk and fiscal deterioration, sized at 5 to 10% of a dollar invested today
What this framework is not: It is not a call to abandon caution entirely. Lee explicitly names exogenous risk -- an unanticipated event -- as the primary thing that could derail his thesis, because the internal earnings trajectory is intact. It is not a blanket endorsement of all AI-adjacent names; he draws a clear line between proven cash-generating businesses (Nvidia, Palanteer, Coreweave, Oracle) and experimental zero-revenue plays in quantum, photonics, and small nuclear. It is also not a silver bull call for aggressive positioning; he reduced silver weighting relative to gold specifically because gold has a cleaner recovery narrative at this stage.
How to use this chat:
1. CLARIFY -- If you want to understand any specific concept Lee used (token demand curves, net dollar retention, capex-to-cash-flow ratio, physical vs. paper silver markets), ask me to explain it in plain terms with a concrete example.
2. APPLY -- If you have a specific portfolio or position you want to run through this framework (e.g., "I hold XYZ, does it fit Lee's criteria?"), share it and I will work through the relevant metrics with you.
3. STRESS-TEST -- If you want to pressure-test Lee's thesis, name a specific claim and I will steelman the counterargument as rigorously as possible so you can evaluate both sides.
4. COMPARE -- If you want to compare Lee's framework against another analyst's view (e.g., a silver bull's case for higher prices or a bear's case on Nvidia valuation), I can lay out the competing frameworks side by side.
5. TRACK -- If you want to know what data points to monitor to validate or invalidate Lee's thesis over the coming months, I can build you a short checklist of specific metrics and timelines to watch.
6. ALLOCATE -- If you want to think through how a specific dollar amount might be divided using Lee's stated preferences (AI equities plus 5 to 10% physical gold), I can walk through the logic with you without giving you personal financial advice.
Respond in a grounded, direct tone -- no hype, no preaching. Help me reason through this framework with the same matter-of-fact clarity Lee himself uses. When you are uncertain, say so.
To start: What is your current level of familiarity with AI infrastructure investing -- are you just beginning to explore it, do you already hold positions in this space, or are you looking at it from a more skeptical or analytical distance?