SEMICON Taiwan is one of the world’s largest semiconductor industry trade shows and conferences, held annually in Taipei.

This year I’m attending to share what I discover here.

This is only Day 1.

Read more

A Future $100B Company or a Disappointment?

A shorter article, not really my style, but why not?

IREN really does bring a lot of attention.

That’s the curse of becoming a retail favorite: as many people will hate you as will love you.

Anyhow, there are valid arguments worth responding to. These are some I found.

Dilution

“IREN is a dilution machine. Way more than Nebius.”

IREN has used more ATM issuance, but Nebius has used more convertibles.

Assuming all convertibles convert, Nebius has increased its fully diluted share count by 45% over the last 14 months, while IREN is at 61%. Over the last six months, it’s 22% and 21%, respectively.

Let’s not exaggerate here.

Financing

“They are raising debt. They’ll probably have to dilute way more to fund $25–30B of capex.”

The fact that IREN can get 50% of GPU value in customer prepayments and another 90% in GPU financing on top of that, while keeping its data centers unencumbered and therefore preserving another pool of collateral against which it can raise financing, is very impressive.

Yet people focus on the absolute $25–30B figure and ignore the outstanding work IREN is doing to raise tens of billions of dollars while being a $15B company.

IREN is performing just as well as Nebius in the financing department while being the smaller company.

GSW sponsorship

“$50M per year GSW sponsorship.”

I agree. It makes no sense to anybody.

At least it’s almost humorous that they got to put such an ugly logo on their shirt.

“Too slow”

“Slow as fuck compared to Nebius and CoreWeave.”

IREN is about a year behind CoreWeave and Nebius in AI cloud monetization, but saying that alone is cherry-picking.

IREN came later, but it’s coming stronger.

• IREN’s operating ARR as of August 26 was $1B. Nebius exited June at $3B.

For a company that was AI-native and had a much stronger team working on this for a decade before IREN, does that really sound like some crazy gap? Especially considering IREN is now growing faster?

IREN went from a $280M annualized revenue run rate based on its June-quarter average to $1B of operating ARR by August 26.

Once you start delivering contracts, ARR explodes. People are turning a timing issue into a fundamental issue when it isn’t.

CoreWeave and Nebius were earlier. That doesn’t mean IREN can’t catch up.

• Moreover, IREN is projecting 300% growth from its current $1B operating ARR to $4B by year-end, already contracted.

• Nebius is targeting $8B of year-end ARR at the midpoint, up about 167% from its $3B June exit ARR.

Again, IREN is growing faster in relative terms.

For 2027, IREN is targeting 0.8 GW of cumulative IT delivery, translating to 0.5GW of added delivered capacity. Nebius has said it wants to add more than 1 GW per year starting in 2027. 1/4 of the market cap, 1/2 of the capacity growth.

That doesn’t seem like a crazy difference to me for a company that started this journey years before IREN.

So is IREN really too slow?

Or is it actually one of the most impressive pivot and acceleration stories in the data-center space?

Vertical integration

And that’s before we even get into power capacity.

It’s easier to grow quickly when you lease data centers, but then you can run into problems like Nebius did with DataOne or CoreWeave did with Core Scientific.

Nebius has recognized this problem and is increasingly focusing on wholly owned sites, while CoreWeave is sticking with its existing model.

IREN has pursued vertical integration from day one.

It has the most impressive power portfolio of the three at a fraction of the market cap.

Monetization

Now let’s talk about monetization.

IREN’s base monetization is basically the same as the others.

IREN had around 90 MW operating as of August 26. Based on $1B of operating ARR, that implies:

• $11.1M ARR/MW.

Nebius reported $12M ACV/MW for its 2026 base economics.

Then you move further out on the curve.

• IREN said its newer three-year contracts are now above $20M/MW, with active discussions around $25M/MW.

• Nebius said its Q2 new deals were above $20M/MW, while some Q3 short-duration deals exceeded $40M/MW.

Nebius is capturing higher rates at the right end of the curve, which makes sense considering the depth of its platform.

But on the base deals, the picture is practically the same.

The bigger picture

Like I said, IREN is behind Nebius in monetization and definitely behind in its software stack.

But in terms of speed? IREN takes the crown.

Look at the moves this company is making:

Securing enormous amounts of power and land. Signing contracts. Bringing in high-tier customers like Figure AI, Perplexity, Microsoft, NVIDIA, Fluidstack, Cohere and Higgsfield. Building its software stack from zero. Building a brand. Expanding globally.

It’s genuinely a feat.

IREN is spending a lot of money. It’s rapidly expanding headcount, making acquisitions, developing sites around the world and spending on sponsorships and branding.

But they’re doing this with the ambition of becoming a $100B company, and you don’t get there by operating like a small, conservative infrastructure company.

While I agree IREN has overspent in some areas, particularly the GSW sponsorship and the massive executive compensation package, I also believe they’re taking the steps required to become a data-center giant.

Pieces falling into place

IREN has sold out its 2026 capacity.

It has already pre-sold around one-quarter of the capacity it plans to bring online in H1 2027, and its growth is staggering.

Demand isn’t the problem.

Power and land aren’t the problem.

It now has a close relationship with NVIDIA that should help with GPU access, while its financing options have improved dramatically.

The pieces are falling into place.

But investors tend to become overly emotional in both directions with popular names, which is why, on X, IREN can only be the next vertically integrated neocloud giant or a house of cards about to collapse.

Nothing in between.

What am I doing?

Read more

In today’s article, I’ll present what I believe is the safest type of investment you can make if you want AI exposure: energy.

Interestingly, it’s also one of the investments that, despite being extremely strong fundamentally, tends to get punished hard by the market at any opportunity. That creates plenty of opportunities to buy.

In fact, right now, I believe there’s a great entry point for some of these companies.

However, not all of them offer the same rock-solid business. Being selective is necessary to separate the wheat from the chaff and get exposure in a risk-aware way that truly makes this one of the safest bets in the AI buildout.

Read more

High-Precision Manufacturing

The Hidden Industry Behind Space, AI, Robotics, and Advanced Hardware

In this article, I’ll touch on a theme you don’t usually hear about: high-precision manufacturing.

Behind every impressive technology humanity has ever invented, there is a machine that allows it to happen.

However, these companies tend to remain under the radar, despite their impressive customer portfolios.

And it’s not because they can’t deliver good returns. Some of these companies have vastly outperformed the SPY over the last few years.

It’s simply a more under-the-radar industry, but that could change through the growth of two key industries that require extremely precise components: aerospace and robotics.

• We live in a new space race. This time, it’s not just about reaching space, but about scaling the space economy. Satellite constellations, defense systems, a Moon base, and orbital data centers could turn space into an eventual trillion-dollar market, and every component sent to space needs one of these machines to make its production possible.

• Secondly, robotics. Some of these companies are already positioning themselves not just in robotic components, but also in the AI infrastructure that powers them. Semiconductor machinery, complex AI accelerator components, humanoid components... all of this increases demand for these machines.

This growth is already palpable, and it’s not mundane. Orders are increasing 30-60% YoY across the industry, all while everything indicates this is only the start of a worldwide manufacturing boom.

Manufacturing capacity for these machines is constrained, the industry is investing in expansion while machine pricing keeps increasing, and companies are forecasting years ahead of growth and expanding margins.

In this article I’ll cover the companies that supply SpaceX, Blue Origin, Airbus, Rocket Lab, NASA, Tesla, Anduril, Lockheed Martin, Northrop Grumman, GE Vernova, Siemens Energy, ASML, Firefly, Boeing, and Hadrian, among many others.

I will go into the industry, the most interesting stocks, and offer insights into names you genuinely may have never heard about before.

Let’s start.

Read more

Portfolio up 90% YTD across just 13 positions in total.

In this article, I’ll cover my only Physical AI bet in the portfolio so far.

Read more

Subscribe now

In this article, I’ll cover an update on a company I’m accumulating.

This company is a semicap equipment supplier benefiting from the advanced-packaging tailwinds tied to the chips powering the AI era, including HBM.

Jensen sees the semiconductor industry needing to 10x production to power the agentic AI boom.

That means, as an investor, you want exposure to this massive expansion in semiconductor manufacturing capacity.

This company holds a quasi-monopoly in a key process within the supply chain for high-end AI chips.

Its dominant position, combined with the tailwinds from the massive expected growth of the semiconductor industry, makes it a solid bet at a $1.2 billion market cap.

Read more

In this article, I’ll build a valuation model for SpaceX to assess what fair value could look like, so we can get a picture of what a rational, down-to-earth buy zone might be.

I’ll be using a sum-of-the-parts valuation, which, in my opinion, is the go-to approach for such a heterogeneous company.

Read more

In this article, I’ll explain what’s happening in the markets and what I’m doing.

Let’s go.

Macro Analysis

We’ve had a few choppy weeks in the stock market, especially in semiconductors, which has greatly affected the more leveraged Korean investors.

There’s now a clear performance gap.

While the S&P 500 has just reached new all-time highs, the KOSPI and semiconductor stocks remain 25% below their yearly highs.

Why the Rally Became Excessive

From March through June, we went through the biggest semiconductor rally in history. It seemed as if the market had suddenly started believing in AI more than ever before, almost inexplicably.

I’d say it had a lot to do with the war in Iran.

While we were getting bearish news from the Middle East in the form of higher oil prices, supply-chain disruptions, and broader war fears, the fundamentals of the AI trade were improving in the background. However, they were being suppressed by the bearish market climate.

Once that weight was lifted from the market’s shoulders, all those improved fundamentals were suddenly reflected in stock prices, like a compressed spring being released.

Excesses Started to Build

However, with excitement often comes excess.

Investors started piling into semiconductors as if they were free money. I mentioned this several times throughout my articles and became increasingly selective with my picks.

When I looked at themes such as ABF substrates, MLCCs, optics, and glass cloth, among others, valuations were getting out of hand, pricing in several years of earnings growth.

As if that weren’t enough, the leverage cycle poured gasoline on the fire.

Korean Leverage Reached Record Levels

• Korean domestic margin loans reached ₩38.63T on June 24, the highest level ever recorded.

• Borrowing specifically against KOSPI stocks reached ₩29.8T, up 71% from ₩17T at the end of 2025.

• Margin debt in Samsung and SK Hynix alone reached ₩7.79T, up 208% from year-end. The two stocks represented 28% of all KOSPI margin debt.

The Problem Wasn’t Limited to Korea

However, this wasn’t just a Korean problem.

While Korean investors abused leverage the most, the same trend was also visible in the United States.

• U.S. margin debt reached $1.502T at the end of June, up 6.1% in one month and 49% in one year.

• Assets held in U.S. leveraged ETFs reached $220B, up 60% from the end of March, with exposure heavily concentrated in technology and semiconductors.

With this level of leverage, any correction can have devastating effects on stock prices.

What Caused the Leverage Unwind?

The decisive catalyst was the June 22 regulatory warning.

On June 22, the Korean regulator acknowledged that authorities may have moved too quickly in approving leveraged single-stock ETFs and warned about market overheating.

With the risk of leverage restrictions increasing, institutions began selling.

Once Samsung and SK Hynix fell approximately 12%, several mechanical sellers entered the market, creating the domino effect we saw.

Oversupply and Financing Fears Added Pressure

After this, on July 1, Meta was reportedly planning to sell excess capacity as a neocloud, which introduced overbuild concerns into a market that had already seen more aggressive bond issuance to finance AI needs.

On July 7, Goldman data showed that hedge funds had sold U.S. technology hardware for four consecutive weeks amid this combination of historic leverage, oversupply fears, and what some investors saw as excessive debt being used to fund AI capex.

What followed was a historic leverage cleanup.

• July became the worst month in the KOSPI’s 46-year history.

• Two million Koreans were margin-called, and more than 500,000 were liquidated.

Where Are We Today?

We can look at what some institutions are saying to get a better idea of where we stand today and how the current situation could affect the next phase of the cycle.

Read more

Performance Update

Despite the drawdown, my portfolio is up 65% YTD, and my six-year return stands at +869%.

This was achieved with no options, no margin, and no high-risk nano-cap investments.

In this article, I’ll compile all my investment theses, so any new investor can instantly understand what I’m invested in, why I own it, and what I’m buying during dips like the one we’re currently experiencing.

I’ll also share the updated portfolio spreadsheet, including all my holdings and price targets.

Read more

Is AI a Bubble?

I just hopped off a call with an analyst at a major firm, and we both came to the same conclusion.

Market sentiment at the moment is composed of many little moving pieces, but they all come together to form the same question:

Is AI a bubble?

In this article, I’ll explore what needs to happen for stocks to go up and, beyond all the noise, what the market is looking at on a fundamental, long-term basis.

Read more

Subscribe now

In this article, I’ll cover a company I’m accumulating that has just presented an industry-shaking update the market has not yet fully digested.

I’ll explain my rationale and present a valuation model to derive my price target.

Let’s get into it.

Read more

Performance Overview

• 2026 YTD: My portfolio is up 89.1% YTD, compared with 9.4% for the SPY.

• My six-year total return now stands at 1,011%, while the SPY is up 154% over the same period.

That means my starting capital has grown by roughly 11x, compared with about 2.5x for the market.

This was achieved without options or leverage, simply by investing in misunderstood companies exposed to the highest-growth themes in the market.

Every position is public. I publish my full portfolio, along with my rationale and price targets, and provide regular updates in my portfolio reviews. The figures are unaudited, but the record is open, and my track record is documented.

Subscribe now

Over the past few weeks, AI-related stocks have experienced a sharp drawdown, creating discounts of more than 30% across some positions I own. As I explained, I believe the sell-off was unfounded, and the recovery I expected is already underway.

Update Structure

This article will follow a different structure from previous updates.

1. First, I’ll explain all the news and recent developments.

2. Secondly, I’ll give my take on each company, providing a more direct and less time-consuming way to understand the key messages and their implications.

3. Then, I’ll provide a snapshot of all my price targets.

4. Finally, I’ll briefly explain what I’m doing with each of my positions.

Read more

Subscribe now

Performance is holding up surprisingly well.

It’s remarkable how much a couple of bad weeks can distort your perception of just how good the year has been.

In this article, I’ll answer some of the more complex questions submitted by my paid subscribers.

Let’s get into it.

Read more

Subscribe now

Introduction

In this article, I’ll examine a micro-cap semiconductor equipment company selling into the PCB, panel-level packaging, and glass-substrate supply chains, with companies such as TSMC, Samsung, and Intel as end customers.

This company is exposed to both glass substrates and panel-level packaging, the two revolutions coming to semiconductor manufacturing that every investor should be aware of.

I’ve been mapping these supply chains across several articles, as I believe they will produce multiple multibaggers as the technologies move from trials, where they are today, to mass-scale production over the coming years.

Read more

Time for an update on memory, optics, and what institutions are saying.

In this article, I’ll also cover exactly where I’m adding and the stocks currently on my watchlist.

Read more

Subscribe now

I hold one position with exposure to drones, laser-based C-UAS systems, robotics and cheap counter-drone missiles.

However, this $10B market cap company also has a less flashy business segment that could prove to be much more significant than the low value the market is assigning to it today.

Concretely, this company has space-based orbital laser communication technology, which so far is being used for defense satellite communications, as well as in missions like the most recent Artemis II.

In 2026, safe and fast defense communication is a much larger business than the nascent orbital data center industry. However, on a 10-year timeframe, this company may have a rough diamond in the form of lasers.

In this article, I’ll explain the business, the great potential I see, and why I’m buying more.

Read more

Subscribe now

In this article, I’ll explain the current situation in the AI trade and what I’m doing.

First, we need to understand what is happening.

The Sell-Off

On Wednesday, the S&P 500 fell 0.2% and the Nasdaq dropped 0.7%.

The semiconductor index fell 6.3% on the first trading day after the sector’s best quarter on record. Micron dropped 11%, wiping out $140 billion of market value in a single session. SanDisk lost 11%, Applied Materials fell 10%, Intel dropped 9%, and AMD fell 7%.

The sell-off was heavily concentrated in semiconductors, which requires some context.

Micron, Intel, and AMD added $2 trillion of combined market value in Q2 alone. The VanEck Semiconductor ETF gained 82% in the first half, its best first-half performance since the fund launched in 2000. Micron was up 260% year to date going into Wednesday.

Bernstein had already published a cautious note on Monday comparing neocloud economics to colocation, flagging GPU depreciation, financing costs, and operating complexity as the key risks to profitability.

The AI buildout trade was priced for zero doubt, then doubt arrived.

Meta: From AI Lab to Neocloud

Meta is reportedly drawing up plans for a cloud business that would sell access to AI computing power and models under an internal initiative called Meta Compute.

Two structures are reportedly under discussion:

• A Bedrock-style service, where developers access models hosted on Meta’s infrastructure, including Muse Spark, the closed-weight model Meta shipped in April.

• A raw compute business, which is effectively the neocloud model.

The plans are still in development and could change. But this comes after Zuckerberg had already said in May that selling excess compute was “definitely on the table.”

The market was so stretched that what moved stocks was not an actual product launch, but a reported intention.

Meta itself closed up 9% on Wednesday. Investors liked the idea of monetizing a capex bill guided as high as $145 billion this year, instead of relying only on LLMs that appear to be going nowhere and are becoming less competitive with Anthropic and OpenAI frontier models.

The companies hit hardest were Meta’s compute suppliers.

CoreWeave fell 14% to $86, while Nebius dropped 17% to $229.

Both count Meta as an anchor customer, with a disclosed $21 billion commitment at CoreWeave and an agreement worth up to $27 billion at Nebius. A Meta that builds enough capacity to sell the excess is a Meta that may eventually rent less capacity from others, while also competing for external workloads.

Beyond the hit to its most direct suppliers, this creates a dangerous narrative:

If Meta has excess compute, maybe the AI buildout is overbuilt.

Once that narrative exists, every long AI hardware trade gets sold at the same time, especially the highest-beta names in data centers, photonics, and memory.

Macro: Not Enough

Thursday’s jobs report, released a day early because of the July 4 holiday, should have been fuel for growth assets.

• Nonfarm payrolls rose by just 57,000 in June, below expectations of 110,000 to 115,000. April and May were revised down by a combined 74,000 jobs.

• The unemployment rate fell to 4.2%, but for the wrong reason: labor force participation dropped 0.3 percentage points to 61.5%, the lowest level since March 2021, while household employment fell by 507,000.

• Leisure and hospitality shed 61,000 jobs, its worst month since 2020.

This was a very soft report.

Consequently, September hike odds fell from the mid-60% range to a coin flip, with the remaining hike bets pushed toward Q4. The two-year yield slipped to 4.12%, the dollar index fell about 0.5%, its biggest one-day drop in two months, and gold moved above $4,100.

With Warsh’s Fed holding rates at 3.50% to 3.75% behind a hawkish June dot plot, the weak print should have been a pressure relief valve for long-duration assets.

Not for the AI trade, though.

The Dow added 595 points, rising 1.1% to a record close of 52,900. The S&P 500 finished flat, even though most of its members advanced.

In contrast, the Nasdaq fell 1%, and the semiconductor ETF dropped another 5%. Teradyne fell 14%, KLA dropped 12%, and Micron lost another 6%. Even Meta gave back 4%.

Across the two sessions, SOXX lost 13%, its sharpest two-day drawdown since March 2025.

Memory: The Most Leveraged Bet

For better or worse, memory is treated by the market as the most cyclical part of the AI trade.

So when the market hears “overcapacity,” memory gets sold first and questions are asked later.

Korea is where that reflex did the most damage.

• The KOSPI closed down 8% at 7,648, falling back below 8,000 only weeks after crossing 9,000 for the first time.

• SK Hynix fell 15% to ₩2,187,000, while Samsung Electronics dropped 9% to ₩286,000. SK Square, Hynix’s largest shareholder, lost 13%. Samsung Electro-Mechanics fell 13%, and the KOSDAQ dropped 7%.

The open was violent enough, down 5.4%, that the exchange halted trading for five minutes.

This was the third Korean trading halt in under three weeks, all triggered by the same two stocks, which now represent half of the KOSPI’s weight, compared with about a quarter at the end of last year.

The difference in sentiment by investor class was clear:

• Foreign investors net sold ₩5.2 trillion of KOSPI stock on the day.

• Domestic retail investors absorbed ₩5.4 trillion.

The drawdown spread through the Asian supply chain. SMIC fell 11% and Hua Hong dropped 14% in Hong Kong, while Kioxia fell 10% in Tokyo after a 600% year-to-date run.

Korea’s own macro backdrop did not help either. June CPI printed 3.2%, the highest level since December 2023, strengthening the case for a Bank of Korea hike ahead of the next monetary policy meeting on July 16.

The Other Side of the Coin

On the same day, SK Hynix CEO Kwak Noh-jung detailed a ₩100 trillion, or $64 billion, domestic investment plan:

• ₩80 trillion for the M17 NAND fab, with construction starting next year and operations targeted for H1 2029.

• ₩20 trillion for the P&T7 advanced packaging facility.

This sits inside the previously announced ₩1.1 quadrillion expansion program, and comes just days after Seoul unveiled a combined ₩800 trillion national semiconductor plan with Samsung.

Does a semiconductor sell-off make sense given the acceleration in fab capacity expansion?

At the micro level, probably not.

But at the macro level, maybe. These capex plans become more fragile if hyperscaler AI capex starts to slow.

Management teams are telling investors they see enormous multi-year demand for HBM, DRAM, and enterprise NAND.

The market is interpreting something else:

Meta excess compute + rapid future supply growth = lower prices.

In sharp contrast with live sentiment, IBK and NH raised their Hynix targets to ₩4.0 million and ₩4.1 million on the same day the stock fell 15%. Morningstar had also lifted fair values on both Korean names on Wednesday, calling the memory cycle “tracking substantially stronger than expected,” while warning that the capacity now being added could set up a correction later in the decade.

In Micron’s case, the median analyst target sits near $1,550, with Cantor and Barclays both at $2,000.

There is also a secondary, less relevant headwind.

A class action was filed on June 25 in California federal court alleging that Samsung, SK Hynix, and Micron coordinated to restrict DRAM supply, citing 700% price increases over four years.

I wouldn’t worry about this, but it’s not ideal for an already noisy industry.

In summary, memory stocks are being treated as an instrument that the market holds when times are good and sells without hesitation at the slightest sign of problems in the AI buildout.

The “Excess Compute” Fear

Meta selling compute can mean several very different things. It does not automatically mean there is a market-wide structural flood of compute.

It could mean:

• Meta overbought or misallocated capacity.

• Meta wants to monetize idle capacity between internal training and inference cycles, which hyperscalers already do routinely.

• Meta’s model and product execution is lagging its infrastructure spend, which is bearish for Meta’s AI ROI, but not automatically bearish for the supply chain.

• Meta is turning infrastructure into a commercial cloud product, which arguably validates that compute is valuable enough to sell.

It would be one thing if Meta were selling compute at liquidation prices, but that is probably not what is happening.

SpaceX/Google Deal

Meta is likely trying to capture rates similar to the deal SpaceX signed with Google, which were very generous. If anything, that signals compute is still scarce and valuable, just not necessarily valuable for underperforming AI labs.

SpaceX secured pricing that was 4x the five-year neocloud average, which makes it easy to understand why Zuckerberg would want to pursue a similarly profitable opportunity.

In fact, just four days before this drawdown, we got highly contradictory rumors suggesting that Google had been rationing Meta’s access to Gemini since March because Meta wanted more capacity than Google could supply.

According to those rumors, Meta was forced to tell staff to conserve tokens, while other Google customers were also affected, though to a lesser degree.

Moreover, Pichai has called Google “compute-constrained in the near term.” Google Cloud’s backlog has doubled to more than $460 billion, capex is running above $180 billion this year, and Google is still reportedly renting bridge capacity from SpaceX at around $920 million per month for 110,000 GPUs.

This is another sign that Meta may have excess compute, but the industry does not, and what matters is the latter.

These numbers do not point to excess compute. Compute is extremely valuable for companies that know how to extract value from it.

What we are seeing is the free market allocating compute toward the players that can monetize it most efficiently.

Meta: Still an AI Lab

Moreover, Meta does not seem to have thrown in the towel on frontier models entirely.

In a recent post from Alexander Wang, Meta’s head of AI confirmed that the company is training an Opus-level model for coding.

This suggests that Meta may enter the neocloud business as another way to profit from AI, while its capex remains strong and its commitment to training frontier models stays alive.

Neoclouds: An Unclear Threat

The risk for neoclouds comes down to basic supply and demand dynamics.

If AI labs like xAI and Meta go from being huge compute demanders to huge compute suppliers, that could shift the supply/demand equilibrium.

It would be similar to AMD or Amazon becoming memory resellers instead of simply using memory themselves. Sure, NVIDIA and Broadcom would still demand memory, but if you are a memory supplier, you want a market with the highest number of potential buyers and the lowest number of potential sellers.

The same logic applies here.

If Meta sells raw capacity, it could become less dependent on CoreWeave and Nebius. And if hyperscalers broadly start reselling unused capacity, GPU-hour pricing could compress unless demand from other AI labs grows even faster on a relative basis.

However, in the case of Nebius, for example, most capacity is already tied to highly profitable contracts. Combined with the blueprint of the SpaceX/Google deal and the rest of the compute shortage across the industry, a high-quality neocloud like Nebius is not highly exposed to spot pricing.

More importantly, there is no reason to assume GPUs will suddenly be sold at liquidation prices just because of this move by Meta.

If Zuckerberg decided to pursue a neocloud strategy, it would be absurd to assume it is because he sees a compute flood coming and GPUs being sold for pennies. He clearly sees a shortage, just not inside Meta, and plans to take advantage of it by supplying the successful AI labs creating that shortage.

• I do not understand why anyone would think Zuckerberg would enter the neocloud business if he believed it was low-margin or that Meta’s entrance would crash prices.

• I also fail to understand why the market took the SpaceX/Google deal so positively, while punishing the AI supply chain so aggressively for Meta potentially doing something very similar.

After a record year in returns, the market preferred to lock in gains before properly questioning whether the AI trade is truly under threat.

That shows how immature this industry still is from a public-market perspective.

What Would Prove the Market Right?

The market would be right to punish the AI supply chain if:

• Hyperscalers cut 2027 capex.

• GPU-hour pricing starts falling across contracted deals.

• Neocloud utilization weakens.

• Meta sells capacity at distressed pricing.

• Memory pricing rolls over before new HBM and DRAM supply is absorbed.

For memory, the key risk is a 2028 supply wave arriving before AI demand is large enough to absorb it.

For neoclouds, the key risk is that contracted revenue proves less profitable after accounting for financing, depreciation, and data-center lease costs, and that steady profitability is largely never achieved.

For power developers, the risk is that they own attractive sites but fail to convert them into high-quality, financed hyperscaler contracts.

So far, nothing even close to this has happened.

In fact, everything points in the opposite direction.

However, it is still worth knowing what should make investors’ warning bells start ringing.

What Fixes This Market Pessimism?

The AI trade needs to mature from a narrative trade into a financial trade.

Right now, the market is still treating large parts of the AI supply chain as unstable and cyclical. Demand is enormous, but the path from demand to revenue, earnings, and free cash flow is still uneven.

That is what creates the violent swings.

So far, the AI trade has often been about finding the next bottleneck supplier, whether that is GPUs, HBM, power, networking, liquid cooling, data centers, packaging, or substrates.

That phase can create huge returns, but it also creates extreme narrative volatility. Stocks move violently because investors are constantly trying to price scarcity before the earnings fully arrive.

The next phase should be more mature.

It should be less about finding the next bottleneck that gets bid to stretched valuations, and more about owning the companies that can turn AI demand into revenue and earnings efficiently.

That is how mature supply chains work:

• Demand becomes revenue.

• Revenue becomes margins.

• Margins become cash flow.

• Cash flow supports higher valuations.

Neoclouds Need to Mature

Neoclouds need to prove they can become durable infrastructure businesses, not just high-beta AI capex vehicles.

The best outcome would be for the strongest neoclouds to earn a status closer to AWS, Azure, and Google Cloud.

They do not need to become as large or dominant, but they do need to become businesses the market can underwrite with the same level of financial confidence.

Until then, the market will keep treating them as risky AI infrastructure intermediaries with heavy fixed costs and uncertain long-term pricing power.

Memory Needs to Prove This Cycle Is Different

The market still assumes high memory prices eventually create too much supply, followed by a crash.

For the pessimism to fade, memory prices need to remain resilient through the next few years, even as new supply comes online.

The key proof will be that AI has structurally changed the demand curve.

That is when Samsung, SK Hynix, and Micron can rerate more meaningfully, not just through aggressive upside swings, but through consistently double-digit forward earnings multiples.

AI Needs Productivity Proof

The biggest proof is productivity.

AI will not feel like a safe long-term investment until the productivity gains clearly justify the multi-trillion-dollar capex cycle.

Companies around the world are spending enormous amounts on AI, which now needs to move from impressive demos to measurable business impact.

Once investors see that AI capex produces real productivity gains, the market will stop asking whether the spending is a bubble and start asking which companies capture the most value from it and which companies support it.

My Take and What I’m Buying

Read more

Subscribe now

In this article, I’ll cover two portfolio companies, with an updated valuation framework for each.

These two companies are:

• A defense technology company focused on drones, counter-UAS systems, directed-energy laser defense, and space laser communications.

• A data center developer and operator with three signed hyperscaler-grade HPC campus leases.

Read more

Portfolio is up 130% YTD.

Subscribe now

In this article, I’ll cover the latest updates on the companies in my portfolio, including:

• What I think about their current valuations

• The upside I see and why

• My price targets for each holding

Read more

Subscribe now

In this article, I’ll cover my new position:

• SpaceX, TSMC, and Intel supplier

• $720M market cap

• 50% gross margins

• More cash than debt

• Heavy exposure to panel-level packaging and glass substrates, the two next revolutions in the semiconductor industry

Read more