Welcome to this week’s HX Weekly!

We are going to take a break from our normal HX Weekly program and instead bring you discussions on what is on everyone’s mind right now... NVDA!

We have a long history of sharing views on NVIDIA Corporation (NVDA). With it being top of the news once again this week, we are providing some new views along with some previous notes that are still relevant.

We hope you find them helpful!

Thoughts on the Market

Who on Earth Can Replace Jensen Huang?

Hi all, James from the HX Research team here!

I'm jumping in with a guest editorial this week on a topic I find fascinating.

While Wall Street has been collectively freaking out over NVIDIA's stock price, a new question has emerged following the stunning resignation of its legendary CEO, Jensen Huang.

Who on earth will replace him?

NVIDIA's board wasted little time naming Chief Financial Officer Colette Kress as interim CEO.

With the company scheduled to report earnings next week, I can't imagine they wanted to head into one of the most closely watched earnings calls in history without someone at the helm.

With all due respect to Kress, who has played an important role in NVIDIA's success, I suspect she'll remain exactly what her new title suggests: an interim CEO.

After all, we're talking about arguably the most powerful CEO position in technology, if not the entire corporate world.

And the speculation has already reached extraordinary levels.

President Donald Trump has weighed in. CNBC's Jim Cramer and his colleagues are throwing around names.

Wall Street analysts and Silicon Valley insiders all have their favorites.

So, I thought it would be fun to explore some of the candidates.

But more importantly, let's revisit two famous CEO transitions in technology history.

Because replacing a legendary leader can either usher in a new golden age or mark the beginning of a company's decline.

The Most Important Job Opening in the World?

First, let's look at the candidates.

President Trump has suggested former Apple CEO Tim Cook, an intriguing choice given his extraordinary track record running one of the world's largest companies.

Cook knows how to manage complex supply chains, deal with world leaders, and oversee a global technology empire.

But here's the question: Does NVIDIA need a world-class operator like Cook, or someone with a deeper understanding of the technology itself?

That brings us to Rene Haas, CEO of Arm Holdings.

Haas previously worked at NVIDIA, knows the semiconductor industry inside and out, and has experience running a major publicly traded technology company. On paper, that's a pretty compelling combination.

Then there's Ian Buck, NVIDIA's vice president of hyperscale and high-performance computing.

Buck is deeply involved in the technology that helped make NVIDIA dominant in AI. He understands the company's products, customers, and engineering culture.

But would a brilliant technical leader necessarily make a great CEO of a company worth trillions of dollars? That's a very different job!

Other names being discussed include AMD CEO Lisa Su, one of the industry's most accomplished leaders, and Apple's chip engineering chief Johny Srouji.

Su would be a fascinating choice, although persuading the CEO of NVIDIA's chief competitor to switch sides would be quite a feat!

Personally, I find the contrast between Cook, Haas, and Buck particularly interesting.

Each represents a different kind of leadership, and the choice could tell us a great deal about where NVIDIA's board believes the company needs to go next.

But history tells us that impressive credentials don't always translate into successful leadership transitions.

Consider two examples.

Apple: The Success Nobody Expected

When Steve Jobs stepped down as Apple's CEO in August 2011, investors had every reason to worry.

Jobs wasn't simply Apple's CEO. He was its creative force, public face, and the visionary behind products that had changed the world.

Could anyone possibly replace him?

His successor, Tim Cook, was an operations expert who had spent years making Apple's supply chain one of the most efficient in the world.

Many questioned whether he could sustain the extraordinary innovation Jobs had delivered.

Well, we know how that turned out!

Under Cook, Apple expanded its services business, introduced new products, and grew into a technology giant worth trillions of dollars.

Apple's stock increased more than tenfold during his tenure, even before accounting for dividends.

Apple Stock Performance During Tim Cook's Tenure (2011 – 2026)

Cook didn't try to become another Steve Jobs. He built on Apple's existing strengths while developing new ones.

The lesson?

A legendary founder doesn't necessarily need another visionary to replace him. Sometimes an exceptional operator is exactly what a company needs.

Intel: How to Lose a Technology Empire

Now, let's look at the other side of the coin.

During the 1990s, Intel was one of the most dominant technology companies on Earth.

Its processors powered the personal computer revolution, and its manufacturing capabilities were the envy of the semiconductor industry.

Its legendary CEO, Andy Grove, sometimes called "the father of Silicon Valley," was widely considered one of the greatest business leaders of his generation.

Grove stepped down in 1998, handing control to Craig Barrett. Paul Otellini followed in 2005.

Intel remained successful for years, but under successive leadership teams, the company made decisions that would eventually cost it dearly.

Perhaps the most famous was failing to capitalize on the smartphone revolution.

Intel had the opportunity to become a major supplier of processors for Apple's iPhone but passed on the deal.

Meanwhile, competitors using Arm-based designs came to dominate mobile computing.

Later, Intel struggled with manufacturing delays while rivals such as AMD and NVIDIA gained ground.

Intel (INTC) vs. Nvidia Corp (NVDA):  Stock Performance (1990 – Today)

A company that once seemed almost impossible to challenge had lost its technological edge.

The lesson?

Dominating one generation of computing doesn't guarantee you'll dominate the next.

The next NVIDIA CEO must anticipate where computing is headed five or ten years from now.

Get that wrong, and even the strongest competitive position can disappear.

What Does This Mean for NVIDIA?

Two technology giants. Two legendary leaders. Two very different outcomes.

Apple proved that a company can thrive long after its visionary founder steps aside.

Intel demonstrated how even the most dominant technology franchise can lose its edge when leadership fails to anticipate the next big shift.

And that's what makes NVIDIA's choice so consequential.

Huang spent more than three decades building NVIDIA into the dominant force in AI computing.

He understood the technology and cultivated relationships with the world's largest customers. He made enormous bets years before their potential became obvious.

Those are exceptionally difficult shoes to fill.

Personally, I think NVIDIA should prioritize someone with deep semiconductor expertise who also understands how to manage a global technology company.

That makes Haas and Buck particularly interesting to me, although Cook's extraordinary record deserves serious consideration.

But here's the most important point for investors.

The identity of NVIDIA's next CEO matters far more than the market's immediate reaction to Huang's departure.

At HX Research, we spend a lot of time studying the relationship between corporate performance and stock prices.

While the market is understandably focused on today's uncertainty surrounding NVIDIA, its long-term future will depend on something much more important.

Can the company maintain its technological lead, continue growing earnings, and successfully navigate the next phase of the AI revolution?

Those are the questions investors should be asking as the board searches for its next leader.

Because the person who gets this job won't simply inherit one of the world's most valuable companies.

They'll inherit the responsibility of guiding an entire industry through what may be its most important decade.

The board doesn't need to find another Jensen Huang.

It needs to find someone who can ensure NVIDIA doesn't become the next Intel.

And THAT may be the most important decision in corporate America today.

A few weeks ago, Enrique published a note as part of his Truth and Trends free newsletter where he laid out the case that NVDA was going to lose investors more money over the next three years than any other stock in human history.

Little did Enrique know that instead of three years, it would take just three weeks.

While we think NVDA is a screaming buy right now in the short-term, our multi-year price targets remains $75 per share or -54% from current levels. Here is the note...

HX Weekly Redux

NVIDIA: The Biggest Stock Massacre in History

Are you sitting down?

You should be.

Because what I’m about to tell you will sound crazy at first.

Maybe even impossible.

But by the time I’m done, I think you’ll see it the way I do.

I believe NVIDIA is about to lose investors more money over the next three years than any stock in human history.

And before you click away thinking, “Good thing I don’t own NVIDIA”… You almost certainly already do.

The chipmaker makes up nearly 8% of the entire S&P 500.

That's one company out of 500.

And it’s more than 8% of the Nasdaq-100, the index behind QQQ, one of the most popular funds on the planet.

So, if you own an index fund… a 401(k)… an IRA… a target-date fund… You have money riding on NVIDIA right now.

Maybe a lot of it. Whether you know it or not.

Millions of Americans are in the exact same position. Their retirements are chained to the fate of a single stock.

And Wall Street is walking them straight into a trap.

NVIDIA's $1 Trillion Problem

Over the past three fiscal years, NVIDIA generated about $407 billion in total revenue.

Now Wall Street expects the company to generate roughly $1.1 trillion in combined sales during fiscal 2027 and 2028.

That's nearly three times the revenue NVIDIA earned in the previous three years combined.

And the forecasts get even more ambitious.

By fiscal 2029, analysts expect annual sales to approach $900 billion. Some believe NVIDIA could reach $1 trillion.

But where will all that money come from?

Most of NVIDIA's growth depends on companies spending vast sums to build and equip AI data centers.

Yet projects face power shortages, delays, and rising costs. And some customers are developing their own chips.

Wall Street is making two enormous bets: that customers can fund this spending on schedule, and that NVIDIA will capture the lion's share.

I think both assumptions are becoming harder to defend.

Billions in Chips With Nowhere to Plug Them in

On Oct. 20, Business Insider reported that Amazon planned to receive about $5 billion in NVIDIA server racks, followed by roughly $3 billion more each month.

But some Amazon data centers don't have enough electricity to run the equipment. So Amazon is storing completed systems in warehouses until those sites are ready.

Amazon calls this “Early Rack Receipt” and says it's normal practice. Not all the equipment is sitting idle.

But here's the risk: NVIDIA can sell billions of dollars in chips long before customers can put them to work.

If Amazon's warehouses keep filling while its data centers wait for electricity, why would it continue ordering at the same pace?

Amazon isn't alone.

Microsoft recently walked away from a planned data center in England run by NVIDIA-backed Nscale.

The power grid reportedly wouldn't be ready until the 2030s.

Oracle and OpenAI also dropped plans to expand their Stargate data-center site in Texas earlier this year.

These changes don't prove NVIDIA lost chip orders. Projects could move or change.

But they show how easily ambitious spending plans can slip behind schedule.

And that's a major problem for NVIDIA's forecasts.

The Money Is Getting More Expensive

Even if the power problems are solved, another question remains.

How will NVIDIA's customers pay for all these chips?

Oracle plans to raise up to $50 billion through debt and equity-related financing.

Meta expects to spend $135 billion to $145 billion on AI infrastructure this year and is exploring new ways to raise capital.

Meanwhile, yields on some CoreWeave bonds recently approached 16%.

That's an extraordinary borrowing cost for a company whose growth depends on building more data centers.

As financing costs rise, companies may have little choice but to delay projects or cut spending.

And there's another concern.

Some AI infrastructure debt is being placed in special-purpose companies and financed through loans or securities tied to data centers and equipment.

In some cases, even NVIDIA chips serve as collateral.

The structure isn't identical to the mortgage-backed securities that helped fuel the 2008 financial crisis.

But the growing web of borrowers, lenders, and investors should sound familiar.

If one major borrower runs into trouble, the damage could spread well beyond that company.

NVIDIA Knows the Risk

Here's what makes this story especially striking.

NVIDIA warns about this very problem in its own SEC filings.

In its latest quarterly report, the company states:

“Customers may postpone purchasing new architectures due to the lack of availability of data center infrastructure to deploy our products.”

And NVIDIA has already made huge bets on future demand.

As of July 26, it had $279 billion in manufacturing and supply commitments, up from $119 billion just one quarter earlier.

These are NVIDIA's commitments to suppliers, not firm customer orders.

But they show how much the company is counting on future sales.

If customers change their plans, NVIDIA could face extra inventory, higher costs, and pressure on profit margins.

It doesn't need to lose its customers for this to happen. It just needs them to buy less than Wall Street expects.

The King Has New Challengers

Some of NVIDIA's biggest customers are finding other ways to power AI.

On Oct. 27, Oracle announced plans to offer Google's custom AI chips, known as TPUs, starting in 2027.

Oracle builds and operates data centers for customers including OpenAI. Its cloud customers will soon have another option beyond NVIDIA.

Oracle is still expanding its NVIDIA business. But the new deal signals change.

Reports on Oct. 28 also indicated that OpenAI plans to buy Google's AI chips.

OpenAI is also working with Broadcom on custom chips.

Amazon has its own Trainium chips. China's Huawei is building rivals. SpaceX and Tesla are also pursuing plans to make their own chips.

NVIDIA still has a powerful advantage in CUDA, the software platform developers use to build and run programs on its GPUs.

That makes it difficult for many existing customers to switch away from NVIDIA.

But Google and Amazon don't need to beat NVIDIA at its own game. They can offer customers an entirely different way to run AI.

None of these efforts will replace NVIDIA overnight.

But consider the timing.

Customers may slow orders while they work through chips they've already bought. And when they start buying again, some of that money could go elsewhere.

NVIDIA could remain the world's top AI chipmaker while winning a smaller share of future spending.

What If Wall Street Is Wrong?

Let's return to NVIDIA's revenue forecasts.

Wall Street expects the company to generate roughly $1.1 trillion in combined sales during fiscal 2027 and 2028.

But what if customers can't keep spending at that pace?

What if power shortages delay more data centers, rising borrowing costs force companies to scale back, and some of their chip budgets shift to competitors?

Instead of $1.1 trillion, imagine NVIDIA generates just $600 billion over those two years.

That would still be an extraordinary business.

But it would mean roughly $500 billion less revenue than Wall Street expects, a shortfall of about 45%.

Now, I'm painting an extreme picture here.

But I'd argue that Wall Street's revenue forecasts are also extreme.

And here's what concerns me most.

NVIDIA might not need to miss those forecasts by anything close to 45% for its stock to suffer serious damage.

Even a much smaller shortfall, paired with weaker guidance, could force analysts to cut future earnings estimates.

Investors might also decide NVIDIA no longer deserves the premium valuation they've assigned to its shares.

Under those circumstances, would a 30% decline in NVIDIA's stock really be unreasonable?

I don't think so.

At a market value of roughly $5.25 trillion, that would erase about $1.6 trillion in shareholder wealth.

But given NVIDIA's enormous market value, it wouldn't take a collapse in its business for shareholders to suffer historic losses.

And the consequences won’t stop there.

With so much of the AI boom tied to NVIDIA's success, a sharp drop in its stock could shake confidence in the entire trade.

Which brings us to the biggest risk of all.

The Biggest Card in the House

NVIDIA sits at the center of a vast network of AI companies, investors, and lenders.

Some firms buy NVIDIA chips while also receiving funding or other financial support from NVIDIA.

The company has even agreed to buy computing power from certain AI cloud firms under specified conditions, helping support their expansion plans.

There's a certain irony in this chart.

Given everything we've just discussed about a possible AI bubble, it's a fitting image.

The biggest bubble in the chart belongs to the company whose success is helping hold the entire AI boom together.

These relationships help fuel the AI boom.

But if chip orders slow, investors may begin questioning whether heavily indebted data-center operators can earn enough to repay their loans.

NVIDIA could also become less willing to fund the firms buying its products.

A slowdown at NVIDIA could put pressure on the same companies that have helped drive its growth.

When the AI King Falls

NVIDIA could remain the world's leading AI chipmaker for years.

But Wall Street has set expectations so high that even this extraordinary company may struggle to meet them.

Its customers face growing financial pressure, data center projects are falling behind schedule, and competition is heating up.

That's why I believe NVIDIA is headed for a historic destruction of shareholder wealth over the next three years.

And remember, you don't need to own NVIDIA directly to get caught in the fallout.

Its shares are a major part of the S&P 500 and Nasdaq-100, putting millions of retirement accounts at risk.

Remember, NVIDIA doesn't have to lose its crown for my prediction to come true. It only has to disappoint Wall Street.

Our skepticism of NVDA is not new. We have been pointing out the risk for several years. Now, this means we have been wrong.

We first wrote a negative note on the stock back in August 2024. On a split-adjusted basis, the stock was trading at $125 per share when we published this note. So, it is up +32% from when we published two years ago.

The argument we made then, though, is still true and even MORE valid now. Here is that note and we look forward to republishing these notes again when it hits our $75 target. Enjoy your weekend.

Why NVDA Will Be a Short

Know Your History!

At the end of February, we published a note on NVIDIA Corporation (NASDAQ: NVDA) cautioning investors to DO NOTHING with the stock.

Wow – we were wrong, right?

At the time, the stock was around $80 per share (split-adjusted), and within a few months (June), it would hit a high of over $140.

If you were a fan of NVDA and followed our advice, you missed a nice money-making opportunity.

We were wrong.

Our calls on stocks, though, are always about probability. Our view then was that the chance that you would make a lot of money soon was low. That view was wrong.

The note we wrote about NVDA at the time, though, was not focused on the short term. It focused on what we think will EVENTUALLY happen.

We have been active participants in the stock market for three decades, and I became a full-time portfolio manager back in 1997. JUST in time to see the “Dot-com” bubble.

While this time is not the same, we are seeing many similarities with the current enthusiasm about artificial intelligence, or "AI.”

What is most interesting to us is not the similarities in stock price movements but the company fundamentals.

In honor of “NVDA Day” today, we are republishing that note along with our recollections from that period.

We don't know what NVDA will report today. They will most likely report another great quarter, and it will (again) soar to new highs.

We think it is an excellent company riding a substantial fundamental wave of demand that will continue for decades, JUST like the Internet twenty-five years ago.

That said, we think it is important that investors know their history because NVDA's growth will eventually slow (and even go negative), and you will need to be prepared.

Here is what we wrote…

We were thinking about what is going on with NVDA and similar situations we have seen in our careers.

As we were doing so, we saw a great note from our old friend Dan Ferris at Stansberry Research. He mentioned a similar situation that neither of us was around for but was as impactful as artificial intelligence—the introduction of radio.

As he says, every "bubble" has that ONE company that is the "it" company for that moment. For radio, it was Radio Corporation of America ("RCA"). This was the "picks and shovels" company of the radio rollout, selling radios and parts to enthusiasts.

As Dan retells it, the first radio station started broadcasting in November 1920, and by 1929, there were nearly 700 stations across the country. Sales of radio equipment went from $60 million a year in 1922 to $842 million in 1929.

Just about any stock with the word “radio” in the name went soaring. Think of what happened with anything with the word "crypto" a few years ago.

RCA was the leader of the pack. It went from a split-adjusted share price of $1.17 in 1921 to $114.75 by 1929. That is a +9700% return!

Here is a chart we found…

THEN it crashed down to $2.62 by 1932. It didn’t reach that old high again until the 1960s…

This isn't to say that NVDA is going to do what RCA did but rather to point out that there are always significant trends that will change the world and a company that is leading the way. When they go to the moon, though, you need to be careful.

While we weren’t around for the advent of radio and RCA, we were around for what happened during the “dot com” era. Folks often forget that the primary catalyst for the boom was the literal physical buildout of the Internet.

Much like in the radio era, the demand for equipment to build the backbone of the Internet was absolutely booming!

Telecommunications companies (whose stocks were soaring) were buying tons of equipment.

Many folks recently have been referencing what happened to the stock of Cisco Systems, Inc. (CSCO) back then.

Dan points out sales grew from $5 million in 1988 to more than $22 billion by 2001. An incredible +4000% increase! That was REAL demand…

When revenue grows like that, the stock price is likely to rise. Here is the stock chart from back then…

The stock peaked on March 27, 2000, at more than $80 per share.

We are guessing you know what happened next, but here is that chart…

Almost 24 years later, the stock STILL has yet to reach those old highs.

We want to share an important chart regarding CSCO. It shows the revenue growth from 1993 through the present day…

We spoke about the huge growth across the 1990s, but look at what happened at the end of this period—CSCO revenue went down.

It would continue to move higher and more than double, but it was going in the other direction for a three-year period.

Could this happen to NVDA?

It certainly doesn't feel like it right now, as they are crushing numbers and growing like crazy. We remember those CSCO reports back in 1999, though, and they felt the same.

Back then, CSCO made the "must have" networking equipment as every telecommunications company on Earth scrambled to build the Internet. These competitive companies would pay virtually anything at the time to get the equipment. In their mind, they couldn’t afford to be left out!

Eventually, they got enough equipment to build out what they initially needed. They built out much MORE than was initially required and didn't need to buy much equipment for a while.

CSCO also built more capacity and was able to satisfy these customers. Unfortunately, they did so right as the customers realized that – in their excitement – they had built more than they needed in the near term.

In the meantime, competitors to CSCO, who were way behind, either caught up with better products or with attractive pricing. The product wasn't as good, but it was a lot cheaper, more available, and "good enough.”

THIS is what we think eventually happens to NVDA also.

It will still be a leader. Revenue will likely grow multi-fold from these levels. It is still an awesome company in an awesome area.

The market, though, catches up, especially in equipment like semiconductors.

Want to know the scariest article we have seen about NVDA recently?

The article discussed how NVDA's AI chips were selling in the secondary market for 2x or 3x their cost from the company itself. These chips were not cheap to begin with!

This means these chips are so in demand that buyers are willing to pay a premium.

What do we think of when we see THIS kind of demand for manufactured goods like semiconductors? SUPPLY.

NVDA is going to create the supply to satisfy this demand. So will their competitors.

We think NVDA is a great company in a great position, but EVENTUALLY, this is going to end badly.

BE PREPARED.

We hope that you’ve enjoyed this week’s issue of HX Weekly…

Do you have any thoughts, questions, or feedback? Tell us more in the comment section or at [email protected].