What if Nvidia’s biggest opportunity for investors isn’t Nvidia itself—but the companies it desperately needs to keep the AI machine running? 🔍

Find out which suppliers are showing the strongest signs of scarcity, where pricing power is emerging, and why today’s AI bottlenecks could become tomorrow’s biggest investment risks.

Read through to the end — the framework at the close is the part most busy investors can reuse every week.

"The big money is not in the buying or selling, but in the waiting.”

— Charlie Munger

A fixed $500 a month ignores the forecast: same deposit, most months, whether the line is quiet or running.

Alphabet Inc. Class C $GOOG ( ▼ 1.24% ) closed at $335.45. Five years earlier it was about $141.46. That is a +$193.99 move, or +137.13% in total — roughly 18.8%/yr on average if you held the whole stretch. That pace is stronger than a typical long-run market baseline. It is unusual, and it is not a number to copy forward blindly.

Look for on the chart: the quieter 2022–2023 stretch, the later rise toward the $404.44 52-week high, and the pullback to $335.45 — DCA would have bought more shares on the earlier dips and fewer into the late strength.

Lesson: Past pace rarely continues. A five-year annualized rate near 19% on a very large company is a result, not a plan. Past results never guarantee the future — especially after a sharp run that has already given some of it back.

Next Horizon: another verified 5-year chart, same $500/month frame, same honest catch.

Want a cleaner look at this name? Open GOOG on Snowball Analytics — price, fundamentals, and history in one place. Context for the chart above, not a buy signal.

Sponsored resource

Nvidia $NVDA ( ▲ 0.57% ) is selling some of the most important technology in the AI boom, but its latest numbers reveal something worth looking beyond the headline growth: even Nvidia is having to spend heavily to secure the supplies needed to keep expanding.

The company’s gross margin is expected to settle around 72% to 73%, down from the exceptionally high levels reached during the AI surge. That does not necessarily signal a broken business. Nvidia was operating at roughly 72.7% gross margin only two years ago.

The more interesting detail is the cost behind that growth. Nvidia’s cost of goods nearly doubled to about $63 billion, while its supply and capacity commitments jumped from approximately $119 billion to $279 billion. Those commitments include data-center infrastructure, particularly memory and manufacturing capacity.

For investors, that raises a better question than simply asking whether Nvidia can keep growing: which companies are receiving all that money, and which of them actually control something Nvidia cannot easily get elsewhere?

AI has several obvious bottlenecks. Electricity, cooling, data-center construction and advanced chips all require enormous investment.

Elon Musk has even argued that power and cooling can be a bigger constraint than AI chips themselves. That makes companies such as Vertiv (VRT), nVent Electric (NVT) and GE Vernova (GEV) logical names to watch.

But the more overlooked shortage may be memory.

One useful way to identify a true bottleneck is to examine pricing power. A supplier experiencing rapidly increasing revenue but declining margins may simply be selling more products that other manufacturers can also produce. A supplier whose revenue and gross margins rise together may have something much harder to replace.

The analysis behind the article tested 24 companies across memory and storage, data-center equipment, power and installation. Thirteen passed the scarcity test. All 24 generated more gross profit than they did two years earlier, but the difference was in how much pricing power they appeared to have.

That distinction matters because AI demand alone does not guarantee extraordinary economics. Scarcity does.

Vertiv $VRT ( ▼ 1.17% ) is one of the clearest beneficiaries of the AI data-center buildout. Its business sits directly in the path of rising demand for power management and cooling.

The numbers are notable. Deferred revenue increased from roughly $1.8 billion to $3.6 billion in six months, while inventory rose about 73%. Revenue grew 26%, and gross margin expanded by approximately 4.3 percentage points.

That is a strong combination. Customers are committing money ahead of shipments, demand is visible, and margins are improving.

But there is an important caveat.

A 4.3-point margin expansion is meaningful, yet it is not the same kind of evidence seen in memory. It suggests Vertiv has pricing power, but customers may still have alternatives. The stock also briefly climbed about 4% following Nvidia’s earnings before giving those gains back, showing how quickly expectations can shift even for a major AI infrastructure supplier.

nVent Electric $NVT ( ▲ 0.11% ) provides an even more useful example.

Its data-center business is targeting approximately $2 billion in sales in 2026, more than double the previous year. For context, the entire company generated only about $3 billion in sales two years earlier.

Yet gross margin declined by roughly 2.2 percentage points while revenue increased 46%.

That does not mean nVent is weak. It means the shortage may not be the equipment itself.

Electrical hardware and liquid-cooling systems are essential, but more factories can eventually manufacture them. The real bottleneck may instead be permits, grid connections, construction schedules and available capacity.

That is an important distinction. A company can sell an essential product without owning the scarce resource behind the demand.

The strongest evidence in the article comes from Micron $MU ( ▲ 0.39% ) and SanDisk $SNDK ( ▼ 1.36% ).

Micron’s gross margin increased from about 37.7% to 84.6%, meaning the company was keeping roughly 85 cents of gross profit for every dollar of revenue at the highlighted margin level.

Micron has also secured 16 long-term supply agreements extending through 2030, covering approximately 25% of revenue at locked-in prices. Management has indicated that some important customers may only be able to obtain around half to two-thirds of the memory they need in the medium term, with no clear visibility on when supply will fully catch up.

That is much stronger evidence of scarcity than simply reporting record demand.

SanDisk provides a similar signal from another part of the memory market. Since being separated from Western Digital roughly 18 months ago, its gross margin has increased from around 26% to 85%, while revenue has nearly tripled.

SanDisk does not produce the high-bandwidth memory positioned directly beside Nvidia’s AI processors. Its focus is flash memory and storage. But modern data centers need enormous amounts of storage as AI systems process and retain more information, so SanDisk demonstrates that tight supply is affecting a much broader part of the memory industry.

Give a Skydive agent an ongoing responsibility and they’ll handle it on schedule, every time.

Have them prep your morning report, research new leads, monitor customer feedback, or keep projects moving overnight. You wake up, the work is already done.

Make your first agent

This is where the opportunity becomes more complicated.

Micron may appear inexpensive at around six times next year’s earnings, while its trailing valuation is in the low teens depending on how the fiscal year is measured. But memory companies are cyclical, and today's extraordinary profits cannot simply be projected forever.

Micron’s previous gross-margin peak was around 59% in 2018. About five years later, the company was losing money on a gross-profit basis.

That history is important.

The current thesis works because the market appears to be pricing in a significant decline from today's earnings. The question is whether the decline will be modest or whether the next memory downturn will be severe.

SanDisk has a similar attraction, with its valuation looking reasonable relative to a conservative bear-case scenario. But because it has only been publicly traded as a standalone company for around 18 months, there is much less historical valuation data to use as a guardrail.

In other words, both stocks may offer substantial upside if memory remains tight, but neither should be treated as a risk-free bargain.

The biggest threat to this thesis is new supply.

China is investing aggressively in domestic memory production. A Chinese memory manufacturer that went public in July raised roughly $8.6 billion, making it the largest equity offering in Asia in 2026. More than half of the equity reportedly came from government-linked sources.

Analysts expect its production to reach roughly nine wafers for every ten wafers Micron starts by year-end.

That comparison needs some caution because wafers do not translate directly into finished memory bits, and the Chinese producer is still using older equipment and producing lower-memory-content products. It also represents less than 8% of the broader memory market compared with Micron's more than 20% share.

Still, the direction matters.

The company is targeting the exact part of the market where Micron has generated extraordinary margins. If Chinese production scales successfully, additional supply could eventually push prices down and compress those margins.

Apple $AAPL ( ▼ 0.52% ) has reportedly been testing the Chinese memory for use in products such as iPhones and Macs.

That has already attracted political attention. Seven bipartisan U.S. senators wrote to Apple in July seeking a written commitment regarding the use of the Chinese memory, while the U.S. commerce secretary indicated that the administration was not supportive of the move.

Interestingly, the Chinese supplier reportedly rejected Apple's request for a discount because its production was already committed elsewhere.

That detail is easy to overlook, but it actually reinforces the scarcity argument. If Apple cannot simply demand a lower price because available production is already spoken for, the supplier still has considerable leverage.

The bigger question is whether that leverage lasts once additional Chinese capacity reaches the market.

The article also highlights several companies that have seen significant share-price declines, including Marvell Technology $MRVL ( ▲ 1.32% ), Western Digital $WDC ( ▼ 3.51% ), Lam Research $LRCX ( ▼ 0.96% ) and KLA $KLAC ( ▼ 0.63% ).

These declines do not automatically make them bargains. But they show how quickly the market can punish companies when expectations become too high, even when the long-term AI demand story remains intact.

The same idea applies to power.

GE Vernova $GEV ( ▲ 0.92% ) passed the scarcity test, with gross margin improving by about 2.2 percentage points to roughly 20%. Its order book is strong, but the company does not appear to have the same pricing power as a supplier controlling a truly scarce resource.

Constellation Energy $CEG ( ▼ 1.77% ) is more unusual. Its gross margin expanded by roughly 24.5 percentage points, making it the strongest non-memory example in the analysis.

The reason is simple: Constellation owns nuclear reactors that were built decades ago and are already licensed and operating. Those existing assets are difficult to replicate quickly.

A new power plant cannot be created simply because an AI data center needs electricity next year. That makes existing nuclear generation fundamentally different from selling another piece of equipment.

The article does not present Constellation as a completed valuation case, but the economics illustrate an important point: sometimes the scarce asset is far more valuable than the equipment used to build it.

There is also a much smaller and far riskier angle in Copper One Resources, a Canadian junior copper explorer.

Copper recently moved above $6.70 per pound on COMEX, up roughly 40% over 12 months, yet Copper One had fallen about 75% in 2026 at the time discussed in the article. The company had a market capitalization of only around C$14.8 million and approximately C$10.4 million in cash.

Its projects include Majuba Hill in Nevada, a former producer, along with the Red and Red Hill projects in British Columbia, with drilling underway at Red.

This is not comparable to Micron, Nvidia or Constellation. Copper One is an extremely early-stage exploration company, meaning the potential reward comes with substantially higher geological, financing and execution risk.

Its inclusion reinforces the broader theme, however: AI requires physical resources far beyond GPUs, and copper is one of the materials needed to build the electrical infrastructure supporting the expansion.

The AI boom has made it easy to focus on the companies with the biggest revenue growth.

A better question may be:

What does the AI industry absolutely need, and who has the power to say no?

That question changes the way companies such as Nvidia, Micron, SanDisk, Vertiv, nVent Electric, GE Vernova and Constellation Energy should be viewed.

It also puts Marvell Technology, Western Digital, Lam Research, KLA, Apple, SK Hynix and even highly speculative names such as Copper One Resources into a much broader supply-chain picture.

The strongest businesses may not simply be those selling the most products. They may be the ones controlling an input that customers cannot easily replace.

For now, memory appears to provide some of the clearest evidence of that scarcity. Micron and SanDisk stand out because their margins are telling the same story that Nvidia’s massive supply commitments are hinting at: the AI industry is not only hungry for chips. It is competing for everything those chips need to function.

That creates an opportunity, but it also creates a warning.

High margins attract new competitors. New competitors eventually create new supply. And in cyclical industries such as memory, the transition from shortage to oversupply can happen much faster than expected.

For a long-term portfolio, the goal is therefore not to chase whichever AI supplier has the most exciting headline. It is to identify the companies sitting closest to the actual bottleneck, understand how long that bottleneck can last, and make sure the valuation still works when conditions eventually become less perfect.

Sponsored resource

Want a simple next step for the $500 habit?

Our Fast Track Blueprint walks through a trusted online broker setup and a clear portfolio starting path.

Thanks for reading. This format is built to be fast to open, clear to understand, and useful enough to act on — without pretending past returns continue forever.

Disclaimer: This newsletter is for informational purposes only and is not financial advice. Consult a qualified advisor before investing.