Listen to the audio version of this article (generated by AI).
Oil is rattling Wall Street, but we’re seeing a different story play out regarding the AI boom… A story told by chip shipments, memory profits, and data-center rents.
Imagine owning a factory where orders keep coming in. Your customers want more equipment, while the landlord next door is filling new buildings almost as quickly as they open.
Then you check your business valuation.
The price has fallen.
That’s how AI investors feel today. Across the infrastructure supply chain, companies are giving us breadcrumbs that this spending boom has substance. Yet stocks remain under pressure. Why? Because oil prices and interest rates are making investors nervous.
Corporate results tell us whether the underlying businesses are growing. Oil and bond yields help determine how much investors are willing to pay for that growth.
Right now, those forces are pulling in opposite directions.
The opportunity comes from figuring out which has more staying power. And increasingly, we think the evidence favors AI earnings.
Here’s what the latest numbers reveal, and where we’re looking as the market sorts through the noise.
The AI Boom Is Showing Up in the Books
Taiwan Semiconductor Manufacturing Co. (TSM), better known as TSMC, reported September revenues up 54.6% year-over-year. That brought third-quarter sales to approximately $46.7 billion, up roughly 50% and above the top of the company’s guidance.
Samsung’s preliminary third-quarter results point to operating profit nearly nine times the year-earlier level, as AI memory demand and constrained supply strengthen pricing.
An AI accelerator needs memory to work effectively, so building more computing capacity creates demand across several components, and shortages can give critical suppliers considerable pricing power.
Then there’s Applied Digital (APLD), which develops and operates the facilities housing that equipment. Its fiscal first-quarter revenue surged 322% to $341.9 million. But recurring base rent from its high-performance computing business reached $65.8 million, up from $44.1 million the previous quarter.
That’s roughly 49% sequential growth in rent.
Construction projects are becoming operating assets, as tenants pay to use them.
Applied Digital’s total revenue includes substantial low-margin tenant fit-out work. Investors shouldn’t confuse every dollar of construction-related revenue with a dollar of recurring rental income.
Still, the direction is encouraging. Applied’s management expects to bring more than 600 megawatts into service over the next 12 months, compared with 250 megawatts over the preceding year.
Taken together, these reports show strength across chip manufacturing, memory, and data-center operations. That breadth makes the case for a healthy investment cycle more persuasive than any single blockbuster earnings release.
Why Good Businesses Can Have Bad Trading Days
So why isn’t Wall Street celebrating? Because Brent crude has pushed toward $105 a barrel.
Renewed Middle East attacks and threats to Saudi energy infrastructure have increased the premium traders demand for potential disruption. Hurricane-related shutdowns in the U.S. Gulf have added supply pressure.
Expensive energy works its way through the economy. Transportation costs rise. Consumers have less money left over. Businesses face pressure on margins. As we know, stubborn energy inflation can delay interest-rate relief. Higher yields make bonds more competitive with stocks and reduce what investors will pay today for future earnings. Companies financing large construction programs also face higher borrowing costs.
That’s how an improving business can have a falling share price.
Let’s consider the following simple hypothetical… Picture a company earning $1 per share and trading at 30x earnings, giving it a $30 stock price. Now, it’s earnings rise 20% to $1.20, but investors lower the valuation to 24 times earnings. That means the stock falls to $28.80. Despite its growing business, the valuation reset makes it seem like an underperformer.
That arithmetic helps explain the tug-of-war playing out across growth stocks. It also explains why the market doesn’t necessarily need a dramatic collapse in oil prices to recover. If yields stabilize and valuations stop contracting, continued earnings growth can begin doing more of the work.
Keep an Eye on What’s Next
Meanwhile, the next wave of AI infrastructure spending is seeking financing.
Broadcom (AVGO) is reportedly pursuing more than $50 billion to finance OpenAI’s custom AI chips, with Apollo and Blackstone among prospective lenders. Alongside SpaceX’s (SPCX) reported $40 billion effort to finance Nvidia-chip purchases, that represents more than $90 billion in potential hardware financing.
The key word here is “potential.” Neither deal has closed yet, and some funding would support previously announced projects. We shouldn’t count the entire amount as newly announced demand.
But securing capital is crucial. A deployment plan becomes much more consequential for suppliers when a customer can actually pay for the equipment. The opportunity extends beyond chips to networking, power, cooling, and the facilities connecting everything.
Of course, debt must be repaid. If customers fail to earn sufficient returns on computing capacity, today’s financing boom could become tomorrow’s spending slowdown. That’s why we’re watching execution alongside ambition: delivered capacity, recurring revenue, financing terms, and evidence that customers need what’s being built.
Applied Digita’s latest Polaris Forge 1 debt tranche carried a 7% coupon, compared with 9.25% on its first notes. Progress on construction and improved credit support can help a project secure better financing, even in a difficult rate environment.
The Bottom Line on a Fall Boom
Our bullish view rests on a straightforward judgment: AI earnings growth is likely to outlast the latest energy shock.
Iran’s reported continued review of U.S. proposals leaves a diplomatic path open. A meaningful agreement would still need to produce safer shipping and fewer attacks, but there is a route toward lower energy risk premiums.
A resolution by November is possible. It is a scenario, not a deadline investors should take for granted.
The risk is that disruption persists long enough to weaken demand and keep financing expensive. Confirmed damage to production capacity would make the oil problem harder to dismiss.
For now, however, the corporate evidence supports using October weakness to selectively accumulate companies delivering on the AI buildout. We favor financial strength, credible execution, and products customers actually need.
And that last point takes us to the next opportunity.
That is, every leap in AI’s capabilities creates another set of physical requirements. More computing demands more power. More automation requires more specialized components. Moving infrastructure into orbit would require an entirely new construction and supply chain effort.
Elon Musk’s businesses sit at the intersection of those ambitions.
In my , I explore how his companies could become more closely integrated – including my thesis about a potential Tesla–SpaceX combination – and what that could mean for the businesses supplying them.
Which companies could to turn these ambitions into operating businesses?
One area I examine is America’s rare-earth and magnet supply chain, where
Today’s earnings reports show what happens when infrastructure spending starts reaching suppliers. XPANSE explores where another wave could emerge.
