Nvidia, AMD, and the AI Spending Debate: What It Means for the Rest of Us

Three tickers have taken over my group texts this week. NVDA. AMD. NBIS. If none of those mean
anything to you yet, stick with me, because there is a decent chance you already own a piece of at
least one of them, even if you have never bought a share on purpose.
Here is the short version of why everyone is talking about them. Nvidia reports earnings on August
26, and Wall Street is expecting revenue around 91.9 billion dollars for the quarter, which tells you
something about the scale we are dealing with. AMD just closed the largest bond offering in its history
to help fund its own AI buildout and is sitting near its 52 week high. And Nebius, a much smaller AI
infrastructure company most people had never heard of a year ago, has seen its stock climb more
than 300 percent in 2026 after landing some of its biggest cloud computing deals yet. At the same
time, a well known investor who correctly called the 2008 housing crash has reportedly taken a short
position against some of these AI infrastructure names. Smart, informed people are on both sides of
this trade, and that alone should tell you something.
Underneath the ticker symbols is a genuinely interesting debate, and it is worth understanding even if
you never plan to buy an individual tech stock in your life. The cost of an AI “token,” which is roughly
the unit that measures how much you pay to run one of these AI models, has been falling fast. Some
companies have cut prices on their newest models by 75 to 80 percent this year alone. You would
think that means the AI spending boom should be cooling off. Instead, the opposite is happening.
Amazon just raised its annual capital spending guidance to 220 billion dollars and says it still expects
to run short on capacity through 2027. Meta spent somewhere between 130 and 145 billion dollars
building out AI infrastructure this year. Nvidia is reportedly helping arrange roughly 250 billion dollars
in financing tied to a single data center project valued north of 500 billion dollars.
So which is it? Is AI getting cheaper, or is it getting more expensive? Economists have a name for this
pattern. It shows up whenever something gets more efficient and, instead of total spending going
down, total spending goes up because cheaper access unlocks a lot more usage. Cheaper electricity
a hundred years ago did not shrink the average power bill, it filled homes with appliances nobody had
budgeted for. Something similar may be happening here. Nobody knows yet whether this AI buildout
ends up looking like the interstate highway system, essential infrastructure that pays for itself for
decades, or like the fiber optic buildout of the late 1990s, where real technology got built but a lot of
the money that built it never came back. Reasonable, well informed people land on both sides of that
question, and I am not going to pretend I know the answer, because nobody actually does yet.
Here is why this matters to you even if you have never looked at a stock chart in your life. If you have
a 401k, an IRA, or really any account invested in an S&P 500 index fund or a target date fund, you
almost certainly have real exposure to this story already, because a small handful of AI related
companies now make up an unusually large share of the major stock indexes. That is not a reason to
panic and it is not a reason to chase. It is a reason to actually know what you own.
This is a “plan long, resist short” moment if I have ever seen one. Nobody can tell you today whether
the AI infrastructure boom is a bubble or a foundation, and anybody who tells you they know for
certain is selling you something. What we can control is whether your own portfolio is built to survive a
few years of “we do not know yet” without betting the farm on any one story, name, or narrative. That
is the whole point of spreading the risk instead of concentrating it.
If this has you wondering how much of your retirement account is quietly riding on a handful of AI
names you never chose, that is a great question to bring to a one on one review. You can reach us at
IronEagleAdvisors.com or 434-465-6485.
None of this is a recommendation to buy, sell, or hold any individual stock. It is meant to help you
understand a conversation you are probably already hearing pieces of, so you can ask better
questions about your own accounts.