The AI Infrastructure Cycle Has More Runway Than the Skeptics Think

Editor’s Note: This is an analysis I put together late last year discussing the still strong potential for AI infrastructure investment and technology demand. I feel those same catalysts are still in play today as Agentic AI has accelerated demand. Recent earnings commentary from SK Hynix, Texas Instruments, and Intel, among others, supports this. So, I wanted to re-publish the commentary this morning (the final chart is updated through yesterdays' close)…

The AI Infrastructure Cycle Has More Runway Than the Skeptics Think

• Past technology infrastructure cycles have peaked at 2% to 5% of GDP.

• The AI infrastructure TAM is expected to triple by 2030.

• Nvidia’s forward P/E is far below Cisco’s dotcom peak.

Every cycle has its skeptics, but history favors the disciplined…

There’s an old saying in the investment world that I’ve grown more fond of with age: “It’s not about timing the market, it’s about time in the market.” In other words, wealth comes from having a plan, sticking to it, and accepting that markets swing. With strategies like diversification, steady contributions, and dividend reinvesting, wealth compounds over time.

Take the S&P 500 Index as an example. Over time, it has averaged a 9.7% total return (dividends reinvested) since 1928. That means that by staying invested, you would double your money about every 7.5 years. And that includes periods when the market faced plenty of turmoil. Yet $100 invested at the end of 1928 would be worth around $590,000 today…

The past few weeks are a sharp reminder of why long-term investing matters. After dropping 4.4% from November 10–20, the S&P 500 has already clawed back most of its losses.

One driver of that whipsaw was speculation about the Fed’s rate cuts. After last month’s decision, several officials warned against easing further in December, sparking a selloff. But late last week, the outlook flipped when FOMC Vice Chair John Williams (dove, voter) suggested a rate cut was appropriate.

Today, though, I want to zero in on another source of turmoil—talk of a bubble in tech stocks. Earlier this year, the sector made up 36% of the S&P 500, surpassing the 34.8% peak during the dot-com boom. That triggered calls for a market top.

I get it. The echoes of the late 1990s are hard to ignore. Some strategists see a chance to make a name by calling the peak. If they’re right, they’ll ride that call all the way to the bank. But based on the data I’ve surveyed, we’re not even close. The S&P 500 still has plenty of room for a steady grind higher.

But don’t take my word for it, let’s look at what the data’s telling us…

AI Infrastructure Spending

If there’s talk of an AI bubble, spending is the first stop. Naysayers point to the sheer volume and amount of investment.

This year, AI capital expenditures on data centers are expected to hit $375 billion. Alphabet, Amazon, Meta, and Microsoft account for $320 billion of that. That’s about 50% higher than Goldman Sachs’ estimate of $250 billion for 2024. No wonder skeptics are worked up.

But history puts this in perspective…

U.S. nominal GDP for 2025 is projected at just over $30 trillion. Compare that to $375 billion in AI capex and you get to roughly 1.2% of output. Based on numbers from the U.S. Bureau of Economic Analysis, OECD, and U.S. Census Bureau, that type of spending is nothing compared to the past. Prior infrastructure cycles peaked at 2%–5% of economic output. In other words, we’re nowhere near bubble territory.

Total Addressable Market Growth

Next, let’s look at growth potential. The total addressable market (“TAM”) is the biggest possible pool of buyers. Companies use it to gauge how much revenue they could capture if everyone bought in.

Here, infrastructure is key. Advanced Micro Devices CEO Lisa Su has been one of the most forthright executives about this data. After all, AI semiconductors are the backbone of what makes data centers work. So, I want to frame her commentary around the growth opportunity out to 2030 to get an idea of the potential…

In 2023, she projected $400 billion in AI infrastructure by 2027. Earlier this year, she raised that to $500 billion by 2028. After the latest earnings release, she said TAM could hit $1 trillion by 2030. (I added in the 2029 numbers to split the difference between 2028 and 2030 and give you an idea of the growth potential.)

AI’s market is set to explode because it’s becoming part of daily life, from healthcare to cars. Generative AI demands massive computing power, while personal devices are driving demand for faster hardware. Put it together, and TAM keeps expanding as AI spreads everywhere.

Infrastructure Darlings: Now vs. Then

Skeptics love comparing Nvidia today to Cisco in the dot-com era. Both were seen as the backbone of their markets. Nvidia dominates AI semiconductors; Cisco built the early internet.

The knock on Nvidia is that its surge may not pay off if demand falls short—just like Cisco, which overbuilt for an internet boom that didn’t fully materialize.

The knock on Nvidia is that the stock surge tied to AI infrastructure spending may not fully pay off if demand falls short. That’s the same trap Cisco fell into during the dot‑com era.

So, let’s take a look at the forward price-to-earnings ratio for Nvidia since 2020 compared to Cisco from 1995-2000…

Nvidia’s multiples in 2020–21 resemble Cisco’s in 1995–96. That’s where the similarities end. Nvidia’s multiple has been shrinking despite the rally, while Cisco’s exploded higher. That doesn’t look like mania buying to me.

The Nasdaq Composite

Finally, let’s compare the Nasdaq now versus then. I used 2023 as the starting point today (ChatGPT launched in late 2022) and 1995 back then (Netscape’s Mosaic browser launched in late 1994)…

Since early 2023, the Nasdaq is up about 135%. From 1995 to the same point in 1997, it gained 130%. But by March 2000, the dot-com Nasdaq had surged nearly 570%. To match that, today’s index would need a fourfold gain from here. We’re nowhere close.

Bringing It All Together

Yes, today’s market has echoes of the dot-com boom, like tech’s heavy weighting in the S&P 500. And yes, some people want to call the top.

But history shows that new technologies drive growth and investment cycles forward. That’s what happened before, and it’s happening again now. Based on precedent, we’re not near a peak. That backdrop supports a steady long-term rally in the S&P 500.

Five Stories Moving the Market:

Intel delivered a blockbuster sales forecast that beat Wall Street expectations, signaling that the long-struggling chipmaker is benefiting from the giant build-out of artificial intelligence computing – Bloomberg. (Why you should care –the solid earnings and revenue guidance is another sign that demand for compute power continues to accelerate)

Europe's largest software maker SAP reported ​a 17% increase in first-quarter profit, beating estimates, ‌due to strong growth in its cloud business; the company maintained its ​full-year outlook for overall revenue growth, including ​cloud growth at 23% to 25% - Reuters. (Why you should care – the company said it remains focused on managing costs and profitability)

U.S. President Donald Trump ordered the U.S. Navy to shoot any boat putting mines in the Strait of Hormuz, while the military said it intercepted two oil supertankers that tried to evade its effort to prevent passage to and from Iran’s ports – Bloomberg. (Why you should care – the White House said it will continue to suppress Iran’s main source of revenue until it leadership comes to bargaining table)

Cisco Systems showed a switching chip that it says will be able to connect quantum computers of different types, another ​step in its effort to eventually connect an internet of quantum machines ‌the way its gear connects the existing internet – Reuters. (Why you should care – the switches are expected to make quantum computing networks more secure)

Microsoft is offering long-tenured employees voluntary buyouts, a first for the software giant as it continues to reorganize staff around its push to accelerate its artificial-intelligence efforts; the voluntary retirement program is part of a broader shift by Microsoft to alter its performance system and how it awards bonuses and stock options – WSJ. (Why you should care – management said it’s looking for areas where it can improve efficiency, likely implying a margin boost) 

Economic Calendar:

Earnings – CHTR, HCA, NSC, PG

Japan – CPI for March

U.K. – Retail Sales for March (2 a.m.)

France – Consumer Confidence for April (2:45 a.m.)

SNB’s Schlegel (Chairman) Speaks (4 a.m.)

Germany - Ifo Business Climate Index for April (4 a.m.)

Canada – Retail Sales for February (8:30 a.m.)

U.S. – University of Michigan Consumer Sentiment (Final) for April (10 a.m.)

U.S. - Baker Hughes Rig Count (1 p.m.)

U.S. - CFTC’s Commitment of Traders Report (3:30 p.m.)

Fed Releases Balance Sheet Updates on Commercial Banks (4:15 p.m.)