AI Data Centers: The $5T Infrastructure Race

Data centers are the physical backbone of the AI revolution, and Marc Chaikin has strong opinions about which companies are positioned to win and lose as the buildout continues. The stakes are enormous: an estimated $5 trillion will be spent on data centers in the U.S. alone.

Oracle (ORCL) is the second stock on Chaikin’s hitlist, rated Bearish. The story is a cautionary tale about what happens when a company bets the farm on a customer that may not be able to pay. Oracle had been building a data center in Texas for Elon Musk. When Musk walked away, a LinkedIn message from an OpenAI executive started a chain of events that led to a $300 billion megadeal between OpenAI and Oracle.

Oracle went from a dependable software and cloud services company to a freewheeling architect of the data-center boom. OpenAI promised to pay Oracle $300 billion from its future revenues. The problem: future revenues are all OpenAI can guarantee, because right now, they don’t make any money. It recently leaked that OpenAI’s own CFO told insiders that OpenAI will not be able to pay back its data-center contracts if revenue doesn’t grow fast enough. It’s not growing fast enough.

Oracle’s stock price got cut by more than half. The company executed one of the largest modern layoffs of the AI era, sending 30,000 employees packing. For context, when Lehman Brothers collapsed in 2008 and sent the global financial system into crisis, they laid off 25,000 people.

On the buy side, Chaikin recommends Fabrinet (FN), a company most investors have never heard of. Fabrinet (FN) is the second free pick, rated Bullish by the Power Gauge. Fabrinet builds the physical infrastructure that AI data travels on: transceivers, photonics systems, and fiber optics. As Chaikin puts it, while everyone talks about who’s building the advanced chips that power AI, no one pays attention to who’s building the high-speed superhighway that all the AI data travels on.

Fabrinet has nearly $1 billion in cash and zero long-term debt. Its partners include Nvidia, Cisco, and Amazon. The key insight: even when frontier AI changes the computing architecture, the data infrastructure doesn’t change, so Fabrinet’s hardware becomes more critical, not less. The faster AI gets, the more critical Fabrinet’s hardware becomes.

The contrast with Oracle is stark. Oracle is drowning in debt building data centers for a startup that can’t pay its bills. Fabrinet is swimming in money building the frontier AI superhighway for some of the biggest companies on earth. The Power Gauge rates Oracle Bearish and Fabrinet Bullish, reflecting the institutional money flow into Fabrinet and out of Oracle.

The contrast between Oracle and Fabrinet illustrates Chaikin’s broader thesis about frontier AI. Oracle is building infrastructure for today’s AI, the kind that waits for human prompts. Fabrinet is building the superhighway that all AI data travels on, regardless of what the computing architecture looks like. When frontier AI changes how computing works, the data infrastructure does not change, so Fabrinet becomes more critical, not less.

The Power Gauge is Marc Chaikin’s proprietary stock-rating system. It uses 20 factors across four categories to rate more than 5,000 stocks as Bullish, Bearish, or Neutral. The system appears on every Bloomberg and Reuters terminal in the world and is used by hundreds of banks, hedge funds, and brokerage platforms.

The four categories are Financials (LT Debt to Equity, Price to Book, Return on Equity, Price to Sales, Free Cash Flow), Earnings (Earnings Growth, Earnings Surprise, Earnings Trend, Projected P/E, Earnings Consistency), Technicals (Relative Strength, Money Flow, Persistency, Price Strength, Price Trend ROC, Volume Trend), and Experts (Earnings Estimate Trend, Short Interest, Insider Activity, Analyst Rating Trend, Industry Rel Strength).

The most important factor is the Chaikin Money Flow indicator, which tracks when institutions are pouring money into or pulling money out of a stock. Back-testing shows the Power Gauge would have issued bullish ratings for at least 8 of the top 10 stocks of the year every single year since 2016.

The offer is for Power Gauge Report, Chaikin’s flagship research service. It includes 1 full year of Power Gauge Report (monthly stock recommendations and analysis, valued at $499), 1 full year of access to the Power Gauge Rating system (look up any of 5,000+ stocks, valued at $1,000), and four bonus reports: 100X Starburst Opportunity ($199 value), The Robotics Hotlist ($299 value), The AI Energy Hotlist ($299 value), and The AI Bottleneck Hotlist ($299 value). A mystery gift valued at $2,999 is also included.

The total stated value is $2,595 plus the mystery gift, but the actual price is described as less than the price of a nice dinner and is revealed on the order page. The guarantee is 30 days, 100% money-back. If you cancel during the trial period, you keep all four bonus reports. The service has grown to over 550,000 subscribers.

Where to Learn More

For the complete analysis, read our CPGR Frontier AI review covering Marc Chaikin’s full thesis on frontier AI and the Power Gauge system.

Read our deep dive on Fabrinet, the hidden AI infrastructure play building the data superhighway.

Read our analysis of AI data centers for the full infrastructure thesis.

Ready to explore Marc Chaikin’s full research? Learn more about Power Gauge Report here.

This is not financial advice. Always do your own research before investing.