The artificial intelligence boom has dominated market narratives for nearly three years, minting trillion‑dollar giants, reshaping corporate spending priorities, and fueling one of the most powerful thematic rallies since the early days of cloud computing. But according to renowned short‑seller Danny Moses, the excitement now carries a familiar echo—one that reminds him of the dot‑com era.
Moses, best known for his role in the 2008 housing‑market short chronicled in The Big Short, told Business Insider that the AI market is showing unmistakable signs of overheating. While he stresses that AI is a legitimate long‑term growth story, he believes the current pace of investment and valuation expansion is pushing the sector toward a mathematical breaking point.
"Growth is real," Moses said, "but the mathematical models no longer hold. I think we’re reaching a tipping point where the math starts to fall apart."
A Bubble—But Not a Short Call
Moses is careful to distinguish between identifying a bubble and betting against it. He is not calling for investors to short the entire AI ecosystem. Instead, he argues that the sector is entering a phase where selectivity matters more than ever.
In his view, the AI boom resembles the late‑stage dot‑com cycle: real technological progress paired with unrealistic expectations, aggressive capital spending, and a widening gap between winners and everyone else.
The message is straightforward: AI is not a monolith, and investors who treat it as one risk getting burned.
Focus on the Giants, Not the Hopefuls
Moses believes the safest way to participate in AI’s long‑term growth is through the largest, most financially resilient tech companies—firms with the balance sheets and cash flow to fund multi‑year AI investments without jeopardizing their core businesses.
He points to Amazon, Google, Meta, and Microsoft as prime examples. These companies can adjust capital expenditures on the fly, absorb short‑term volatility, and still generate billions in free cash flow.
"These companies can adjust capital expenditures at any time and still maintain positive cash flow," Moses said. "In contrast, other companies relying on AI investments must continuously spend."
In other words, the mega‑caps can afford to play the long game. Smaller firms—or even mid‑sized tech names—may not have that luxury.
The Riskier Side of the AI Trade
Moses is far less optimistic about companies that have become AI darlings despite fragile financial foundations. He specifically highlighted Oracle, noting its high debt load and the substantial cash required to fulfill AI‑related infrastructure orders.
Oracle has leaned heavily into cloud and AI infrastructure spending, but Moses argues that the company’s balance sheet leaves little room for error. If AI demand slows or capital expenditures rise faster than expected, the financial strain could become more visible.
He also flagged Supermicro and CoreWeave—two of the most volatile AI‑linked names—as examples of companies where investor enthusiasm may be outpacing fundamentals. Supermicro’s meteoric rise has been fueled by demand for AI servers, while CoreWeave has expanded aggressively through debt‑financed data‑center buildouts.
These companies, Moses warns, represent the speculative edge of the AI trade.
Investors Are Finally Separating Winners and Losers
Despite his concerns, Moses sees a positive development: investors are beginning to differentiate between strong and weak AI plays. The market is no longer rewarding every company that mentions "AI" in an earnings call. Instead, it is gravitating toward firms with durable balance sheets, consistent profitability, and clear competitive advantages.
"I think this proves that investors are starting to distinguish between winners and losers," Moses said. "They are more willing to rely on companies with stronger financials to participate in AI‑related projects."
This shift marks a departure from the early stages of the AI boom, when nearly every stock with AI exposure saw outsized gains. Now, the market is rewarding discipline—and punishing overreach.
Uranium: An Unlikely AI Infrastructure Play
In a surprising twist, Moses is also bullish on uranium, arguing that it will become a critical component of the AI industry’s long‑term infrastructure. As AI workloads grow exponentially, so does the demand for stable, large‑scale energy sources. Nuclear power, he believes, will play a central role in meeting that demand.
Uranium, therefore, becomes a long‑duration investment tied indirectly to AI’s expansion.
But Moses cautions that the uranium theme requires patience. The timeline for nuclear infrastructure buildout is long, regulatory hurdles are significant, and the payoff may not materialize for years.
Investors, he warns, often underestimate the lag between AI‑driven demand and the infrastructure required to support it.
A Market at a Crossroads
Moses’s analysis arrives at a moment when AI stocks are experiencing heightened volatility. After a blistering rally in 2023–2024, the sector has faced growing skepticism about valuations, capital‑expenditure cycles, and the sustainability of demand.
His comments echo a broader shift in sentiment: the AI story is still compelling, but the easy money phase may be over.
The next chapter, he argues, will be defined by:
- Balance‑sheet strength
- Cash‑flow durability
- Capital‑expenditure flexibility
- Clear competitive moats
- Real, not theoretical, AI monetization
Companies that check these boxes will continue to thrive. Those that don’t may struggle as the market becomes more discerning.
Conclusion: A Bubble Worth Navigating Carefully
Danny Moses is not predicting an AI crash. Instead, he’s urging investors to approach the sector with discipline. The AI boom is real, but so are the risks. The market is transitioning from a phase driven by hype to one driven by fundamentals—and that shift will separate the durable winners from the speculative pretenders.
For investors, the takeaway is clear: AI remains a powerful long‑term theme, but selectivity is no longer optional.