The AI Bubble Has Already Popped — Five AI Stocks at Post-Bubble Valuations

TL;DR Nvidia trades at a P/E of 21 (67% below its 5-year average of 64), AppLovin at 25 (72% below 90), Zscaler at a P/S of 6.75 (60% discount), Oracle's fair value sits at $229 (56% upside), and Broadcom's at $443 (41% upside). Surging earnings combined with falling prices have pushed five AI stocks into post-bubble valuations.

Nvidia up 1,200%. Micron up 300%.

Those are the five-year gains for AI stocks. It is natural to draw parallels with the 2000 dot-com bubble, when the NASDAQ crashed 75% and $5 trillion in market value evaporated overnight. The fear that AI stocks will follow the same path dominates investor sentiment right now.

But the numbers tell a completely different story.

The Core Thesis — This Is Post-Bubble, Not Pre-Crash

AI leaders have indeed fallen 12 to 20% from their peaks. Yet earnings are surging and forecasts keep getting revised upward. Valuation multiples — the price investors pay relative to earnings — have dropped dramatically.

My analysis points to a clear conclusion: these five stocks are already trading at post-bubble prices.

1. Nvidia (NVDA) — P/E of 21, a 67% Discount to History

Down 20% from its 52-week high of $212.

The four major hyperscalers — Meta, Alphabet, Amazon, and Microsoft — are spending $115 billion per quarter on AI capex. Every quarter. Total AI spending is projected to exceed $700 billion this year, and the primary beneficiary is Nvidia's GPU lineup.

Revenue forecast for this year: $370 billion (up 71%). EPS: $8.29.

Divide the $4.3 trillion market cap by expected revenue and you get a P/S of 11.7 — less than half the five-year average of 24. The P/E of 21 is one-third of the five-year average of 64.

A P/E of 21 for a company growing earnings at 73% annually is the opposite of bubble territory.

Applying the five-year average P/S yields a market cap of $8.88 trillion, or $363 per share. The average P/E gives $530 per share — 200% upside from here.

2. AppLovin (APP) — Trading at a 63% Discount to Last Year

Down 48% from its peak on fears that AI will replace software providers.

Some long-term risk exists, but disruption does not happen overnight. The company is still expected to deliver $8 billion in revenue (up 46%) and $15.29 in EPS this year.

At a $130 billion market cap, the P/S of 16 exceeds the five-year average. But it represents a 63% discount from where this stock traded just last year. On a P/E basis, the stock trades at 25 — against a four-year average of 90.

Even a recovery to just 50 times earnings would double the stock.

3. Zscaler (ZS) — Hardest-Hit Cybersecurity Name, P/S at a 60% Discount

Down 59% from its $337 peak. The deepest casualty among cybersecurity stocks.

The culprit is fear that AI will replace cybersecurity vendors. Reality points in the opposite direction. Among 100 large companies surveyed, 77% are using AI agents, but only 4% have secured them. AI is expanding the attack surface, not shrinking it.

Revenue forecast: $3.3 billion (up 24%). EPS: $4.

At a $22 billion market cap, the P/S sits at 6.75 — a 60% discount to the five-year average. A return to just 12 times sales would take shares to $246, up 78%.

4. Oracle (ORCL) — the Market's Contradiction Creates an Opportunity

Oracle has become the poster child for the bubble burst. Shares surged 40% in a single day to $345 last year, only to give it all back and then some. Now down 58%, essentially flat for the year.

Investors worry about funding the cloud infrastructure buildout and whether promised revenue from OpenAI and other AI customers will materialize. But consider the contradiction: the market simultaneously believes AI is powerful enough to replace software and jobs, yet not compelling enough to attract the spending needed to build it out. Both fears cannot be true at the same time.

Revenue forecast: $67 billion. EPS: $7.45.

The five-year average P/S of 6.5 yields a $435 billion market cap, only 3% above current levels. But next year's revenue growth is projected at 30%. On a P/E basis, the five-year average of 30 times $7.45 gives a fair value of $229 per share — 56% upside. And that is before growth truly ramps.

5. Broadcom (AVGO) — the Strongest Defense in AI

Down just 24% from its peak — the smallest decline among the five, a testament to its entrenched industry position.

Revenue is forecasted at $105 billion this year (up 64%), driven by accelerator chips and networking. EPS dips slightly to $1.32 as R&D spending ramps, but is expected to jump 57% next year.

The five-year average P/S gives a market cap of $1.1 trillion and $250 per share. But that average itself is 57% below what investors paid last year. A more realistic P/S of 20 against $105 billion in revenue yields a $2.1 trillion market cap, or roughly $443 per share — 41% upside.

The Bottom Line

All five stocks share the same pattern: earnings surging, share prices falling, valuations 51 to 67% below historical averages. The "AI bubble" narrative does not match the actual numbers.

Further near-term weakness over the next month or two is possible. But the data makes one thing clear — these prices are post-bubble, not pre-crash.