‘Hard to Imagine It Getting Much Cheaper’: Cramer Makes Bold Case for NVDA Right Now

Jim Cramer made a pointed call on air recently about NVIDIA (NASDAQ:NVDA | NVDA Price Prediction), and here is what he said.

“Look, there are a lot of people that just joined … This is their chance to buy Nvidia. It’s not going to get much cheaper. It’s hard to imagine it getting much cheaper,” Cramer said, addressing newer investors who’ve watched the stock from the sidelines.

So what’s his actual case? It comes down to two things: valuation and competitive positioning.

The Valuation Argument

Cramer cited Morgan Stanley’s view that NVDA trades at roughly 18 times 2027 earnings, which the firm called a “surprisingly good entry point.” For context, the broader market typically trades at 20 to 22 times forward earnings. If that framing holds, Nvidia is actually trading at a discount to the S&P 500 on a forward basis, despite being one of the most dominant growth companies on the planet.

The fundamentals back that up. Nvidia just reported Q4 fiscal 2026 revenue of $68.1 billion, beating estimates by a meaningful margin. That follows a revenue progression of $39.3 billion, $44.1 billion, $46.7 billion, and $57 billion in the four prior quarters. This is not a company decelerating.

The analyst community agrees. 94% of analysts covering NVDA are bullish, with a consensus price target of $263.39, compared to the current price of $182.05 as of March 2, 2026.

The Inference Chip Edge

Beyond valuation, the conversation touched on something that gets less attention: Nvidia’s competitive positioning in the inference market. The discussion noted a tie-up with Groq and highlighted that Nvidia’s inference chip benchmarks significantly faster than comparable offerings from Google and Amazon. Crucially, Nvidia also has a cost-competitive inference option, meaning they’re not just winning the high-end arms race. They’re competing across the full stack.

CFO Colette Kress put it plainly on the last earnings call: “Our inference demand is accelerating, driven by test time scaling and new reasoning models. Long-thinking reasoning AI can require 100x more compute per task compared to one-shot inferences.”

That’s the moat. Every time a reasoning model gets more sophisticated, it needs more compute. Nvidia built Blackwell specifically for this moment.

The Pullback Is Real

The stock is down about 5% over the past week and roughly 14% below its 52-week high of $212.18. Macro headwinds, including oil price drops and Fed rate uncertainty, have weighed on sentiment broadly.

Analysts who view the AI infrastructure buildout as a multi-year trend point to Nvidia as the essential infrastructure provider. The company has reported revenue growth of 73% year over year and has beaten estimates each quarter, while Morgan Stanley characterizes its valuation as below-market multiples on 2027 earnings.