If you bought Nvidia at the start of last year, you’ve basically spent the last twelve months on a rollercoaster that only goes up, except for that heart-stopping drop in April. People love to talk about bubbles. They’ve been screaming "dot-com crash" since the stock hit $100. Yet, here we are in early 2026, looking back at a year where NVDA managed to jump nearly 39%, soundly beating the S&P 500 for the third year in a row.
It wasn't a straight line. Honestly, 2025 was kind of a mess for the nerves. We saw the stock tank to a 52-week low of $86.62 in April when everyone freaked out about China export restrictions and a massive $5.5 billion charge. But then Jensen Huang did what he always does: he showed the world the receipts.
Why nvda stock prediction 2025 held up despite the skeptics
The big question everyone had was whether the "hyperscalers"—the Googles and Microsofts of the world—would keep spending billions on chips. They did. In fact, they spent more. By the time the third quarter of fiscal 2026 rolled around (which ended in October 2025), Nvidia was pulling in $57 billion in a single quarter. That’s not just a "good" result; it’s basically an entire industry's worth of revenue concentrated in one company.
The thing most people got wrong about the nvda stock prediction 2025 was the "DeepSeek" scare. Remember when that Chinese model supposedly proved you could do AI for cheap? Investors panicked. They thought GPU demand would evaporate. Instead, we saw a pivot to "reasoning models" like OpenAI’s o1, which actually require more compute power, not less. It’s the classic Jevons Paradox: the more efficient you make something, the more people use it.
💡 You might also like: Pfizer Stock Today Price: Why Everyone Is Watching the 25 Dollar Mark
The Blackwell and Rubin momentum
While everyone was staring at the stock price, the actual engineers were staring at Blackwell. Jensen described demand as "off the charts," and he wasn't exaggerating for the cameras.
- Blackwell Ramp: The GB200 units started shipping in volume mid-2025.
- Order Book: By late October, the company had $500 billion in booked orders for Blackwell and the upcoming Rubin architecture.
- Supply Constraints: Even with TSMC cranking out wafers, Nvidia was essentially sold out of cloud GPUs for the foreseeable future.
Price targets shifted wildly. At the start of 2025, Morgan Stanley was looking at $166. By the end of the year, Evercore ISI was pounding the table with a street-high target of $352 for the following cycle.
Revenue growth vs. valuation
Is it expensive? That’s the trillion-dollar question. When the stock was at its peak in October 2025, it hit an intraday high of $212.21, pushing the market cap past $5 trillion.
But here is the weird part: because the earnings grew so fast—jumping 67% in Q3—the forward price-to-earnings (P/E) ratio actually stayed somewhat reasonable. It sat around 24 to 25 times next year's expected earnings for much of the latter half of the year. For a company growing revenue at 60% plus, that’s actually "cheaper" than many slow-growing consumer software stocks.
The China problem and the tariff "poker game"
You can’t talk about the 2025 performance without mentioning the tariffs. The Trump administration’s trade policies kept everyone on edge. Nvidia ended up raising prices on gaming GPUs by about 10% and high-end AI chips by as much as 15% just to keep their 75% gross margins intact.
The market eventually swallowed the price hikes. Why? Because there is no second place. AMD’s MI300 and Intel’s Gaudi chips are fine, but they aren't Blackwell. Google’s TPU is the only real threat in the data center, and even then, most developers are locked into Nvidia's CUDA software. It’s a moat made of code, not just silicon.
Looking at the 2026 hand-off
As we move out of 2025, the focus is shifting to Rubin. This is the next-gen architecture scheduled for a late 2026 ramp. The "Rubin windfall" is what analysts are now pricing in. They are looking at fiscal 2027 earnings per share (EPS) estimates of $7.60. If you do the math on that, it’s easy to see why some think Nvidia could become a $6 trillion company before 2026 is over.
📖 Related: Where Do I Cash a Money Order From Western Union Without Getting Ripped Off?
Practical Insights for Your Portfolio:
- Watch the Hyperscaler Capex: If Microsoft or Meta suddenly say they are cutting AI spending in their quarterly calls, Nvidia will sneeze, and the whole market will catch a cold.
- Monitor the "Reasoning" Shift: As AI moves from just "predicting the next word" to "thinking through problems," the demand for H200 and Blackwell Ultra chips stays sticky.
- Don't Fear the Pullbacks: 2025 showed us that the 20% drops are usually driven by macro fears (tariffs, China, Fed rates) rather than a breakdown in Nvidia’s actual business.
- Mind the Valuation: While the P/E looks okay, a $5 trillion valuation requires perfection. Any delay in the Rubin roadmap could cause a massive correction.
The 2025 prediction wasn't about a bubble bursting; it was about a company proving it could outrun its own hype. Now, the goalposts have moved again.
Actionable Next Steps:
- Review your exposure: Ensure Nvidia doesn't represent more than 10-15% of your total portfolio to manage the volatility that 2025 proved is inevitable.
- Track quarterly data center revenue: Specifically, watch the "Blackwell Ultra" shipments in the first half of 2026 to see if the $500 billion order book is converting to actual cash flow.
- Set stop-losses near the $140-150 support levels: This was a key area of consolidation during the late 2025 rally.