After DeepSeek's R1 model was released, it ignited a panic among stocks associated with artificial intelligence (AI). The Chinese startup reportedly trained its model for only $5.6 million, making many question why U.S. companies are pouring billions of dollars into computing power.
This caused Nvidia's (NVDA 1.04%) stock to fall by nearly 20%. Nvidia is the primary beneficiary of AI spending in the U.S., as most AI models are trained on its hardware. In fact, DeepSeek's R1 model was trained on Nvidia's H800 chips (the more powerful H100 chips aren't available in China due to export bans).
I see this drop as a buying opportunity, not a time to sell. There's a specific behavior that occurs after a breakthrough in efficiency known as the Jevons paradox, and I think all investors should familiarize themselves with it.
What is the Jevons paradox?
The Jevons paradox was developed in 1865 by William Jevons, an economist from England. He observed that as a resource becomes more efficient thanks to technological advancements, the consumption of the resource doesn't drop. Instead, use rises because the cost has fallen enough to make it more accessible.
NASDAQ: NVDA
Key Data Points
This paradox was originally observed in coal usage during the Industrial Revolution but can also be seen in other areas, like fuel consumption in vehicles. If your car could instantly get 10 times better gas mileage, would you use it the same amount, or would you take it on more road trips because it's now cheaper to do so?
This same mindset is currently being applied to the AI arms race, as DeepSeek's innovative breakthrough to train models more efficiently could increase the computing power needed. As models become more efficient, they will become more cost-effective for consumers and businesses alike. This will drive further demand for the product as the price falls, requiring more AI computing power.
This is a counterargument to the knee-jerk sell-off that occurred after DeepSeek's R1 model gained popularity. I think it's a valid point and can be used to scoop up Nvidia shares for a much cheaper price tag.
How cheap is Nvidia stock now?
Despite the sell-off, Nvidia's stock still doesn't trade for bargain bin prices. As of the time of writing, Nvidia trades for 51 times trailing earnings and 44 times forward earnings.
NVDA PE Ratio data by YCharts
Those aren't historically cheap prices for most stocks, but most stocks aren't growing at the same pace Nvidia is. During the third quarter of Nvidia's fiscal 2025 (ended Oct. 27, 2024), its revenue rose 94%, which helps justify its price tag.
Management expects fourth-quarter revenue to be around $37.5 billion, indicating 70% revenue growth. Nvidia's quarter just ended this week, so the effect (if any) of DeepSeek's R1 model won't be known until Q1 of fiscal 2026.
We'll likely get a Q4 earnings announcement for Nvidia in late February when management will undoubtedly field questions about the strength of its business looking forward. Currently, Wall Street analysts still have strong growth in mind, with revenue expected to rise 52% in fiscal 2026. I think this is still a fair assumption, as I don't think the overall AI investing landscape will be affected by DeepSeek's breakthrough.
Despite some market scares from DeepSeek's new AI model, I still believe that Nvidia has a strong investment thesis, as its GPUs are training models worldwide. Nvidia has a dominant market position, and its strength should allow it to continue its success well past 2025.
