Unlock stock picks and a broker-level newsfeed that powers Wall Street. Upgrade Now
Better Buy in 2025: Nvidia Stock or Bitcoin?

In This Article:

Nvidia (NASDAQ: NVDA) and Bitcoin (CRYPTO: BTC) don't have much in common, but they have been two of the best-performing investments over the last five years, obliterating the S&P 500 (SNPINDEX: ^GSPC) index by several orders of magnitude:

NVDA Chart
NVDA data by YCharts

Had you taken $10,000 and split it equally between Nvidia stock and Bitcoin exactly five years ago, it would be worth a whopping $139,000 today. The same investment in the S&P 500 would be worth just $19,306.

A series of unique tailwinds could drive further upside in both Nvidia stock and Bitcoin this year, but I think one is a better buy than the other. Read on.

Nvidia's headquarters with a black Nvidia sign in the foreground.
Image source: Nvidia.

1. The case for Nvidia

Nvidia supplies the world's most powerful graphics processing units (GPUs) for data centers, which are used to develop artificial intelligence (AI) models. It recently started shipping its latest GB200 GPU, which is based on the new Blackwell architecture and can deliver up to 30 times more performance (in certain configurations) compared to the company's old flagship H100 chip.

But recent innovations in the AI space have triggered concerns that chip demand could collapse in the future. For instance, a China-based start-up called DeepSeek used a series of clever techniques to train its V3 and R1 AI models with a fraction of the computing power of its U.S. peers like OpenAI, and yet they deliver comparable performance in some benchmarks.

However, many of the industry's top developers are now moving away from traditional pre-training workloads, and allocating their computing capacity to test-time scaling to improve their models instead. This involves an AI model spending extra time "reasoning" or "thinking" during the inference phase before generating a response to a user's query, which means it makes better use of the knowledge it already has.

DeepSeek, OpenAI, Anthropic, Meta Platforms, and more are now focused on developing reasoning models that utilize test-time scaling. According to Nvidia CEO Jensen Huang, these models consume 100 times more compute than their predecessors, which means any drop in traditional training workloads will probably be offset by significantly higher inference workloads. As a result, Nvidia's GPUs should remain a hot product for years to come.

Nvidia recently reported its financial results for its fiscal year 2025 (ended Jan. 26), and it delivered a record $130.5 billion in revenue, which was a 114% increase from the prior year. Over $115 billion of that total came from the data center segment, which represented growth of 142%.