ChatGPT took the world by storm when it launched in November of 2022. By now, readers likely know this technology and what it is capable of. On almost any metric, the launch of ChatGPT was a huge success.

In less than two months, the site reached 100 million users. The fastest growth any website had ever experienced. For comparison, it took TikTok nine months to hit 100 million users. It took Instagram two and a half years.

AI quickly became an undeniable trend set to reshape the global economy.

And there was no bigger beneficiary than Nvidia (NVDA). Nvidia GPUs, along with its proprietary software and networking cables, are the preferred method to train Large Language Models (LLMs) that are behind popular chatbots like ChatGPT.

Microsoft and OpenAI reportedly spent hundreds of millions of dollars to buy thousands of Nvidia’s GPUs to build ChatGPT ahead of launch.

You already know this. But what you might not know was the power struggle happening behind the scenes.

In September of last year, the U.S. government asked Nvidia to stop selling its A100 and H100 GPU chips to Chinese companies.

By October, the U.S. government outright banned Nvidia from selling its top AI chips to China.

Nvidia estimated that it’ll lose $400 million in quarterly chip sales to China. That’s about 6% of its total sales.

The U.S. government has cut off China’s access to the entire supply chain to manufacture AI chips.

That includes:

  • Chip design software

  • AMD’s MI250 AI chip

  • Equipment capable of making advanced AI chips being sold to China

  • Chip manufacturers, like TSMC, which are banned from accepting orders from China-based companies

It begs the question: Why is the U.S. government so eager to slow China’s development of artificial intelligence?

The answer will have a huge impact on how the AI trend plays out over the next several years. And as investors, we need to be paying attention.

A Matter of National Security

Governments have always placed limitations on the exportation of technology deemed to be vital for national security. Typically, this meant military technology. What’s interesting is that it is now bleeding over into advanced tech like artificial intelligence.

The White House doesn’t want China to have any alternatives or be able to reverse engineer this technology. That’s why the Nvidia bans went into effect late last year.

Government sanctions show the potential behind AI.

This isn’t just about getting better search results or using chatbots to answer customers’ questions.

AI tech has advanced to the point that it’s now a part of the national defense strategy.

Even without selling to China, companies like Nvidia, AMD, and TSMC will see plenty of demand from U.S. and non-Chinese tech companies.

AMD’s CEO Lisa Su sees demand for AI chips growing from $30 billion this year to $150 billion by 2027.

That hardware will support thousands of companies adopting AI over the coming years.

The consulting company McKinsey estimates that generative AI will deliver $2.6 trillion to $4.4 trillion in value.

Banning sales to China will make it easier for U.S. companies to get these chips and adopt AI even faster.

The fact that governments are taking such an active role in the development of this technology shows us what’s at stake. As I said, this isn’t just about moderately better search results. It is being viewed in existential terms. And whichever country becomes the leader in artificial intelligence will have an economic – and perhaps military – advantage for decades.

What should investors do with this information?

The reality is that even with sanctions, companies like Nvidia will have plenty of business in the years ahead. But that doesn’t mean Nvidia is my favorite artificial intelligence stock to own right now.

I’ve identified AMD as the best opportunity among the chipmakers at the moment.

Shares of AMD are up 77% this year. That’s a fraction of Nvidia’s 192% rally. And looking at the valuation, AMD trades for a relative bargain.

As I wrote in a previous essay (catch up here), shares of NVDA trade at 40X current sales. That’s the most expensive the company has ever been on a valuation basis. Perhaps the stock can “grow into” that valuation. But for my money, I’d be a buyer at AMD right now.

AMD’s MI300 will be released later this year. Production will take a few months to ramp up. But it will be even more powerful than Nvidia’s A100 chip series.

Wall Street analysts are still rating AMD like it will struggle to sell the MI300 with Chinese buyers on the sideline. Perhaps that’s why the company trades for only seven times sales, well below its all-time high of 14X.

But my research suggests that corporate AI adoption will be more rapid than most realize. AI developers want an alternative to Nvidia, and AMD has it.

That will be a huge win for AMD.

To read more about the AI opportunity with AMD and two other stocks, click here.

Regards,

Colin Tedards
Editor, The Bleeding Edge