By now, you’ve undoubtedly heard of DeepSeek, China’s new large language model AI that’s sending shockwaves through the financial and tech sectors.

Touted as an alternative to OpenAI’s ChatGPT-4, DeepSeek claims to offer the same level of intelligence while being significantly cheaper and more energy-efficient.

However, many industry experts are skeptical of these claims…

In this article, we’ll break down what you need to know about DeepSeek, whether it really is the tech disruptor China claims, and how investors should view the situation.

DeepSeek claims several impressive figures—are they true?

DeepSeek’s developers came out of the gate with several bold claims, including:

• It costs 250 times less to operate than ChatGPT-4
• It uses one-tenth of the power compared to OpenAI’s model
• It was trained with only 2,048 GPUs in two months at a cost of $5.58 million (for context, ChatGPT-4 required 25,000 A100 GPUs, 100 days of training, and cost $100 million).

The news caused a massive selloff across the market, particularly in AI-related stocks.

• AI leader Nvidia (NVDA) dropped 17%, losing $600 billion in a single day
• The tech sector as a whole lost over $1 trillion in market cap
• AI-related energy stocks, data center REITs, and chipmakers also suffered declines, as DeepSeek’s claims suggested a lower need for expensive infrastructure.

But several factors suggest these claims might not be legitimate.

For one thing, there are reasons to believe DeepSeek might be a copy of ChatGPT-4.

Let’s take a closer look at the details…

Was DeepSeek built with stolen OpenAI data?

An investigative study by Daily Economic News has shown that if you ask DeepSeek, “Who are you?” it often replies with “I am ChatGPT-4.” And when asked for its API documentation, it provides OpenAI’s API manual.

What’s more, many of its responses, including jokes, were identical to those of ChatGPT-4.

(Interestingly, though, DeepSeek often struggles with logical problems that ChatGPT-4 easily solves.) 

Now, many are speculating that DeepSeek was built on stolen OpenAI tech.

And there’s a dark twist to this story…

Suchir Balaji was an OpenAI researcher responsible for training its language model. In August 2024, he resigned after raising concerns about the company’s technology violating U.S. copyright laws. (He also went to The New York Times about his concerns.)

Not long after, he traveled to China with some friends.

Then, in November, one day after celebrating his birthday, he was found dead in his apartment in San Francisco. His death was ruled a suicide… but his family isn’t buying that story, and a private investigator found inconsistencies in the case. 

There’s still a lot we don’t know about the situation… but suffice it to say, it’s become a red flag. The circumstances surrounding Balaji’s death are striking many as suspicious—especially considering the timing in relation to his trip to China. Even Elon Musk has called for a formal investigation into Balaji’s death.

Possible political motives and market manipulation

There’s another major red flag surrounding DeepSeek: China’s tight control of the narrative. 

For instance, after the launch, DeepSeek’s developers refused all external media interviews, saying they were “focusing on research.”

They also locked access to overseas users, requiring a Chinese-registered phone number to use the model. They claimed a “malicious virus” required them to restrict access, preventing international researchers from verifying their claims.

Meanwhile, the Chinese government prominently featured DeepSeek on national TV, declaring it a technological victory over the U.S. And it suppressed many videos that didn’t align with its messaging.

These factors are raising questions about whether China orchestrated DeepSeek’s announcement to disrupt U.S. markets and show it’s not falling behind in the AI race.

There could also be a motive related to market manipulation.

Billionaire hedge fund manager Bill Ackman posted on X:

The bottom line: DeepSeek’s lack of transparency, sudden media blackout, and restrictions on access suggest its tech is not as advanced as claimed. And the Chinese government’s involvement raises questions about propaganda and political motives rather than true technological breakthroughs.

Investment takeaways: Should investors buy the dip?

Regardless of whether the DeepSeek hype is indeed exaggerated, the situation could present a buying opportunity for Nvidia and AI infrastructure stocks.

If DeepSeek isn’t legitimate, the AI sector will rebound.

If it is legitimate, it will still rely on Nvidia’s chips and significant energy to operate.

What’s more, hyperscalers like Microsoft, Meta, and Amazon show no signs of slowing down their AI spending—which will continue to propel AI and energy demand higher.

Further investigations may reveal whether DeepSeek is stolen technology or a sophisticated AI leap. Make sure to tune into WSU Premium each week for the latest details as the story develops.