DeepSeek Is Destroying, Disrupting, And Dictating US AI Dreams With New Developments

A day ago, we covered the initial reactions of users, businesses, and experts to the newly found capabilities of DeepSeek’s R1 generative artificial intelligence (GenAI) model.

The move saw investors question if artificial intelligence (AI) companies really needed to spend the copious amounts they’ve been claiming to spend to advance AI. DeepSeek’s V3 model was made for just about $5.6 million, as compared to the millions and billions spent by United States-based AI companies.

Moreover, DeepSeek built its AI assistant in just around two months—faster than competitors—using lower-capacity NVIDIA H800 chips. Also, its tool is available for free.

This question was also reflected in the stock market, as NVIDIA’s stocks fell by over 17%, resulting in a wipeout of around $600 billion from its market capitalization at its peak, marking the biggest market value drop in U.S. stock market history.

At the same time, the Nasdaq fell by over 3%, dropping over $1 trillion off the index of technology companies at one point.

It also affected almost every other technology company including Broadcom (17%), Microsoft (2.5%), Alphabet (4.2%), and Tesla (3.5%). Even Japanese and Australian share markets were affected.

After the mauling of the first day, the second day saw tech company shares steadying.


The interest gathered the attention of OpenAI CEO Sam Altman also—obviously—who said in an X post, “DeepSeek's R1 is an impressive model, particularly around what they're able to deliver for the price. We will obviously deliver much better models and also it's legit invigorating to have a new competitor! We will pull up some releases.”

Altman added, “But mostly we are excited to continue to execute on our research roadmap and believe more compute is more important now than ever before to succeed at our mission. The world is going to want to use a LOT of AI, and really be quite amazed by the next-gen models coming.”

NVIDIA noted that DeepSeek’s success reiterated the need for more advanced AI chips, and also said, “DeepSeek is an excellent AI advancement and a perfect example of Test Time Scaling. DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely-available models and compute that is fully export control compliant.”

On the other hand, users migrated to the new tool so rapidly that DeepSeek dethroned ChatGPT on Apple’s App Store to clinch pole position for free applications.

However, this wasn’t just observed in the U.S. chart, DeepSeek also grabbed the top position in App Stores in China, India, and 51 other countries.

Yes, 51.

It also features in the top 10 free apps lists in 111 countries and in 18 countries on Google’s Play Store, according to Appfigures.

However, all this attention came at a price for DeepSeek.

The sudden growth of popularity led to outages and a large-scale malicious attack on its services that eventually saw the company temporarily limit registrations. Only users who were already registered were provided the option to log in.

Interestingly, these outages were the company’s longest in around three months.

In this regard, DeepSeek’s sign-in page blared a message that read, “Due to large-scale malicious attacks on DeepSeek's services, registration may be busy. Please wait and try again. Registered users can log in normally. Thank you for your understanding and support.”


Ahead of this, DeepSeek drew concerns related to its ties to China.

One of these includes the heaps of user data it collects, including chat messages, which could potentially be sent back to its home country. As the company doesn’t have a press contact executive, it’s unclear what its data privacy initiatives are.

On the other hand, the AI model has also been observed to censor its replies when questioned about China, its President Xi Jinping, or the country’s policies. This also extends to topics commonly censored by the Chinese government, which includes government criticism, human rights violations, and more.

At least this characteristic offers a better reason for its existence, especially when compared to the one where DeepSeek’s V3 model thinks it's ChatGPT.

As per some users on X, when the AI assistant is asked which model it is, its response says it’s ChatGPT, an LLM developed by OpenAI.

Furthermore, questions about its API provide instructions on how to use OpenAI’s API.

Mike Cook, a research fellow at King’s College London specializing in AI, explained, “Obviously, the model is seeing raw responses from ChatGPT at some point, but it’s not clear where that is. It could be ‘accidental’ … but unfortunately, we have seen instances of people directly training their models on the outputs of other models to try and piggyback off their knowledge.”


Despite these concerns, one can’t deny DeepSeek made a big impact on the AI industry, and as if that wasn’t enough, the company released a new AI model—Janus-Pro.

Janus-Pro is an image generator (available for download) that’s reported to be able to outperform Stability AI's Stable Diffusion and OpenAI’s DALL-E 3. It comes as an upgrade to the previously released Janus, which was launched late last year, around the same time the AI company revealed its DeepSeek-V3 model.

Currently, Janus-Pro's input resolution is limited to 384x384 resolution, however, its largest model—Janus-Pro-7b—could beat the two abovementioned image generating models on two AI benchmark tests.

OpenAI and Stability AI didn’t provide a comment on the matter.

Do you think DeepSeek will become the new leader of the AI industry? Do you think OpenAI will be able to challenge DeepSeek’s newly acquired pole position?

Let us know in the comments below!