'Lmao', 'Obviously' To Emoji - How Elon Musk Mocked DeepSeek After It Beat ChatGPT
Tesla CEO Elon Musk has voiced doubts about DeepSeek’ app's much touted capabilities that beat ChatGPT on Apple App Store and made Nvidia and other tech stocks bleed.

Tesla CEO Elon Musk has voiced serious doubt on X (formerly Twitter) about the claims made by DeepSeek, a Chinese artificial intelligence startup that has made headlines for beating OpenAI's ChatGPT and turning into the most downloaded free app in the US on the Apple App Store.
The “new kid on the AI block” has knocked the tech giants of Silicon Valley to their knees. Notably, Wall Street indexes plunged over the past few days after a major selloff of tech stocks following the seemingly unstoppable rise of DeepSeek’s AI programme. Nvidia Corp. and ASML Holding NV plummeted with the former witnessing its market cap losing over $500 billion in value on Monday.
Even as the stock markets bled, Musk, the owner Grok AI, remained unimpressed and refuted DeepSeek’s claims.
Musk responded to a post on X by a user named Chubby, which showed a video of billionaire and Scale AICEO Alexandr Wang saying “DeepSeek has about 50,000 NVIDIA H100s that they can't talk about because of the US export controls that are in place.” Musk responded to the post with a single word “Obviously.” This was still a tame reaction, but he cranked up the condemnation further in his other reactions even taking to mocking DeepSeek.
When Salesforce CEO Marc Benioff implied that DeepSeek’s recent success showed that AI development is not exclusively dependent on costly GPUs, a user responded about whether he was taking DeepSeek claims about creating this hugely successful app on a shoestring budget literally. Musk reacted dismissively with a “Lmao no”.
DeepSeek's AI Assistant, which is based on the DeepSeek-V3 model, has experienced exceptional growth since its Jan. 10 release. According to DeepSeek creators, the DeepSeek-V3 model outperforms several open-source models and can compete with some of the most advanced closed-source models available.
DeepSeek researchers have also shockingly claimed that the DeepSeek-V3 model was trained at a cost of less than $6 million utilising Nvidia’s H800 CPUs. DeepSeek uses open-source technology and less expensive chips, eliminating the need for costly hardware that is subject to US export restrictions.
Musk even mocked the startup as being a product of a China-based laboratory, with subtle references to how Coronavirus also emerged from a Chinese lab. In a reaction on X, Musk responded with a “face with tears of joy” emoji.