OpenAI CEO Sam Altman said that the ChatGPT maker will make much better AI (artificial intelligence) models than China-made DeepSeek’s R1 model. China’s DeepSeek took the world by storm when it launched the cheaper generative AI model (V3) last week and said that it used Nvidia’s (NVDA) older hardware to run the model. The news sent Nvidia’s stock price spiraling down nearly 17%, making it lose close to $600 billion. This marked the largest one-day share price drop in the stock of any company on Wall Street.
Invest with Confidence:
- Follow TipRanks' Top Wall Street Analysts to uncover their success rate and average return.
- Join thousands of data-driven investors – Build your Smart Portfolio for personalized insights.
DeepSeek’s V3 AI model needed less than $6 million in computing power and uses Nvidia’s H800 chips to run the model. The Chinese AI startup’s DeepSeek-R1 model is supposedly 20 to 50 times cheaper than OpenAI’s o1 model, the company said on DeepSeek’s WeChat post.
Altman Says OpenAI Models are Better Than DeepSeek
Yesterday, Altman replied to a post on social media platform X, where he was asked about the competition from the Chinese AI startup and what is the further course for OpenAI. The CEO acknowledged that DeepSeek’s R1 is commendable at the price point that it is offering. However, his comments denoted a sense of superiority, since Altman believes that ChatGPT models are far better and it is fun to have a new competitor on board.
Furthermore, Altman said that they will pull up some new model releases in competition and will continue to research newer applications. Also, the CEO said that superior generative AI models are going to require more computing power. The world is going to use a lot of AI in the future in every task, and thus, more compute and enhanced models will be required.
The Search for Artificial General Intelligence
OpenAI is on a mission to launch an AGI model, which stands for artificial general intelligence model. These models are expected to perform cognitive tasks just like or even better than humans in some cases. AGI models are still in the development phase. OpenAI was the first to talk about such models or even think of them. AGI models are expected to perform with more human-like intelligence, with the ability to self-teach and improve with each learning. The current generation of AI models can only perform those tasks that they are trained to perform.
DeepSeek’s R1 Mocks U.S. AI Leadership
DeepSeek’s R1 model is capable of performing nearly the same tasks as most of the America-made AI models, that too at a cheaper cost and with older hardware. Its launch is sort of mocking the U.S. AI leadership. Many American companies are investing billions of dollars in AI based on the notion that enhanced computing power and advanced data center infrastructure are required to train larger-scale AI models. Nvidia CEO Jensen Huang also commented on DeepSeek yesterday, saying that DeepSeek models will still require a lot of its GPUs (graphic processing units) for training.
Importantly, R1’s launch puts into doubt the ambitious “Stargate” project unveiled by newly instated U.S. President Donald Trump last week. OpenAI, along with other big tech firms such as Oracle (ORCL), Softbank (SFTBY), Nvidia, and more, have pledged to invest $500 billion in the project in a span of four years. The money will be used to develop advanced data centers and improve the AI infrastructure in the country. The U.S. administration wants to become the superpower in the AI realm and has implemented a strict export ban on advanced AI chips to China, fearing its use in military tactics.
What Are the Best AI Stocks to Buy?
For now, all eyes are on DeepSeek and its future. DeepSeek R1 model’s launch has surely sent a shockwave on Wall Street, dragging down all major U.S. tech stocks.
We leveraged the TipRanks Stock Comparison Tool for Best Artificial Intelligence Stocks to see which one performs better compared to rivals. Investors can choose to invest in these stocks after thorough research.
