Skip to content
Investor: strategies, tips, and opportunities. — Investing — Finance — Money

On a Monday morning, shares of Nvidia, Broadcom, Microsoft, and various other Artificial Intelligence (AI) related companies experienced a decline.

Surprising outcomes from a budding AI company have led to a downturn in these stock prices.

,   and  Administrator
3 min read
A dissatisfied individual with sprawled arms fixated on a digital screen.
A dissatisfied individual with sprawled arms fixated on a digital screen.

The tech sector took a hit recently, with stocks plummeting, due to a Chinese start-up named DeepSeek and its new AI model, dubbed R1. This one-year-old company has managed to make waves in the tech world, as the R1 model climbed its way to become one of the top 10 AI systems globally. What's even more noteworthy is that this impressive feat was achieved using older-generation processors at a significantly lower cost, according to the company.

DeepSeek has shaken up the AI landscape by adopting a new approach to training its AI models. This technique, known as reinforcement learning or reward-driven optimization, involves refining strategies for problem-solving and trying different approaches to achieve desired results. This method is known for being more adept at learning and evolving.

One of the main challenges with AI has been its opaque nature, or what is often referred to as its "black box" nature. DeepSeek's R1 model removes this uncertainty by demonstrating its thought process, thus providing clarity and transparency. Venture capitalist Marc Andreessen even called DeepSeek R1 "one of the most amazing and impressive breakthroughs" he's ever witnessed.

However, it's important to note that DeepSeek's R1 currently trails the performance capability of models produced by OpenAI and Alphabet. But the fact that they were able to achieve such results with a smaller number of less powerful chips has raised concerns within the industry.

The Potential Implications

The news of DeepSeek's achievement sent stocks of tech companies like Nvidia, Broadcom, Microsoft, and Alphabet tumbling. Nvidia, a leader in providing the graphics processing units (GPUs) used for AI systems, has a 98% market share in data center GPUs, according to semiconductor analyst firm TechInsights. If AI models can be trained on lower-cost, inferior chips, Nvidia stands to lose significant business.

Broadcom, a provider of networking products used in data centers, could also face diminished demand for its high-end chips if the lower-cost option gains popularity. Microsoft, open-source champion, and Alphabet, both committed to developing next-generation AI models, also have much to lose if there are more cost-effective alternatives.

But, all may not be lost. Some analysts view the current market reaction as an overreaction and a golden buying opportunity. According to Wedbush tech analyst Dan Ives, US Global 2000 companies may not opt for a Chinese start-up for their AI infrastructure. Moreover, while prices seem high based on common metrics like the P/E ratio, they are much more reasonable when looking at forward earnings.

That being said, the development of generative AI is still in its early stages. The current disruption is simply a prime example of a knee-jerk reaction from some investors, and astute investors will wait for all the evidence to come in before making any hasty decisions.

Enrichment Data

DeepSeek's new model, R1, has significantly impacted the training and performance capabilities of AI models by adopting an efficient and cost-effective approach. Some key aspects of R1 include:

  1. Resource Efficiency: R1 utilizes a Mixture of Experts (MoE) framework, activating only 37 billion parameters out of 671 billion during inference, making it more resource-efficient than many other large models.
  2. Cost Efficiency: R1 is reportedly 20 to 50 times cheaper to use than OpenAI's comparable models, with operational expenses estimated at around 15%-50% of what users typically spend on OpenAI's o1 model.
  3. Performance and Accuracy: R1 matches the performance of leading AI models like OpenAI's GPT-4 and Google's Gemini in tasks such as logical inference, multilingual comprehension, and real-world reasoning.
  4. Multi-Stage Training and RL-Based Approach: DeepSeek R1 employs a multi-stage training process and an RL-first approach to foster advanced reasoning capabilities.

The implications for tech stocks like Nvidia, Broadcom, Microsoft, and Alphabet are significant and may include market disruption, reduced demand for AI infrastructure, and a shift towards democratized AI access.

DeepSeek's success in achieving top AI performance with lower-cost processors has sparked discussions about the potential impact on technology companies. With R1's resource and cost efficiency, rivals like Nvidia, Broadcom, Microsoft, and Alphabet might face pressure in the AI market if this cost-effective approach gains traction.

Investors are closely monitoring the situation, with some analysts viewing the market reaction as an overreaction and considering it as a buying opportunity, given the long-term prospects and potential return on investment in this emerging technology.

Read also:

    Related

    Latest