A new AI disruptor has arrived, shaking up Silicon Valley’s dominance. DeepSeek’s low-budget, high-performance model not only challenges ChatGPT but also forces investors to question the billions poured into AI’s future.
A Cost-Effective Alternative to ChatGPT
On January 20, 2025, Chinese AI startup DeepSeek released its open-source reasoning model, DeepSeek-R1. This model, comprising 671 billion parameters, was trained using approximately 2,048 Nvidia H800 GPUs for $5.6 million and offered performance comparable to leading models like OpenAI's o1. Unlike ChatGPT, it is open source, allowing developers worldwide to access, modify, and build upon its models. It offers an API at significantly lower rates—$0.14 per million tokens compared to OpenAI's $7.5 per million.
Following its release, DeepSeek-R1's chatbot application quickly became the most downloaded free app on the US iOS App Store, surpassing OpenAI's ChatGPT, thereby challenging OpenAI's market dominance. If enterprises and developers begin favoring DeepSeek for its affordability and flexibility, OpenAI will need to adjust its pricing strategy or develop more advanced models to stay competitive and justify its pricing. Additionally, DeepSeek's open-source approach will likely pressure OpenAI to reconsider its proprietary model, potentially leading to greater accessibility and openness in its AI offerings.
DeepSeek used a more efficient architecture to come up with this model, which does not deploy full computing power at all times but instead activates only the portion relevant to the task at hand. A model like OpenAI’s has 1.8 trillion parameters, which are active simultaneously, whereas DeepSeek has 671 billion parameters, with only 37 billion active at any given time. This allowed them to train their model using significantly fewer GPUs. Also, GPT-4 generates text by predicting one token at a time, where each token depends on the previously generated ones. DeepSeek’s Multi-Token Prediction (MTP), however, predicts multiple tokens simultaneously, enabling faster and more efficient text generation.
Read more: Tech Industry Outlook 2025: What’s on the Horizon?
Initially, the reduced need for GPUs seems like a challenge for AI giant Nvidia, but it will ultimately benefit the company. If more companies adopt efficient AI training methods like DeepSeek, it will trigger a Jevons’ Paradox effect, democratizing AI. Models that once required billions of dollars to build can now be developed at a much lower cost. As AI becomes more affordable, small businesses that previously couldn't afford it will adopt it. This increased adoption will drive demand for more computing power, ultimately leading to greater demand for chips.
Investor Concerns Over AI Companies Funding and Valuations
The cost of $5.6 million, which is only a fraction of billions of dollars spent by leading US AI companies like OpenAI and Anthropic, raises doubt among investors about the enormous amounts being spent on computing power in the AI arms race. The launch sent ripples through the US market, with big tech stocks like Nvidia down nearly 17% on January 27, 2025, wiping off $600 billion in market capitalization.
Venture capitalists should be alarmed by the rapid advancements made by DeepSeek at a fraction of the cost incurred by its competitors. This raises significant concerns about the sustainability of the massive valuations and capital requirements that AI startups have commanded. In 2024, foundational model startups raised a staggering $40 billion, double the previous year. Investors who have funneled billions into LLM companies should question whether these bets will pay off, especially as lower-cost alternatives emerge, challenging the assumption that scale and spending alone drive AI dominance.
Alleged Use of OpenAI Data Sparks IP Concerns
OpenAI has raised concerns that DeepSeek has likely used its proprietary models to train its own open-source competitor through a technique called distillation. This process involves using outputs from advanced models, like GPT-4, to enhance the performance of smaller models at a much lower cost. While distillation is a widely accepted industry practice, the main concern is that it has been used in a way that violates its terms of service, which explicitly prohibits copying services or using OpenAI's outputs to create competing models.
Read more: AI in National Security: Protector or Threat?
US technical experts are investigating whether OpenAI's data was improperly accessed by individuals allegedly linked to DeepSeek. Microsoft's security researchers detected unauthorized data transfers involving individuals believed to be associated with DeepSeek. The researchers noted that a significant amount of data was extracted in the last quarter of 2024 through OpenAI's API, the primary method for software developers and businesses to access its services, raising concerns about potential breaches and intellectual property violations.
DeepSeek's rise signals a shift in the AI landscape, proving that innovation isn’t just about deep pockets but also efficiency and execution. But if the allegations against DeepSeek hold, it will likely spark legal battles and stricter regulations around AI training data. Regardless of the outcome, the case highlights the growing tensions in the global AI race, where intellectual property and competitive edge are increasingly under scrutiny.
Partner of choice for lower middle market-focused investment banks and private equity firms, SG Analytics, provides offshore analysts with support across the deal life cycle. Our complimentary access to a full back-office research ecosystem (database access, graphics team, sector & and domain experts, and technology-driven automation of tactical processes) positions our clients to win more deal mandates and execute these deals in the most efficient manner.
About SG Analytics
SG Analytics (SGA) is an industry-leading global data solutions firm providing data-centric research and contextual analytics services to its clients, including Fortune 500 companies, across BFSI, Technology, Media & Entertainment, and Healthcare sectors. Established in 2007, SG Analytics is a Great Place to Work® (GPTW) certified company with a team of over 1200 employees and a presence across the U.S.A., the UK, Switzerland, Poland, and India.
Apart from being recognized by reputed firms such as Gartner, Everest Group, and ISG, SGA has been featured in the elite Deloitte Technology Fast 50 India 2023 and APAC 2024 High Growth Companies by the Financial Times & Statista.
