The Chinese AI company DeepSeek recently gained attention for claiming to achieve significant cost and energy efficiencies in training advanced AI models. According to Barron’s, DeepSeek’s models rival those of OpenAI and Google, yet were trained for just $6 million using older Nvidia chips. This is a stark contrast with the more than $100 million cost to train OpenAI’s ChatGPT and future models expected to cost upwards of $1 billion.
While these claims raise questions about the future of AI development, several caveats remain, including reliance on stockpiled hardware and unverified R&D costs. For enterprises today, these developments represent both challenges and opportunities, underscoring the need for strategic guidance in navigating the rapidly evolving AI landscape.
What Is DeepSeek Claiming?
The contrast between DeepSeek’s cost-effective approach and the large-scale investments by U.S.-based AI companies underscores a fundamental shift in how AI is likely to evolve. But DeepSeek’s reported claims are not without limitations. As noted by Barron’s, DeepSeek’s figures do not account for research and development expenses, which are a significant part of AI innovation. Additionally, DeepSeek’s reliance on older Nvidia chips highlights a key vulnerability, as U.S. export restrictions on advanced semiconductors to China present a growing challenge.
DeepSeek’s reported ability to train models on older hardware at significantly lower costs has the potential to reshape how AI is developed and deployed. If these efficiencies are proven to be scalable and sustainable, they could accelerate the adoption of AI technologies across a variety of industries. However, the lack of verified data and the exclusion of R&D costs from their reported figures make it difficult to determine whether this model can be broadly applied. DeepSeek’s reliance on non-sustainable hardware strategies adds another layer of uncertainty to their claims.
Key Considerations for the AI Market
Realistically speaking, Western enterprises are unlikely to work directly with DeepSeek. Security concerns, compliance requirements and geopolitical tensions related to the company’s ties to the Chinese government present significant barriers to adoption in markets such as the U.S. and Europe.
A more likely scenario is that U.S. hyperscalers and other providers will incorporate similar cost and energy-saving innovations into their own platforms. This would allow enterprises to benefit from new efficiencies without compromising on trust, compliance or security.
The broader market dynamics are also evolving rapidly. As Barron’s has observed, technological advancements in efficiency often lead to increased demand, a phenomenon known as the Jevons Paradox. This dynamic suggests that, as AI becomes more affordable and accessible, its adoption across industries accelerates. This trend is further reinforced by the rise of industry-specific solutions and the growing role of small and large language models, which are increasingly tailored to meet unique business needs. These developments are likely to expand the scope of AI applications, particularly in sectors such as Manufacturing, Life Sciences and Finance.
What Enterprises Should Do Now
DeepSeek’s advancements serve as a clear example of the rapid changes shaping the AI industry. As AI continues to evolve, organizations will need a deeper understanding of emerging solutions, market dynamics and the competitive landscape of service providers.
One of the most significant implications of declining AI training costs is the potential for new use cases that were previously considered too expensive or impractical. This shift likely means enterprises will need to find third-party providers that are best suited to help them implement solutions that align with their strategic goals. While lower costs may democratize access to AI, the complexity of the AI stack remains a challenge, and enterprises will need expert guidance to ensure successful adoption.
The integration of energy-efficient AI capabilities into hyperscaler platforms will enable faster time-to-market and improved return on investment, but enterprises will need to carefully balance innovation with operational efficiency. Only by evaluating the trade-offs between cost savings and long-term scalability and security will enterprises be able to drive measurable business outcomes. Industries such as Healthcare, Retail and Supply Chain Management are particularly well-positioned to benefit from these advancements, as they increasingly adopt tailored AI solutions to address their specific challenges.
Now is the time for enterprises to build a comprehensive AI roadmap that incorporates cost-efficiency trends and aligns with emerging market dynamics. It will be increasingly important for organizations to select providers that align with their regulatory and data sovereignty requirements, particularly as global AI supply chains become more complex.
ISG helps organizations navigate the rapidly changing AI market to capitalize on emerging opportunities while addressing the complexities of AI adoption. Contact us to find out how we can help.
About the author
Steve Hall
As the leader of ISG’s business in EMEA and an Executive Board Member, Steve provides strategic insight and advice to help ISG’s clients solve their most critical business challenges, helping them adopt and optimize the technology and operating models they need to compete successfully. In particular, he uses his long experience and broad expertise to challenge and inspire them to think about their risks and opportunities in new and unexpected ways.
Past achievements for clients
Steve leads his team’s engagement with clients with an industry-recognized and highly valued perspective on the most important trends in business and technology. He asks and answers the big questions: Why do you need to transform? What’s your best way forward? What do you need to accelerate? And where should you invest your technology dollars to make it all happen?
Among his many client success stories, his ability to take in the big picture, define the problem and connect the dots to the right solutions helped one legacy postal and shipping giant transform itself into a modern logistics powerhouse. He also guided a global energy industry leader through a complex operating model and IT provider transition, helping them see past the obvious cost cutting measures to identify the root causes of their challenges—and delivering savings far beyond what they had imagined.
