In the first installment of our Artificial Intelligence & Real Estate Series, published last December, we introduced a framework to analyze the impact of the rising adoption of AI technologies on the commercial real estate environment. We also warned that as the AI revolution remains in its infancy, investors should brace for unanticipated breakthrough advancements that are difficult to predict and precisely time. Indeed, over the past month, a pivotal leap took place, intensifying the competition for AI industry leadership and market dominance.

The shifting economics of AI development and deployment

On the 20th of January, a little-known Chinese AI start-up, DeepSeek, released a free and open-source generative AI chatbot able to rival the US industry giants. The release went unnoticed until a few days later when more details about its large language model (LLM) became widespread.

LLMs are general-purpose models trained to understand human language and perform tasks such as generating text or answering questions. These serve as the basis for various AI applications, including, but not limited to chatbots, search engines, and document summarization. Developing a state-of-the-art LLM system typically requires significant capital expenditures and ongoing operational costs. As a result, this field is dominated by a handful of tech giants—such as Alphabet, Amazon, Meta, and Microsoft—investing billions in data center infrastructure, training proprietary AI models, and reinforcing their industry leadership.

Against this backdrop, the rise of DeepSeek may be viewed as a black swan event, or at the very least a watershed moment in the current AI landscape. Indeed, compared to the previous LLMs, DeepSeek claimed to build its application at a small fraction of the cost (measured in millions of dollars rather than billions) compared to other leading LLMs, using far less computation power and fewer chips. Although questions still arise on the validity of these claims, the achievement has cast doubts on several economic and technical assumptions taken for granted when developing AI models: massive scale, billions of dollars in capital expenditures, substantial power requirements, and costly NVIDIA graphics processing units (GPUs). As a result, these requirements may no longer be essential, lowering entry barriers for new competitors and signaling a profound shift in the economics driving future AI advancements.

DeepSeek innovation could represent a significant milestone of the current AI paradigm. However, it should not be seen in isolation but as part of a broader and expanding ecosystem of businesses competing to push the technology frontier further and gain market shares. Indeed, in the weeks following DeepSeek’s announcement, several companies, including OpenAI, ByteDance, Alibaba, Minstral, and Moonshot, released new versions of their own AI models with either more complex reasoning capabilities, enhanced performance, improved functionalities, or lower customer fees. This suggests the AI race is just in its infancy, and its rapid evolution will likely continue at a breakneck speed.

The impact on the real estate sector varies by property type

In December 2024, we introduced a framework to analyze the impact of AI technologies on the commercial real estate environment. We believe the recent case of DeepSeek, alongside the broader industry trends of lower barriers to entry and wider competition, does not change but instead accelerates the path toward our earlier, sector-specific conclusions. We expect AI to continue to have a significant and positive impact on data centers, life sciences, healthcare, and logistics; a marginally positive effect on hotels, student housing, and residential; and a neutral to a negative impact on offices, high-street retail, and shopping centers.

• Data centers: While the sector could experience some near-term demand moderation following the DeepSeek announcement, long-term prospects remain strong. Increased technological advancements will likely generate more data and, in turn, demand for space. In the U.S., major tech companies currently face years-long backlogs in model testing due to near-zero vacancy rates. Demand has outpaced supply, with rental growth approaching 30% annually in some markets. If DeepSeek’s model delivers on its promises, it could alleviate some immediate pressure by improving access to much-needed space while also fueling AI demand in the medium to long term—a net positive for the sector.
• Industrial: The sector could benefit from further supply chain modernization. While automation has already reshaped logistics, advancements in AI-driven operations could enhance efficiency and create new investment opportunities in next-generation distribution centers.
• Residential: Stronger productivity growth, typically associated with wage growth, could improve housing affordability and living standards. If AI-driven efficiencies translate into higher earnings, this could support demand for residential real estate.
• Retail: Stronger economic growth, driven by AI-enabled efficiencies, can strengthen business and household balance sheets. Additionally, AI applications in inventory management and operational optimization may help retailers reduce costs and improve profitability. However, these benefits could be offset by increased e-commerce penetration enabled by AI tools.
• Office: This sector may face the most significant challenges if AI accelerates the elimination of redundant positions. While AI has the potential to drive growth in emerging industries and improve corporate balance sheets, we feel that structural dislocations may act as a governor on a more robust demand recovery in an already challenged sector.

Conclusion

As history has shown with other transformative technologies, greater accessibility fosters widespread innovation and economic gains. AI is already delivering tangible productivity improvements, and its impact will undoubtedly expand. While the implications for commercial real estate vary by asset class, we expect the continued advancements of this technology to produce a positive effect overall. Ultimately, we recommend incorporating AI considerations into the investment decision process alongside other critical factors we have addressed in our Inside Real Estate – 2025 Annual Real Estate Outlook, such as macroeconomics, geopolitics, ESG, and demographics. We expect the approaching recovery in real estate to be primarily driven by income and rental growth. Thus, selecting property types with strong underlying fundamentals and favorable secular tailwinds is paramount.