When Chinese startup DeepSeek announced a free generative AI tool to rival other major players in the market, we released a flash survey to capture reaction on the new open-source large language model (LLM). The study analyzed organizational adoption of AI, awareness and evaluation of DeepSeek-R1, perceptions of its capabilities, and potential impact on AI investment strategies.
The survey findings indicated a complex landscape of interest, skepticism, and caution among companies as they examine the value of DeepSeek-R1 compared to proprietary LLMs such as OpenAI’s models, including ChatGPT. DeepSeek-R1 has undoubtedly generated interest; however, it grapples with significant trust and security concerns that impede its widespread adoption. Numerous organizations are carefully evaluating its cost benefits; nevertheless, proprietary models currently dominate the market. Future adoption of DeepSeek-R1 will largely hinge on rigorous performance validation, robust security and governance assurances, and wider industry acceptance.
Watch the following video to get complete survey highlights.
ETR recently conducted a flash survey to gather insights on DeepSeek. A total of 106 respondents participated, including 23 from Global 2000 enterprises. The majority of respondents, 57%, came from large enterprises, while 44% represented small and midsize companies, providing a fairly balanced perspective across different business sizes.
The survey attracted a diverse group of professionals, including 30% in CxO-level roles, nearly half of respondents (49%) held senior leadership positions, and 21% were IT practitioners offering technical and operational perspectives. The majority of respondents, 71%, were based in North America, while 21% came from EMEA, and APAC accounted for 8% of participants.
This respondent composition highlights a broad and well-rounded mix of industries, company sizes, job functions, and regions, ensuring valuable insights into DeepSeek’s impact and reception.
Key Takeaways:
- AI Adoption Trends:
- 80% of organizations use commercial LLM subscriptions (e.g., ChatGPT, Microsoft Copilot).
- 63% integrate LLMs via APIs and web services, while 39% deploy open-source models on internal infrastructure.
- 27% are training proprietary LLMs, indicating an increasing push for in-house AI solutions.
- DeepSeek-R1 Awareness & Interest:
- 75% of respondents learned about DeepSeek-R1 only in January 2025.
- Awareness remains low, with 47% ranking their familiarity as neutral (3 out of 5).
- Despite this, 54% expressed strong interest in testing the model, with 35% extremely likely to do so in the next six months.
- Perceptions of DeepSeek-R1:
- 65% believe DeepSeek-R1 will reduce integration costs, but skepticism remains.
- 57% agree it performs near-equivalently to OpenAI, but 30% remain neutral or skeptical.
- Trust issues persist, with only 26% trusting its accuracy and 9% confident in its data privacy.
- Impact on AI Investment:
- 52% of respondents say DeepSeek-R1 will not change their AI investment plans.
- 25% say it will impact investment decisions, citing cost savings (44%) and increased competition (19%).
- Organizations prioritizing security and vendor commitments (35% and 16%, respectively) are hesitant to change strategies.
- AI Investment Strategy Outlook:
- 68% agree AI investments would increase if costs were lower.
- 51% will continue using pay-per-prompt models like OpenAI’s, despite open-source alternatives.
- Only 35% think DeepSeek-R1 will trigger an AI investment re-evaluation, while 42% disagree.
- Most organizations (73%) are not pausing AI projects due to DeepSeek-R1, suggesting it is not yet seen as disruptive enough to warrant major strategic shifts.
Among other aspects of the survey, ETR gauged the sentiment and possible skepticism about the tool. This analysis was intended to measure how organizations perceive DeepSeek-R1, focusing on its cost efficiency, performance, accuracy, and data privacy compared to other LLMs like OpenAI’s models and ChatGPT.
This portion of the survey shows that while organizations see cost savings and competitive performance potential in DeepSeek-R1, trust in accuracy and data privacy remains a meaningful concern. Companies may need further validation and real-world testing before fully embracing it as an alternative to proprietary LLMs.
ETR also examined how organizations view broader AI investment strategies, cost considerations, and the potential impact of DeepSeek-R1 on their decision-making. The full sample size of respondents rated their level of agreement on various statements in this question (based on a scale of 1 to 5), revealing a mix of financial constraints, openness to change, and cautious evaluation.
While cost remains a major limiting factor for AI investment, DeepSeek-R1 has prompted a mix of hesitation and strategic evaluation. Some organizations remain committed to pay-per-prompt AI models, while others see potential in self-hosting. Despite the buzz, the fact that nearly three-quarters of respondents will not pause AI projects suggests that DeepSeek-R1 has not yet made a compelling enough case to make companies reassess their AI plans.