Perplexity using US hosting for DeepSeek R1

Perplexity’s answer to my prompt about DeepSeek R1 being hosted in the US.

Perplexity AI has integrated DeepSeek R1 into its platform, offering users access to advanced AI capabilities with a focus on data security and freedom from censorship [1, 3]. This move allows users to leverage DeepSeek’s technology without the privacy concerns associated with data processing in China [5, 6].

Data Security and Privacy

A key advantage of using DeepSeek R1 through Perplexity is that all data processing occurs on servers located in the U.S. and Europe [3, 12, 15]. Perplexity affirms that user data remains within Western servers, ensuring compliance with stringent data protection standards and mitigating concerns about data sharing with the model provider or the Chinese government [3, 5, 11]. This commitment to data security makes Perplexity an attractive option for users who want to utilize powerful AI tools while maintaining control over their data[6].

Uncensored Access to Information

Perplexity has removed censorship from the DeepSeek R1 model, allowing users to inquire about any topic without restrictions [1, 3]. This ensures users receive unbiased and accurate information, empowering them to form their own opinions and conduct independent analysis [3, 5]. Perplexity’s CEO, Aravind Srinivas, demonstrated this by using a search about Tiananmen Square, showing the AI is uncensored within Perplexity[1].

DeepSeek R1 Features and Capabilities

DeepSeek R1 is known for its advanced reasoning capabilities, excelling in mathematical, coding, and logical challenges[2]. It utilizes a unique reinforcement learning framework, achieving impressive scores in benchmarks such as AIME 2024 and MATH-500[2]. The model also offers a “chain of thought” visibility, providing users with real-time insights into the AI’s reasoning process, enhancing understanding and transparency [5, 6].

Accessibility and Pricing

Perplexity AI offers different tiers of access to DeepSeek R1. Perplexity Pro subscribers have access to 500 DeepSeek R1 Pro searches per day, while free users receive five free uses per day [3, 4]. A Perplexity Pro subscription costs $20 per month or $200 per year and also grants access to other models like OpenAI’s o1[1]. DeepSeek R1 is available on the web and iOS, with the feature yet to arrive on Android[1].

Strategic Significance

Perplexity AI’s integration of DeepSeek R1 highlights the company’s commitment to providing users with cutting-edge AI technology while addressing concerns about data security and censorship[6]. By hosting the model on US and European servers and removing content restrictions, Perplexity offers a secure and open platform for users to explore the capabilities of DeepSeek R1 [3, 12, 15]. This move positions Perplexity as a key player in the AI landscape, bridging international technological advancements with local data privacy needs[6].

Sources [1] Want to try DeepSeek without the privacy worries? Perplexity AI just … https://www.techradar.com/computing/artificial-intelligence/want-to-try-deepseek-without-the-privacy-worries-perplexity-ai-just-launched-it-on-its-ios-and-web-apps [2] DeepSeek R1 w/ web search might be better than Perplexity - Reddit https://www.reddit.com/r/perplexity_ai/comments/1iacrzl/deepseek_r1_w_web_search_might_be_better_than/ [3] Perplexity: DeepSeek R1 Without Censorship - Azat TV https://azat.tv/en/perplexity-deepseek-r1-censorship/ [4] Perplexity comes out with its own AI assistant and DeepSeek R1 … https://nxtli.com/en/perplexity-ai-assistent/ [5] You can try DeepSeek’s R1 through Perplexity—without the security … https://www.fastcompany.com/91272776/you-can-try-deepseeks-r1-through-perplexity-without-the-security-risk [6] Perplexity AI Debuts DeepSeek’s R1: Bridging AI Power and Security https://opentools.ai/news/perplexity-ai-debuts-deepseeks-r1-bridging-ai-power-and-security [7] DeepSeek AI is free, but it’s literally down 99% of the time. - Reddit https://www.reddit.com/r/ChatGPT/comments/1igtnzw/deepseek_ai_is_free_but_its_literally_down_99_of/ [8] How a top Chinese AI model overcame US sanctions https://www.technologyreview.com/2025/01/24/1110526/china-deepseek-top-ai-despite-sanctions/ [9] Perplexity AI o1 vs DeepSeek R1: Who Wins at Research? - YouTube https://www.youtube.com/watch?v=4OxyDSXGzoc [10] DeepSeek Is Here to Stay as Microsoft, Perplexity Integrate Its Model https://www.pcmag.com/news/deepseek-is-here-to-stay-as-microsoft-perplexity-integrate-its-model [11] About the R1 model used for Pro Search on Perplexity https://www.perplexity.ai/help-center/en/articles/10496197-about-the-r1-model-used-for-pro-search-on-perplexity [12] DeepSeek Now In Perplexity’s AI Search—U.S. AI … - Forbes https://www.forbes.com/sites/luisromero/2025/01/28/deepseek-now-in-perplexitys-ai-search-us-ai-dominance-challenged/ [13] Perplexity lets you try DeepSeek R1 without the security risk, but it’s … https://www.zdnet.com/article/perplexity-lets-you-try-deepseek-r1-without-the-security-risk-but-its-still-censored/ [14] DeepSeek’s Popular AI App Is Explicitly Sending US Data to China https://www.wired.com/story/deepseek-ai-china-privacy-data/ [15] Is Perplexity Complicit? An Analysis of DeepSeek R1 Integration … https://www.linkedin.com/pulse/perplexity-complicit-analysis-deepseek-r1-integration-nikravesh-hfs8c