China’s DeepSeek AI Disrupts Tech Industry, Opening New Collaborations with Insurtechs
China’s DeepSeek AI Disrupts Tech Industry, Opening New Collaborations with Insurtechs
DeepSeek, the Chinese artificial intelligence (AI) model that has taken the tech world by storm, continues to make waves with its latest industry partnership. 

The AI model, which has surged to the top of Apple’s App Store downloads, is now being integrated into Roadzen Inc’s (a leader in AI-driven insurance and mobility solutions) MixtapeAI platform, marking a major step in its expansion into global business applications.

Speaking to Insurtech Insights about the move, Rohan Malhotra, CEO and Founder of Roadzen, said: “It’s very important to understand what DeepSeek’s models are and are not. DeepSeek R1 and V3 are fully open-source models, with both their weights and methodologies published. R1 is the only open-source reasoning model available. Security or data sovereignty is not a concern if hosted on your own servers or deployed in Amazon Bedrock or the like, as most companies will do. The data stays in the US, or wherever you host it, does not go to China, and you can easily fine-tune the base model for your own use case.”

“R1 is the only open-source reasoning model available. Security or data sovereignty is not a concern if hosted on your own servers or deployed in Amazon Bedrock or the like, as most companies will do. The data stays in the US, or wherever you host it, does not go to China, and you can easily fine-tune the base model for your own use case.”

Rohan Malhotra, CEO of Roadzen

DeepSeek explained

Malhotra described the technology as, “a win for both foundational AI companies and applied AI businesses,” saying: “Foundational AI companies can now replicate DeepSeek’s methodologies in their own models since they have been openly published. Meanwhile, applied AI companies gain access to an open-source alternative with significantly lower inference costs compared to other foundation model providers.”

Malhotra also clarified Roadzen’s relationship with the technology, saying: “Lastly, Roadzen is not collaborating with DeepSeek, in the same way that integrating Google Maps into your app will not be called collaborating with Google. We are offering DeepSeek’s advanced reasoning capabilities and lower inference costs in our agentic platform Mixtape to our insurance, auto and fleet customers.”

DeepSeek Undercuts OpenAI Market

DeepSeek first gained global attention after its January 20 launch, shaking financial markets and challenging the dominance of U.S. tech giants. Its ability to perform at levels comparable to OpenAI’s o1 model—while operating at a fraction of the cost—has raised questions about the long-term advantage of high-performance chips. 

Reports suggest DeepSeek’s training was significantly cheaper than OpenAI’s GPT-4, with its founder, Liang Wenfeng, stockpiling Nvidia A100 chips before the U.S. imposed export restrictions on advanced semiconductor technology to China.

The model’s impact has been felt beyond AI research, triggering a historic $600 billion market loss for Nvidia and a broader sell-off in AI-related tech stocks. Analysts say DeepSeek’s success underscores the rapid evolution of China’s AI sector, which has been a key focus for President Xi Jinping’s push for technological self-sufficiency.

DeepSeek censorship

While DeepSeek has drawn comparisons to ChatGPT, its application within China remains subject to strict government censorship. Like other Chinese AI models, including Baidu’s Ernie and ByteDance’s Doubao, it filters politically sensitive content, adhering to Beijing’s stringent regulations. Nonetheless, its technological prowess has positioned it as a serious competitor in the global AI landscape.

In a test, a user requested DeepSeek to explain what happened in Tiananmen Square. DeepSeek started collating information – then stopped, telling the user, ‘Sorry, that’s beyond my current scope. Let’s talk about something else.’

By contrast, when asked the same question, ChatGPT delivered a contextual and fairly in-depth response to the political event that has been a blot on the CCP’s political history since it happened in 1989.

Speaking to Insurance Edge, Rory Yates, Global Strategic Lead at EIS, suggests a cautious approach to the technology. 

He says: The big question that’s causing most of the concern is whether the entire industry has been wildly overspending and investing? It’s also raising profound questions about how China may have undercut America’s (and emerging UK, Europe and likely other markets) most critical economic advantage on A.I. by making its technology free.”

He continued: “There’s so much to unpack and some massive questions to answer from the true cost, model implications, trust, how it can be harnessed from a business PoV, the censorship people are already reporting and how this will work going forward etc. etc. etc. Until we have more information, it will be very difficult to get to the bottom of any of these things.”

Market excitement at new contender

But political issues aside, the technology is an attractive option for companies seeking the latest in generative AI technology. 

Roots Automation CEO Chaz Perera agrees, and says the advent of DeepSeek is an opportunity for more competitiveness to enter the GenAI marketplace. “Further democratization of this technology is a win for the DIY community, but doing it at scale still requires a deep understanding of AI and engineering that most insurance companies don’t have.”

“What’s exciting about DeepSeek is that they’ve identified a way to reduce operating costs on richer models, but it’s early days for DeepSeek models and there are questions about its security that need to be addressed before a regulated industry can use it.”

“We are excited to experiment with and consider building upon similar models, hosted in-house at Roots. Roots would not send our customers’ data to third-party systems outside our own cloud.”

DeepSeek an opportunity?

The technology also provides an opening to those exploring the Agentic AI space – as automation advances continue to speed up customer claims processes and more. 

Yates adds: “It is certainly way too early for any stock market calls. Lets hope the trading algo doesn’t get carried away. Another “fake” market move is exactly what we don’t need. Once we start getting more information and answers we can start to form more considered views.

“For now we need to keep exploring all of these models. For most of us harnessing multi-AI agent strategies means looking in-depth at everything in this space and asking more questions of our technology foundations, data models, security and so on.”

With China’s AI advancements accelerating and global companies eager to harness its capabilities, DeepSeek’s rise is reshaping the future of artificial intelligence in unexpected ways.

Reporting by Joanna England

Share this article:

APPLY TO SPONSOR

Gain access to the most senior audience of insurance executives, entrepreneurs, and investors. We offer a wide range of opportunities for you to engage with our attendees from networking to thought leadership.

Sponsorship packages provide a wide range of opportunities developed for almost any budget and are designed to help achieve your branding, networking, and/or thought leadership goals. 

Insurtech Insights Europe 2025

Join us at Europe's largest insurtech conference at InterContinental London - The O2
on March 19-20th, uniting over 6,000 senior insurance professionals!