China enters the AI race with the release of DeepSeek, prompting conversations about what happens when AI tools take data from each other (rather than just the general public)

China enters the AI race with the release of DeepSeek, prompting conversations about what happens when AI tools take data from each other (rather than just the general public)

The race for domination continues to heat up at China’s AI model “DeepSeek” enters the fray, just days after newly inaugurated President Trump announced his plans to invest 500 billion in AI infrastructure during the course of his term.

Established as a startup under the same umbrella as the quantitive hedge fund High-Flyer, which is primarily owned by AI enthusiast Liang WenFeng (who built his fortune during the 2007-2008 financial crisis), little has been verified about how DeepSeek came to be.

That has not stopped endless speculation since it’s launch was announced, including how much of it is modeled after existing AI models such OpenAI’s ubiquitous model, ChatGPT.

Also being questioned is how the chips it was trained on were sourced, chip restrictions were placed in on China in 2019 which continued under President Biden specifically to curtail China’s ability to access infrastructure used in the advancement of AI technology. This restriction not only covered the chips themselves, but the technology used to manufacture them.

According to Liang, he sourced the the 10,000 Nvidia A100 GPUs prior to the federally imposed ban.

At present time the founders of DeepSeek are indicating that their goal is to continue the research and advancement of AI infrastructure with their model and not seek commercialization. To back these claims, you can currently download the first series of their model for free open source whether you’re a researcher or a commercial user.

It should also be noted that DeepSeek has an updated data set as compared to ChatGPT, which is currently capped to data from 2023, what this means is its most recent data is from 2023 and before and anything that occurred in 2024 and beyond would not be available so if you were to example ask ChatGPT "Who won the 2024 Presidential Election?" it may not give you a correct answer.

There have also been claims that DeepSeek is much cheaper to train, although training costs for existing AI models are largely inflated. These costs are based on the cost of cloud computing rental prices, which have a wide range of variance.

AI training costs vary wildly depending on a range of factors.

AI and cloud computing are both worthy investments for businesses looking to strategically position themselves for technology growth in 2025 and beyond, and Valley Techlogic is at the forefront of utilizing these technologies.

Whether it be initializing AI tools like Microsoft’s Co-Pilot in your business or migrating more of your operations to the cloud to reduce overhead spending on physical hardware, we’ve got you covered. Reach out today for a consultation and learn how you can catapult your business forward with technology advancements through Valley Techlogic.

Looking for more to read? We suggest these other articles from our site.

This article was powered by Valley Techlogic, leading provider of trouble free IT services for businesses in California including Merced, Fresno, Stockton & More. You can find more information at https://www.valleytechlogic.com/ or on Facebook at https://www.facebook.com/valleytechlogic/ . Follow us on X at https://x.com/valleytechlogic and LinkedIn at https://www.linkedin.com/company/valley-techlogic-inc/.