The Geopolitical Race Of LLMs

We are now a community of 235! Thank you❤️

This newsletter is free and I don’t use paid advertising. I completely rely on organic growth through users who like my content and share it.

So, if you like today’s edition, please take a moment to share this newsletter on social media or forward this email to someone you know.

If this email was forwarded to you, you can subscribe here.

If you want to create a newsletter with Beehiiv, you can sign up here.

If you are in tech and use social media, you must already be tired of the word ‘Deepseek’. In case you aren’t aware, you can read my last edition.

What followed the launch of Deepseek R1 is something that no one predicted.

Back in December, Deepseek already launched the Deepseek V3, another extremely powerful AI model which was allegedly trained on a much lower budget. It did create waves in the AI and trendy tech circles but didn’t reach out to the layman like dozens of other AI models that have been launched in the recent past.

The Deepseek R1 has shaken the tech industry and even major governments around the world.

There are multiple reasons why the R1 got much more hype than previous launches from Deepseek but one major reason is the rightful claim that the R1 performs better than OpenAI’s best reasoning models and is trained on a much smaller budget. Of course, we do not know what was the actual training budget for any of the recent OpenAI models because the word ‘Open’ is only in the name but with the money the company has been raising it can be a safe bet that their training budget is greater than $6 million.

The China Impact

The biggest reason why Deepseek has suddenly become a topic of discussion even among non-tech people is because it is made in China.

This positions China publicly as an AI leader. Of course, this is not unexpected but the US was not ready to cede its place as the best country for AI development; definitely not-so-quickly.

The problem is people draw conclusions based on whatever half-baked knowledge they have. Without reading the research paper or without even using R1, the general public reacted to “much lesser budget, better than OpenAI”

The market panicked. Shareholders sold any company associated with AI. NVIDIA fell the hardest, 17% in a day, a record $593 billion loss, the highest for any company in a single day.

Broadcom, Microsoft, and Google were all contributors to NASDAQ’s 3.1% fall on Monday.

Just a week ago, President Donald Trump announced a $500 billion investment in AI infrastructure through a venture known as Stargate.

For China, this was described as the “Sputnik moment” in AI. Marc Andreessen, the popular Silicon Valley VC described the Deepseek R1 as “one of the most amazing and impressive breakthroughs” and “a profound gift to the world.”

Apart from being as good as the best models in the world and allegedly being trained for 98% lesser budget, the Deepseek R1 is also open-source.

This is China saying - “we can make something better than you, cheaper than you, and give it away for free.”

Worldwide Reactions

With the markets tumbling, President Trump had to react publicly too.

I agree with most experts that the reaction of the market was unjustified and more panic than understanding of the situation. R1 being cheaply trained does not mean that the chips by NVIDIA or any other company are of any lesser value. It is just an indication that there is a scope for better optimizations in AI training. This does not mean that inference does not require high-performance networking and chips.

We will require faster and more powerful computing infrastructure to train better AI models in the future.

Of course, corporate teams at OpenAI, Anthropic, and Meta did go into damage control mode because their priority is the investors more than anything.

For example, this article by Dario Amodei, the CEO of Anthropic reads like ‘don’t let China do good for the world before we do good for the world.’ It’s more like a letter to convince the investors to pressure the government to impose more sanctions on China. It would have been better if Dario focused on how Anthropic has better projects in its pipeline than any company in the world and how it welcomes competition. Perhaps, the free market principles disappear when the competition is external.

Back in India, the reaction was more on the ‘if China can, why can’t we?’ Nationalist sentiments do tend to suddenly erupt whenever China does something positive that makes international headlines. India has this in common with the USA.

The difference here is the US government will do something about the reaction to Deepseek; the Indian government will pretend.

Like this announcement by India’s Electronics and IT Minister, which sounds okay on the surface but highlights a deeper problem of India’s slow culture. Even with one of the highest engineer populations in the world, India does not have a single decent manufacturing plant where AI chips can be made. India has to import all the hardware required to build AI. It is a model in which you are guaranteed not to win.

The Near Future

What this mic-drop moment by China has done is make AI a serious topic of conversation. A few more announcements like this over 2025, and AI will start creeping into daily conversations, not just of IT professionals and VCs, but also of traders, teachers, writers, scammers, students, and more.

This is just the beginning. AI becoming better, faster, and cheaper is inevitable. Companies and even governments around the world must be pressured to contribute as much as possible to making AI models open source. This is the only way to ensure access to the best technology in the world for the common man at affordable prices.

AI must not always be a technology guarded behind gates that require a $200/month subscription.

There’s a reason 400,000 professionals read this daily.

Join The AI Report, trusted by 400,000+ professionals at Google, Microsoft, and OpenAI. Get daily insights, tools, and strategies to master practical AI skills that drive results.

Sign up now for free and work smarter, not harder.

Did you like today’s newsletter? Feel free to reply to this mail.

This newsletter is free but you can support me here.

I’d be happy to connect on Medium, LinkedIn and X.