INDmoney

A little-known Chinese AI startup, DeepSeek, has sent shockwaves through global markets.

It is challenging giants like Nvidia, OpenAI, and Google.

Its low-cost AI has triggered a selloff in US tech stocks and wiped out over $1 trillion in market value!

Here’s the FULL story 🧵👇

Founded in 2023 by Liang Wenfeng, DeepSeek first made waves with its DeepSeek Coder and LLM models.

Its real breakthrough came in May 2024 with DeepSeek-V2 — a high-performance AI developed at a fraction of the cost of its competitors.

Their latest model, R1, has pushed the boundaries even further.

👉 DeepSeek claims that R1 matches OpenAI’s O1 model, achieving a 97% success rate — but at 95% lower costs 👈

To put it into perspective

🔹 OpenAI O1 cost per million tokens: $75
🔹 DeepSeek R1: $10.24

That’s 86% cheaper!

A pricing disruption that’s already impacting global markets.

Globally, it has fueled fears of reduced demand for expensive GPUs.

👉 This has lead to a historic $600M drop in Nvidia’s stock — the largest single-day loss ever for a US company.

Adding to the drama, DeepSeek achieved all this with just 220 employees and $50M in funding.

Meanwhile, OpenAI employs 4,500 people and relies on billions in investment.

DeepSeek isn’t using cutting-edge hardware like Nvidia’s H100 GPUs.

Instead, they’ve innovated with H800 GPUs, sidestepping US tech sanctions.

DeepSeek has also partnered with AMD to build even more efficient AI models.

But it’s not all smooth sailing.

👉 Despite its success, DeepSeek has raised concerns over heavy censorship.

Discussions on sensitive topics related to China are met with responses like: “Let’s talk about something else.”

As DeepSeek overtakes ChatGPT in app downloads, the big question remains.

Can US tech giants fight back? Or is this just the beginning of a new era for AI?