🪴okay.. so what changed with deepseek?

Responsibility for the last bombshell goes to … deepseek. One of the biggest shifts in technological power, since a release of ChatGPT in 2022.

With the Chinese New Year, deepseek totally snake-d OpenAI away. So what does this mean for OpenAI, startups building in AI, and consumers?

Put your billing operations on autopilot using today’s partner: Paddle

Overwhelmed by complicated billing and compliance headaches? Give Paddle a try. All-in-one commerce platform handles payments, taxes, and subscriptions so you can focus on building a great product instead.

📝 Quick summary

You can go ahead and skip this section if you’ve followed the news closely:

• Last week Deepseek (Chinese company) released an open source foundation that nearly matches performance of OpenAI’s best model.

  

• They did it for ~5M USD, making use of Deepseek’s API 27 times cheaper than OpenAI (Deepseek charges $10 for a task that would cost $270 with OpenAI's API).

• Nvidia and US index funds took a massive hit.

  

• As if this wasn’t enough, there are rumours all Deepseek’s training has been done on outputs of OpenAI’s models...

📌 Why this matters

If you are building a startup, you need to avoid being just another AI app on the market at all costs. How? It’s more than ever about capturing unique advantages.

Markets typically evolve in a predictable way: once everyone can access the same core capabilities, the real battle moves to a higher level.

Take smartphones era as an example. 20 years back, manufacturers poured massive resources into phone hardware and custom operating systems. As soon as smartphones and standardised components became widespread, competition shifted from “Who has the most advanced hardware?” to “Who’s building the most compelling user experience?” 

That’s how companies like Apple managed to stand out. Through ecosystem, integration, and brand, rather than raw specs alone. Right now, AI is following a similar script. It’s more about how you stand out, than standing out on its own.

When everyone has access to the same “smartphone components” of AI, it’s the unique ecosystem around it that truly matters.

📩 Get a free pitch deck

Raising soon? I’ll send you over Passionfroot’s pitch deck that raised over $3M, just refer one friend via your personal referral link: https://foundersfuel.beehiiv.com/subscribe?ref=PLACEHOLDER

🔮 What happens next

OpenAI is still fine, it’s unlikely to go under anytime soon. It’s not ideal that they lose money per subscriber right now (around $200 per month), but they try to focus on a different approach anyway.

They are raising $40B and have turned their eyes towards an infrastructure project called Stargate. Their leadership is smart and have (probably) identified in the past that their model is replicable, even though Sam Altman didn’t admit it at the conference last year.

What about other startups?

I mentioned importance on how you stand out with your startups, and I can see these couple of points being key in startup differentiation in cheap AI age.

1. Proprietary Data: If you have a dataset no one else can replicate easily, your outputs stay unique.

2. Real Infrastructure: A global network of data centres, or some physical resource that others can’t copy overnight.

3. Customers: Who already has user relationships and distribution? If you can seamlessly integrate AI into a massive existing user base, you’ve got a leg up.

4. Speed & Execution: Back to the good old speed of iteration. The faster you are exploring something revolutionary, the more market share you capture from the get go. Get coding.

And consumers?

Just keep using the one you find working out the best. Prices will go down on all competitors, while the functionalities will increase. What a time to be alive.

✈️ How We Can Help