I studied 5,000+ Series A and Series B stage companies in the USA to identify the next unicorns

[lovelycoffeexx]

I researched over 5000 Series A and Series B stage companies in the USA to identify the next unicorns.

I identified 10 companies that I feel could become unicorns very soon based on several factors including growth rates, hiring trends, and leadership. I then ranked them on a few metrics and presented the data to you below.

I used Crustdata's database for this research.

Top 10 Fastest Growing Soonicorn Companies in the USA​

1. LangChain is a language model application development library that develops a language model framework to power applications.
2. Vilya is a biotechnology company developing a novel class of drugs that precisely target the biology of disease.
3. Duckbill is an execution engine for daily tasks that functions as a personal assistant copilot using AI-powered technology.
4. Gutsy is a data-driven security governance platform that applies process mining to secure enterprises.
5. rabbit inc. is a tech company specializing in creating a customized operating system using a natural language interface. They most recently launched the rabbit r1 that got a lot of attention worldwide.
6. Babylon is building a new public Cosmos-based PoS blockchain called Babylon with a native token as the bridge between Bitcoin and the PoS world.
7. Saronic builds scalable, fully integrated unmanned surface vehicles and vessels for naval and maritime forces.
8. Unstructured transforms natural language data from raw to machine learning-ready through its open-source libraries and APIs.
9. Yurts is a secure, deployable, full-stack generative AI platform for Large Language Models.
10. Luminopia is developing a new class of treatments through digital therapeutics for significant neuro-visual disorders.

You can read the full research (with some cool pictures & graphs) here

 

[mousearts]

@lovelycoffeexx Langchain needs to figure out how to actually monetize their services. They have langsmith but unclear how much it’ll actually bring in.

 

[shefmasurj]

@mousearts Langchain is like jQuery for LLMs...

Except jQuery actually works and had wide adoption in production.

 

[giboradonai]

@shefmasurj lol you had me in the first half.

Langchain is imo best suited for rapid prototyping and POC validation.

When you need to prove out a concept really quickly, use Langchain.

When you need something production ready…I can’t imagine not building your own.

 

[lovelycoffeexx]

@mousearts Initially I was a bit wary about the library wrappers like Langchain but with the race to the bottom of foundational models there might be something very big for these wrappers. If they can manage to seamlessly integrate with all model providers and let developers use any provider with just a parameter change, I think this could be very big.

 

[hutchison76]

@lovelycoffeexx You must have never used Langchain if you believe that about them - “seamless” it is not, especially when you deviate even a bit from the very opinionated way their code is structured.

 

[caeman]

@lovelycoffeexx A lot of developers at this point don’t like langchain. It’s like the prototype sdk for RAG.

It’s basically the MySpace to what the someday Facebook will be.

 

[emma82]

@lovelycoffeexx I doubt Rabbit R1 is going anywhere

 

[wantsgirl]

@emma82 Apple has spent years on a new Siri that can do tasks and control other apps.

It's launching at WWDC this year and will be coming to the Apple Watch/Airpods as well as iPhone.

Pretty sure it will be a death blow to Rabbit and Humane.

 

[ajr86]

@lovelycoffeexx I did an analysis a few years ago (~2020) where I looked back 5 years and asked what companies existed back then, and what factors predicted success.

Things that predicted success:

- They described their company in terms of Security / Platform / Analytics / Teams / Real time / Enterprises

- Getting a government-backed monopoly on something

- They didn't use vocabulary like Human, Opportunities, Enable, Developer, Infastructure or Cloud

- If they do some open source thing, they generally started 1-3 years after the project started (when the source code base had 40,000-60,000 lines of code in it, corresponding to about $1m-$2m in development costs)

- Regular 125% annual revenue growth rates

So, based on past trends #4 (Gutsy) probably has a reasonable chance. If #7 (Saronic) can get a government contract, they'll do great. #2 and #10 are out of my area of expertise and analysis so I just don't know. If the world is still similar to 2015-2020, then the rest are unlikely to have any significant success.

 

[averyzuleger]

@ajr86 The thing is, the world is not similar to 2015-2020, it's most similar to 2003-2007 (edit). Someone above wrote that LangChain is the MySpace to the Facebook, and there's a way to explain this. LLMs created a way to skip training models and substitute it for traditional forms of AI. Obviously this doesn't mean other areas in AI are useless, but this basically was the catalyst to pushing us into a platform shift. First, everyone is fighting over what platforms to use, and now we're fighting over what developer tools to use. Those tools will be used to build the next big AI Consumer startups which will probably be freaking scary. Anyway, long winded way of saying that 2015-2020 was where you had to find areas of the world that hadn't been immediately eaten by software (edit) and provide some sort of B2B service. Right now, it's more like we're fighting over who will make the best jQuery.

 

[kenosis]

@lovelycoffeexx They might become unicorns but I see 6 and 10 having longterm growth like 20 years; the others are more niche and their growth may depend more on corporations and government contracts.

 

[packers213]

@lovelycoffeexx Wild that I’m a customer for one of these companies. Used Duckbill for about 4 months. It’s impressive how much work it will do for you, but I haven’t really leveraged it for much beyond scheduling appointments and requesting vendor details. On the fence for if I’d actually recommend it to other people.

 

[wynelle]

@lovelycoffeexx Interesting list. Have you looked at even earlier stage? Like DagsHub and Quandri?