For the longest time, I’ve been scared of heights.
So we jumped out of a plane to launch Zorbe.
Because anything can be rented. But not what your product depends on.
Zorbe builds specialized AI models that companies own, and keeps them state of the art over time.
A lot of what companies run every day is repetitive. Those workloads deserve models built specifically for them.
We’re the first company doing multimodal model operations across language, speech, vision, image and video.
And the first to deliver a distilled video model into production for a customer.
Smaller. Cheaper. Faster. Smarter. Yours.
Own Your Models. Own Your Moat.
a trend we’re seeing right now: more companies want to own and control their models, rather than rent their intelligence from the big labs
Speedrun team Zorbe has gone from 0 to $1.6M cARR in just 6 weeks - building custom distilled models that give companies frontier-like performance while letting them own the models themselves
founders @yash_iyer21@SadaniPranay are cooking 🔥
For the longest time, I’ve been scared of heights.
So we jumped out of a plane to launch Zorbe.
Because anything can be rented. But not what your product depends on.
Zorbe builds specialized AI models that companies own, and keeps them state of the art over time.
A lot of what companies run every day is repetitive. Those workloads deserve models built specifically for them.
We’re the first company doing multimodal model operations across language, speech, vision, image and video.
And the first to deliver a distilled video model into production for a customer.
Smaller. Cheaper. Faster. Smarter. Yours.
Own Your Models. Own Your Moat.
We’ve been messing around with how far you can push a video model once you optimize it for production.
Here are a few generations from our base model with Zorbe’s proprietary optimizations (with no specific training done on a workload).
More on what we’re doing across different modalities soon.
I’m 18
And Zorbe is now backed by @speedrun
Custom models for repetitive workloads across all modalities
Text. Video. Image. Vision. Speech.
Cheaper, faster, smarter, yours
@yash_iyer21 and I are grateful to @far33d and @Tocelot for their confidence in us!
Let's talk!
I got into a16z speedrun as an 18-year-old UPenn dropout.
On our first day, @pmarca told us: owning your model is owning your moat.
Zorbe is an AI lab specializing foundation models for production workloads.
Sovereign video, image, TTS, and language models that YOU OWN.
Your weights. Your traffic. Your workload.
And making the models MUCH smarter, cheaper, and faster is just the side benefit ;)
@SadaniPranay and I are super grateful to @far33d and @Tocelot for their confidence in us and we're excited for what's ahead!
In SF? Let's get coffee.
Want better models that are actually yours? https://t.co/pYAkO7zdOi