We're excited to announce that we've raised a $7.5M seed round led by 20VC alongside an incredible group of founders (Canva, ElevenLabs, Datadog, Huggingface, Lovable, Synthesia, Pigment, and many more), to bring efficient, architecture-first foundational LLMs to the world.

Large language models are improving fast, but are lacking efficiency. The dominant mindset is still: add more compute to reason better. That works… until it doesn't - costs explode, and true reasoning barely improves.

At BottleCap AI, we pair foundational research with real product development harnessing architecture-first LLMs with efficiency as a core principle.

If something works in the lab, we turn it into an application for anyone to use.

And today, we're launching our first product on iOS called "Pulse: Community News", all built on top of CAP1, our own foundational model.

Pulse began as an internal tool designed to help the BottleCap team aggregate information and AI news, supporting us in prioritising research focus. News headlines are short and don't tell the full story, so we added 'community insights' on how people are reacting in real-time to specific events. As they say - there's wisdom in the crowd, and by summarizing opinions of the whole community we can distinguish important signals from noise faster.

Pulse already runs on our efficiency focused foundational model CAP1, and it's just the first of many applications we plan to build on top of our own technology.

This is only the beginning.