Dylan Rose - Founder and CEO of Evergreen
I am excited to announce the pre-orders for Evergreen II, a personal Supercomputer. Our company has thrived by empowering consumers with easy to use personal computers, and Evergreen II is our vision for a world where consumers have access to the latest innovations in AI.
Some Background on Me
I've always had an obsession with advancing personal computers. By the time I was 16 I put together >1000 Macbooks working in a computer repair warehouse. In high school and college I built my take on smart glasses and wearable cameras and launched Kai (2017) and GlimpseCam (2019). After shutting down the wearables startup I joined Amazon and worked with the creator of Google Glass in their special hardware projects division and learned how revolutionary products start from zero. In 2021 I was eager to get back to company building and founded Evergreen.
Some Background on Evergreen
Evergreen's first product was a small, quiet crypto miner based on a Raspberry Pi and Hard Disk Drive. Since launching in October 2022 we've sold and shipped over $5M of devices to thousands of people all over the world. We built out a multi-platform mobile app, custom operating system, production capacity, and a customer support team. This was all out of a garage in LA and it was absolute brutal building out reliable and repeatable production with a very small crew. Over the last 6 months we expanded into GPUs and launched a shared ownership model where individuals bought a cluster that we operated for them in a DC. We’ve sold over 800 GPUs in the last 3 months and have been renting out the cluster for AI inference. Now that we have several foundational pieces in place, we’re launching a new line of consumer products.
Moving from Personal to Supercomputer
When it comes to high performance compute there are really only 2 options. On one end of the spectrum you have M Series chips, and on the other end you have A/H100 systems. There isn't anything in between that delivers high performance AI with the ease of use as a modern PC. Evergreen's computers are built from the ground up to provide users with powerful AI tools that eliminates the pain of set up and maintenance. At first our computers will be high end, where customers will pay a premium, and then we'll drive down cost while maintaining the best cost per performance out of any AI PC in the market. Each generation will get more accessible, and more powerful.
The Sentient PC
Evergreen OS v1 will be conversational and mobile first, meaning you can access it anywhere, and easily share your applications with the internet. We're maintaining a library of the latest AI apps with 1-click install, as well as optimizing system drivers to get the most out of the system. We envision Evergreen II will have an operating system that is not just functional but alive, learning and adapting to its user. The OS will evolve around LLMs as a first class, low level object within the system. This OS will be uniquely yours, ensuring that when your LLM evolves, it evolves specifically for you.
Network Effects
Evergreen is a distributed computing company. We already have over 100 Petabytes of storage distributed all over the world. When you own an Evergreen device, you join a network of powerful, connected systems. Unused GPU capacity will be monetized, reducing the customer's cost of ownership to zero over time. Distributed training will build bigger and bigger models that we every customer can contribute to and profit from. These crowdsourced models will be free from big company or government intervention and contribute to flywheel growth. Every new user brings more compute to the platform, which creates more income, that pay developer to make better apps, which brings more users. We will need to keep driving down the cost of the units to reach mass scale, and eventually sell the machines at or below cost, which will be subsidized by Evergreen selling access to spare compute on the network.
Our Near Term Roadmap
Evergreen’s strategy takes is inspired by Tesla. The Evergreen II is our Roadster, a luxury, high performance machine that will create an initial customer base, and allow us to rapidly iterate with off-the-shelf parts. Its clear that AI PCs will be common place, and we have to ship fast to keep up with the pace of large consumer hardware companies. Here’s the general roadmap for the next couple of years.
- Recent Progress: Initial R&D and deployment of our 1000+ GPU cluster.
- Today: Pre-orders launch with a refundable down payment.
- Near Future: Begin shipping to early customers with OS v1.
- Next Year: Scale to 10,000 shipped GPUs and deploy distributed computing capabilities, develop top-tier OSS applications for end users.
- Next 3 Years: Create best price per performance AI PC (under $2,000) and ship 100,000 units
Evergreen II is a big swing, but the time to challenge legacy companies is now. I hope you'll follow along on our journey to reimagine personal computing. If your interested, Pre-Order Now for just 1% down.