Synclave

Yeah, so we actually started as, TEESQL, but Synclave was more interesting to the teammates. So I'm Dan. This is Crystal. We also have another teammate, Fran, who unfortunately couldn't be here. But rather than talk about us, I'd actually like to talk about Albie. So Albie, as you heard, is shipping Pitch Rotator, and obviously, right, like, your pitch is your most important information. You haven't actually raised yet, or maybe you have, but a small amount. Your competitive advantage is being given away to this platform. So obviously, various customers care about privacy. I know I do, right? I trust Albie, but I don't necessarily trust Anthropic or OpenAI. So with a product like this, she really needs to focus on making sure that her customers' data stays private, right?

So enter confidential compute. Unfortunately, trusted execution environments and CVMs are incredibly, incredibly difficult to set up and run. If you want to buy your own machine, you're talking probably hundreds of thousands of dollars. If you want to rent it, probably a couple thousand a month. For the base level entry CVM, the type of thing that would cost you five dollars on a typical cloud platform, you're going to spend at least fifty dollars a month. So enter Synclave, right? Synclave starts with a product very similar to, Vercel, right? So if Albie has her marketing platform and she wants to go ahead and collect email, she can deploy for free, hook up GitHub to the app in Synclave, and Synclave will deploy for her automatically inside of a shared CVM environment.

Again, this is free. When she's ready to go to, actual, like, MVP stage and get real customers on board, she can upgrade to a five dollar a month shared enclave. So a shared enclave is essentially a sandbox running with other apps inside of a CVM, and what the shared enclave gives her is convenience without as much trust, right? So at the beginning with an MVP, she's bringing on people like me who know her, who trust her. Critically, she can use private inference in this environment. So we've partnered with RedPill, a Phala product, to make sure that GLM 5Q and a number of other frontier open source models are available to Albie and her team. So moving on, as Albie produces a product that she knows is right and she wants to actually launch it and spread it to her network, one click of a button is going to allow her to promote it to a full CVM.

So this five dollar shared enclave becomes a private environment fully attested for Albie's use and the confidence of her customers.

So critically, when Albie gets to the point where she's ready to scale, she's gotten a few viral moments, she doesn't have to re-platform. We support deploying to GCP. We support deploying to AWS all through dstack. We've got, partnerships with Phala, who built dstack in collaboration with Andrew Miller and Flashbots, and we've, moved on to partner with another of, a number of hardware providers who will help to provide the data center substrate if Albie doesn't want to think about her own accounts. But the second that Albie wants to use her own account, she can.