Show HN: GPU Perpetual Futures Prototype(github.com)
github.com
Show HN: GPU Perpetual Futures Prototype
https://github.com/zacharyfrederick/compex
4 comments
It can be quite easy to manipulate GPU prices - especially given only Vast and some small cloud support dynamic pricing of GPU on scale of 1hr, and you are using this for price feed. I can just rent every GPU on those platforms with like $1000/hr to wipe the whole short? Any thoughts on mitigating this?
Great question! I have been thinking about this. Reminds me of how Enron's traders would cause rolling blackouts in CA to drive up electricity prices. The current implementation winsorizes outliers but that isn't enough. Some fixes I see going forward
1. Implement additional providers and perhaps weight their contribution to the index by their inventory (harder for small providers to cause massive swings)
2. Use a Time-weighted settlement over 1-4hr periods to prevent rapid changes in spot
3. Verify data with bulk gpu compute marketplaces like ComputeExchange
4. Have third party services verify/or require some kind of "proof of rental" that gets sent to a registry. WTI/Brent has Platts and Argus that are price reporters who cross validate (similar mental model)
Do you have any more info on manipulating GPU rental prices? I would like to read more on that.
Do you have any more info on manipulating GPU rental prices? I would like to read more on that.
Could look into Uniswap's price oracle mechanism: https://hacken.io/discover/uniswap-v4-truncated-oracle/ to see how they mitigate oracle price manipulation.
The idea is airlines hedge jet fuel, starbucks hedges coffee beans - as GPU compute becomes critical infrastructure the same hedging tools should exist. Not sure if anyone actually needs this but it was interesting to build.
How it works: - Pulls live H200 spot prices from Vast.ai every 15s into a tradeable index - Full perp mechanics: funding rates, mark price calc, real-time P&L - Event-driven Rust backend with supervisor pattern and circuit breakers - Next.js frontend with TradingView charts, real-time WebSocket updates
What's real vs simulated: - Real: Index construction, funding rate engine, forward curve, state persistence - Simulated: Order book depth and trade matching (its a single-client demo)
The backend is the part I'm most proud of - isolated tasks coordinated by a supervisor, each has it's own state machine so if one component fails it doesn't take down the others. Tried to build it with production patterns in mind even though its just a demo.
Also made a 15-page derivatives pricing doc that covers the economic model and hedging scenarios. Basically: rental prices = f(CAPEX, utilization, depreciation) so futures pricing reveals market expectations about GPU supply/demand.
GitHub: https://github.com/zacharyfrederick/compex
Would love feedback on the architecture or if the market mechanics actually make sense. First time building something like this.