We’re finishing up the last preparations for the public release of Inferra Perps. Public launch is coming in the next 2-3 days.

If you hold more than 1% of solana:9ufM9TJd1UEmi9awnGfxCkCHAgQ3JZ5Sw6YxeSeEASY we want to talk. We’re opening private OTC buybacks for large holders. If you’re sitting on a sizeable position and would rather exit without selling into the open market, we’ll buy it back directly, off exchange, at a price we agree on. Together. This keeps large sells off the chart and out of everyone else’s way. Cleaner for you, cleaner for the market. DM's are open.

The Inferra Perps open beta is doing exactly what we wanted it to do – putting the product in front of real traders and showing us what needs to get better. We’ve already started fixing UX issues around the Pool on desktop and improving how positions are managed and understood. We’re also testing edge cases around limit orders and existing positions, while multiple-position support is now something we’re considering for future updates. Feedback keeps coming in, we keep shipping. We want Inferra Perps to feel familiar from the first trade – just with GPU compute as the market you’re trading.

Inferra Perps on mainnet are now in open beta. Everyone in our Telegram community now has access. You can trade, test the Perps, and try everything out. We’re looking for your feedback, bug reports if you run into any issues, and ideas on what we can improve. This is the final stage before the public release of Inferra Perps.

We’re wrapping up Inferra Perps testing with the Whale Chat and finishing a few small things. Tomorrow, we’ll open access to everyone in our Telegram community.

While we’re improving Inferra Perps on mainnet and collecting feedback from Whale Chat members, you can try trading on devnet. Check out the interface, test the features, and share your feedback with us. It’ll help us make Inferra Perps even better before the public release. Inferra Perps devnet: app.inferra.trade/pro

Where does the price for Inferra Perps actually come from? Not from a number we set ourselves. Perps use the same public benchmark as Inferra Index. GPU rental prices are collected from multiple platforms, filtered and aggregated, then verified through Chainlink and published onchain on Solana. The perp market syncs to that same feed. So the price you trade against is the same price anyone can inspect and verify independently. That’s the point of building the Index first. index.inferra.trade/

What does it actually mean to trade a GPU-hour perp? Take H100 as an example. If you think the market price of an H100 hour will rise, you can go long. If you expect it to fall, you can go short. The position follows the price of compute itself, not NVIDIA stock, not a GPU token, and not the price of the hardware. You’re trading the market price of one hour of H100 compute. That’s the idea behind Inferra Perps.

Getting Inferra Perps on mainnet ready for testing with everyone in the main chat. We’re finishing some UI fixes and improving how positions are opened and closed with limit orders. Perps are working and getting better. We’ll test everything as fast as we can, finish the last updates, and then open it up to everyone.

First feedback from the Inferra Perps beta is already coming in. Whale Chat members have started opening the first H100 positions on mainnet, testing the flow and giving us feedback on how the product feels in real trading. That’s exactly what this first stage is for. We’re watching how the market behaves, collecting the issues and suggestions, and making changes where they’re needed before opening access wider. So far, the beta is doing what we wanted it to do – getting the product into real hands and giving us real feedback instead of more internal assumptions. More updates as the first trading activity builds.

Inferra is integrated with @SwitchesBoard . solana:9ufM9TJd1UEmi9awnGfxCkCHAgQ3JZ5Sw6YxeSeEASY can now be used to pay for Switch Board services. You can already try it out.

Inferra GPU-hour Perps go live on mainnet this Monday. This has been one of the main products we’ve been building toward since launching the Index. The Index gave us a reliable on-chain price for GPU compute. Now we’re putting a market on top of it. More details on Monday.

Inferra Index is now available to AI agents too. We’ve released an MCP server that lets MCP-compatible agents access Inferra GPU pricing directly as tools. Instead of opening the Index and copying data into a prompt, an agent can request the data it needs itself and use it inside its workflow. For example, an agent working with compute infrastructure can check the current GPU market before making a decision. github.com/InferraTrade/infe…

A lot has changed at Inferra over the past few weeks, and the product is in a very different place from where it started. Inferra Index is live on mainnet, with GPU-hour prices validated through Chainlink infrastructure before they reach the on-chain feed. We’ve expanded the Index itself, improved the data layer and history, and started publishing regular GPU Price Reports so the market can follow how H100, H200, B200 and other models move over time. We also released @inferra/sdk, making it much easier for developers to use Inferra pricing inside their own products, and launched a private Telegram chat for top solana:9ufM9TJd1UEmi9awnGfxCkCHAgQ3JZ5Sw6YxeSeEASY holders. At the same time, we’ve been pushing onboarding, talking to potential users and partners, and continuing development on Inferra Perps. Tomorrow, another piece goes public: our MCP server, giving AI agents a direct way to access GPU pricing data from Inferra. And there is still a lot ahead. More integrations, broader access to Index data, more tools for developers, continued work on Perps and new ways to turn GPU pricing into infrastructure other products can actually use. The product has grown a lot in a short time. We’re nowhere near done.

Inferra GPU Price Report #5 This week across the selected datacenter GPUs: H100 – $2.8165/h · +14.2% H200 – $4.4139/h · -1.7% B200 – $6.3583/h · -0.1% Biggest move: H100, +14.2% We’ll keep publishing these reports regularly to track how the GPU compute market moves over time. Live data and full history: index.inferra.trade

Integrating Inferra Index no longer means working with the raw feed yourself. @inferra/sdk is now available for developers. Install the package, connect it to your app and use Inferra GPU pricing directly from code without rebuilding the integration from scratch. This makes the Index much easier to plug into dashboards, marketplaces and other products built around compute pricing. SDK: github.com/InferraTrade/infe…