We've committed another $46,000 to Dot's infrastructure, bringing our total infrastructure spend to over $160,000+. Why? Because we know true privacy can only be achieved when the infrastructure powering our platform is native, and in-house, with zero provider trust required. 8x NVIDIA B300 Blackwell Ultra GPUs are now powering MiMo V2.6 Pro, a model which scored 46 on the Artificial Analysis Intelligence Index, just six points behind GPT‑6.1 Sol at 52. Handling complex reasoning, programming, and long-context tasks. Every layer is sustained by our fully custom, in-house infrastructure stack. This infrastructure gives us the foundation to host larger, more complex models and bring more of Dot’s inference in-house. At Dot we do not simply look to hand off our infrastructure to providers, but retain ownership of our entire stack. Try MiMo on Dot at: app.usedot.xyz/models?model=…

Here's everything you need to know about Dot's upcoming staking mechanics, and why this model is one of the best we have seen: 1) Our staking system is tied directly real product usage and revenue. When you stake $DOT, you earn inference credits. 1 credit represents $1 of usage on Dot, meaning your staking rewards are tied to something users actively want and already pay for. 2) The amount of credits you earn are directly tied to Dot's profitability. The current credit model ranges from 6% to 16% yield, with higher profits allowing more value (credits) to flow back to stakers. Platform profit directly determines two things: -How many AI credits stakers receive -How much USDC Dot sets aside to buy those credits 3) The 6%–16% AI yield is a parameter of the system. The current design starts with a 6% base AI allocation (never dropping below this) and increases as Dot's monthly profit grows, reaching the maximum 16% yield at $30,000 in monthly profit. The additional 10% is therefore the profit-driven upside. For example, if Dot makes $18,000 in monthly profit, the AI yield would be 12%. A user staking $10,000 of $DOT would then earn $100 of AI credits that month. These parameters can be adjusted over time based on Dot's actual revenue, margins, inference costs, staking participation, and overall economics. 3) The buyback side works separately from the AI yield. Dot allocates 40% of monthly profit toward buying back the AI credits earned by stakers. For example, if Dot makes $18,000 in monthly profit, the buyback budget is $7,200. If all stakers collectively earned $10,000 worth of AI credits that month, the IRM would have $7,200 available to redeem those $10,000 of credits, giving each $1 credit a cashout value of $0.72 USDC. So a staker with $10,000 of $DOT earning $100 of AI credits could either use the full $100 of AI on Dot or cash those credits out for $72 USDC. The result is a staking system built around sustainability. Stakers do not need to find a third-party buyer to monetize their rewards, and Dot provides a direct path to liquidity through the IRM, with the cashout pool funded by actual platform profits. This creates deep, protocol-controlled liquidity for earned credits while keeping the entire system directly connected to the performance of the business. As Dot generates more revenue and profit, the capacity of the staking and redemption system can grow with it, rather than relying on perpetual token emissions or creating an unlimited external liability. 7 days.