Illia Polosukhin described a new mechanism where staking NEAR tokens grants users proportional access to AI inference power. For individuals, this replaces a monthly subscription of $20 or $200 with an owned asset that provides continuous access to chatbots and coding tools. For businesses, it transforms an expense line item into a balance-sheet asset. The model becomes especially powerful for autonomous agents, which can hold a small amount of NEAR to sustain their existence through continuous data output rather than paying per inference. Illia Polosukhin also revealed that the team is exploring liquid staking around this token, where a liquid staking token could be added to a liquidity pool, enabling new tokens to receive AI data output that feeds back into the project, effectively giving it intelligence and life.
Illia Polosukhin detailed the rationale behind the community's vote to reduce NEAR's annual token issuance from 2.5% to 1.6% over two years. He explained that at launch, NEAR had 5% inflation to bootstrap the validator network and ecosystem economy. After six years, the ecosystem felt that 5% was too high a burden. Last year, a vote halved it to 2.5%. The latest reduction brings NEAR closer to Bitcoin's 0.8% inflation rate, ahead of Bitcoin's schedule even though NEAR is 10 years behind Bitcoin in lifecycle. Illia Polosukhin framed the long-term goal as transitioning to a fixed supply, but noted this requires more research because a fixed supply like Bitcoin's creates an unstable economy despite its powerful meme. NEAR's vertically integrated stack, generating revenue that can pay validators and fund AI computations, provides the flexibility to solve this tension.
Illia Polosukhin argued that NEAR's decentralized, permissionless architecture gives it a fundamental advantage over tech giants like Meta or centralized platforms like Coinbase and Robinhood. He pointed out that large companies cannot build truly open, non-custodial, and uncensored platforms because they must protect their existing business models and are constrained by organizational pressure to prevent abuse. NEAR, by contrast, has already launched uncensored models, which Illia Polosukhin says is critical for cybersecurity and defense professionals. He highlighted the NEAR AI RAD program, which provides discounts and early model access to security professionals, as an example of how the decentralized stack can give good actors an advantage that centralized competitors cannot replicate.