We want blockmachine to take over the free public RPC the Bittensor ecosystem depends on - restructured to need no account or API key, rate-limited by IP, with the right to serve it gated behind an on-chain conviction lock. Two reasons:
This vote authorizes the restructure and the exclusion of free-tier traffic from PSR buybacks - a deviation from the whitepaper, which is why it goes to holders. Paid-tier buybacks, pricing, and the work-based payout model are unchanged.
The Opentensor Foundation has long run the free RPC endpoint that much of the Bittensor ecosystem depends on - wallets, explorers, scripts, and tooling. It is critical infrastructure that OTF has carried as a cost.
Subnet 19 already operates production RPC infrastructure with routing, request logging, and correctness verification. This proposal is for SN19 to take over the free public RPC burden for Bittensor: the same service the ecosystem already relies on, provided by one of its own subnets.
The current free tier requires an account and an API key. The restructured free tier would:
This is the part that requires holder authorization.
The whitepaper commits the PSR to match miner payouts with alpha buybacks, funded by customer revenue. Free-tier traffic generates no customer revenue. Matching it with buybacks would mean buying back alpha against revenue that does not exist — which is precisely what the PSR is designed not to do. The buyback exists to be a verifiable signal of real revenue; matching it to free traffic would break that.
Under this proposal:
Excluding the free tier keeps the buyback honest as a proof-of-revenue signal. It is nonetheless a change from the whitepaper as written, and a material change to economic policy should not be made without putting it to holders.
Ecosystem positioning. Being the free RPC that Bittensor runs on is a standing, visible position in the ecosystem. The aim is placement alongside the existing OTF endpoint, and potentially as a default public RPC option in ecosystem documentation and tooling over time. That is durable attention for SN19 that does not depend on a marketing cycle.
Emission-worthiness. For any subnet, "is this doing something real?" is the question that matters most. A subnet serving real, public, externally-used infrastructure is doing visible, verifiable work. Running the ecosystem's free RPC is a concrete answer to that question.
Productive emissions instead of penalized burn. SN19 currently burns the large majority of its alpha emissions, because paid demand has not yet caught up to the emission pool. Under the emission rules introduced in chain spec 421 (June 2026), the share of miner incentive a subnet burns is now one of the factors in its share of network TAO emission — a high burn rate reduces that share. Directing a carved-out portion of emissions to serving real public traffic, rather than burning it, both puts the emission to visible use and improves how the network weights the subnet's emission.
Conviction reduces liquid supply. Serving free-tier traffic requires miners to lock alpha. Locked alpha is removed from circulating supply for the duration of the lock.
Removing the API key opens the free tier to abuse: an operator could generate free-tier traffic to themselves to farm emissions. The conviction lock is designed to make that uneconomic.
To be eligible for free-tier routing, a miner must maintain an active SN19 conviction lock sized to their cumulative free-tier earnings - the lock must scale with what they have earned from the free tier:
locked_alpha >= ratio * cumulative_free_tier_alpha_earned
The exact ratio is a subnet parameter, set at or near full coverage. If a miner's cumulative free-tier earnings outrun their active lock, the miner becomes ineligible for further free-tier routing until the lock is topped up.
The locked alpha is the miner's own and remains exposed to price. A miner cannot scale free-tier farming without also increasing their exposed alpha position, so farming the free tier and selling means selling into - and devaluing - their own locked stack. This does not make abuse mathematically impossible, but it makes farm-and-dump self-exposing and economically unattractive. The mechanism's only job is to keep free-tier spam from becoming a low-risk emissions-extraction strategy.
Free-tier work is accounted separately and does not contribute to paid-tier routing, paid-tier quality standing, or buyback-eligible revenue. Generating free-tier volume cannot be used to win paid-tier traffic or paid-tier emissions, and free-tier emissions are carved out rather than drawn from the paid payout pool. This protects miners serving real paid demand from being diluted by free-tier activity.
Free-tier work earns emissions with no buyback offset, and there are effects in both directions. Directing a carved-out portion of emissions to free-tier work rather than burning it puts more alpha in participants' hands, which on its own is mildly inflationary on that portion. Running the other way: under the spec-421 emission rules, a lower miner-burn rate raises the subnet's share of network TAO emission, and given SN19's pool a meaningful part of that additional emission is used by the chain to buy and recycle alpha — which is buy-side and supply-reducing. We do not claim the net effect is deflationary. The honest statement is that the burn reduction has a countervailing benefit the simple supply view misses, and that the magnitude is modest at current volumes. The structural lever for SN19's economics remains real paid demand, not this change.
The case for this proposal is positioning and emission standing, not a price argument.
Approval authorizes:
On approval, the whitepaper will be amended to reflect the revised buyback scope.
30-day average alpha
Weight is calculated at first vote and remains fixed if you change option later.