Sprint: How Gonka's Consensus Works

Consensus is a mechanism by which all nodes in a network agree on the current state of the blockchain. Without it, each node would see its own version of the "truth," and the network could not function. In the history of blockchain, there have been three generations of consensus:

• PoW (Bitcoin, 2009) — miners search for meaningless SHA-256 hashes. Safe, but wasteful: 100% of energy goes into a "digital lottery."
• PoS (Ethereum, 2022) — validators stake tokens as collateral. Energy-efficient, but sacrifices decentralization: large stakers (Lido, Coinbase) control a significant share of the network.
• Sprint / PoUW (Gonka, 2025) — the third way. GPUs perform real AI computations that simultaneously serve users and confirm blocks.

Sprint is Transformer PoW 2.0, the unique Gonka consensus. The name reflects its essence: the network's work is organized into sprints (epochs), each of which represents a cycle of execution, verification, and reward distribution. Unlike Bitcoin, where "work" is iterating through random numbers, in Sprint, work is a forward pass through the Kimi K2.6 neural network.

Network scale: currently, Gonka operates ~4,648 GPUs, combined into ~582 MLNodes. Each MLNode is a GPU server with at least 40 GB of VRAM, capable of processing requests to the neural network. To service modern MoE models, the network requires a GPU cluster with a total VRAM of 640 GB. Sprint coordinates all these resources in real-time, distributing tasks and tracking each participant's contribution.

In the Gonka network, participants perform different roles, and Sprint coordinates their interaction. Hosts (miners) provide GPUs for inference and model training—they receive GNK both from genesis emission and for payment for processed requests. Transfer Agents are specialized dispatcher nodes that accept incoming AI requests, verify cryptographic signatures, and route requests to suitable MLNodes, taking into account load, available models, and latency. Validators provide cryptographic audits—verifying that nodes honestly performed the computations. All roles are economically incentivized through GNK rewards, and Sprint ensures their coordination without central control.

The network is fully permissionless—any owner of a suitable GPU can join without KYC by installing cosmovisor and MLNode. Task routing between nodes occurs based on hardware specifications, availability, and node reputation. Voting power in the network (Proof of Compute) is determined by the volume of computational work: "one unit of computing power = one vote." This is fundamentally different from PoS, where the vote is determined by capital.

Key difference from predecessors: Sprint is not an "empty" protocol tied to abstract difficulty. Each block contains proofs of actually performed work—neural network responses that were sent to users. This creates a direct link between network security and its utility: the more AI requests the network processes, the more secure it becomes.

To understand why Sprint is an evolution rather than just "another blockchain," it is useful to compare it with the classic Proof of Work in Bitcoin:

The main difference lies in the source of value. Bitcoin's value relies on "digital gold" — limited supply and consensus on value. GNK value is tied to the real demand for AI computing. When someone sends a request through the Gonka API, they pay GNK for a specific result — a neural network response. This is utilitarian demand, not speculative.

The second critical difference is scalability. Bitcoin cannot scale: more ASICs = more wasted energy, but not more transactions. In Sprint, every new GPU increases the network's bandwidth — more requests are processed in parallel, and blocks are formed faster. Network growth does not lead to energy waste but increases its useful power.

The third difference is hardware versatility. Bitcoin ASIC miners are useless outside of mining — they are specialized chips capable only of calculating SHA-256. The GPUs used in Gonka (H100, H200, A100) are universal computational accelerators. If a host decides to leave the network, their equipment retains its full value — it can be used for rendering, scientific calculations, model training, or sold on the secondary market.

The fourth difference is the economic reward model. In Bitcoin, the reward is fixed and halves every ~4 years (halving), regardless of how much real work is done. In Sprint, rewards are distributed proportionally to the Active Proof-of-Compute weight — the more AI requests a host has processed, the more GNK they receive. Additionally, hosts receive two parallel income streams: tokens from the genesis emission (80% of the total 1 billion GNK supply is distributed to hosts) and payments from users for inference (80% of each AI request). To maximize rewards, hosts lock GNK as collateral — without collateral, node weight is reduced by 5 times. This creates an economic incentive for long-term network participation rather than rapid "in-and-out" behavior.

Sprint security is confirmed by a CertiK audit — the leading Web3 audit firm, which completed the check in September 2025. The project has attracted ~$80M in investment from Coatue, Bitfury ($50M Series B), Insight Partners, and Benchmark — this is institutional confirmation of the viability of the Sprint model as a next-generation consensus.

Let's look at the full cycle of an AI request processing in the Gonka network — from the moment the user clicks "Send" to receiving the response and distributing rewards:

1. Request: a user (or application) sends a standard POST /v1/chat/completions via an OpenAI-compatible API. The request is signed with the wallet's cryptographic key — this proves that the sender has the funds to pay.
2. Routing: the request reaches a Transfer Agent — a specialized dispatcher node. The Transfer Agent verifies the signature, determines the required model, and finds an available MLNode with the appropriate specifications (sufficient VRAM, correct model loaded, minimum latency). Several Transfer Agents operate simultaneously in the network for fault tolerance.
3. Inference (calculation phase): the selected MLNode performs a forward pass through the Kimi K2.6 neural network. The GPU generates the response token by token in streaming mode. The user receives the response in real-time — latency is minimal.
4. Verification (PoC V2): in parallel with request processing, the network verifies the honesty of the nodes. 1–10% of tasks are randomly sent for re-execution by another node. The results are compared. If they match — both nodes confirm their honesty. If not — arbitration begins, and the dishonest node loses 20% of its stake. BLS signatures allow verification of results in less than 10 milliseconds.
5. Block creation: upon completion of an epoch, all proofs of work performed are aggregated into a block. The block contains: hashes of executed requests, BLS verification signatures, and data about each node's contribution.
6. Reward distribution: GNK rewards are accrued in proportion to each node's contribution. 80% of the payment for inference goes to the host that processed the request. 20% is directed to the Community Pool — a fund for ecosystem development (model training, grants). Additionally, hosts receive tokens from genesis emissions — these are two parallel income streams.

The entire process takes seconds. The user does not see the blockchain mechanics — for them, it is a regular API request to a neural network, just like to ChatGPT, only hundreds of times cheaper. The current network inference price is about $0.005 per million tokens, compared to $2.50—15 per million tokens at OpenAI.

Dynamic pricing — another feature of Sprint. The price for inference is recalculated every block depending on network load. In the stability zone (40—60% utilization), the price does not change. Below 40% — the price automatically decreases to attract users. Above 60% — it rises, stimulating the connection of new GPUs. The maximum change is 2% per block. This creates a market mechanism where supply and demand are balanced automatically, without manual intervention.

Security via DiLoCo: besides inference, Sprint coordinates distributed model training. Clusters of GPUs around the world train locally and synchronize every ~1000 steps via the DiLoCo protocol. This allows Gonka not only to serve but also to train AI models — without the need to collect all GPUs in one data center. The roadmap for 2026—2027 includes multi-model inference, where hosts will be able to serve different models depending on their GPUs. Sprint turns Gonka from an inference network into a full-fledged AI platform.

Protocol-level security: to protect against consensus attacks, Gonka utilizes Guardian nodes — 3 special nodes controlling 34% of the votes. This ensures that a malicious actor cannot seize control of the network, even with significant computational resources. Guardian nodes were introduced in version v0.2.7 (January 2026) as an additional security layer during the early stage of network development. As the number of independent nodes grows, the role of Guardian nodes will gradually decrease via on-chain governance — a management process in which all participants vote with tokens and PoC weight. This is standard practice for young blockchain networks: start with a controlled launch and gradually transition to full decentralization.