The End of the Auction Era: Polkadot’s Agile Coretime Shift

Polkadot’s parachain slot auctions ceased on-chain with runtime upgrade 1.2.0 on September 19, 2024. This date marks the definitive end of the "candle auction" model that had governed network access since November 2021. Under the old system, projects bid DOT to lock up slots for fixed terms ranging from six months to two years. The mechanism relied on a randomized end time to prevent last-minute sniping, ensuring fair price discovery through competitive bidding. However, the rigidity of these leases created significant inefficiencies for the ecosystem. Projects often secured more compute power than they needed, leaving valuable resources idle while smaller teams struggled to afford entry. The transition to Agile Coretime replaced this deposit-based lease with a direct payment model, fundamentally altering how developers interact with the Relay Chain.

From Leases to Direct Purchase

The shift to Agile Coretime transforms blockspace into a commodity purchased via two distinct methods: bulk and on-demand. Bulk coretime functions as a monthly subscription, granting consistent access for 28 days at a fixed price. This option suits active chains like Astar, which previously faced high capital requirements and uncertainty when renewing their legacy leases. In contrast, on-demand coretime operates on a pay-as-you-go basis, allowing projects to buy execution time block-by-block through real-time bidding. This flexibility benefits new or smaller chains with irregular traffic patterns, removing the need for massive upfront crowdloans. I find the removal of the two-year lockup particularly compelling because it eliminates the opportunity cost of bonded tokens. Previously, contributors locked DOT for up to 96 weeks in exchange for native project tokens, a structure that tied capital to specific outcomes. Now, DOT is burned upon purchase, introducing deflationary pressure directly linked to network usage.

The economic implications extend beyond simple cost savings. By burning DOT instead of locking it, the network creates a permanent reduction in circulating supply. This contrasts sharply with the previous model where locked tokens were eventually returned, limiting long-term value capture. Developers now pay only for the computational resources they actually consume, rather than reserving entire cores regardless of utilization. This precision allows for elastic scaling; if transaction volumes spike, a rollup can instantly purchase additional on-demand coretime to process blocks faster. Conversely, during quieter periods, reliance on bulk allocations keeps costs predictable. The system also enables sharing mechanisms where multiple projects can split a single core’s capacity, further optimizing resource distribution across the multi-core network.

Operational Realities and Risks

Despite the theoretical elegance, the current implementation reveals critical operational gaps. Two parachains, Mythos and Xcavate, suffered abrupt halts in block production after failing to manually renew their coretime within the required window. The system automatically released their resources back to the public market, where a third party quickly acquired them. When the teams attempted to buy back their cores, they faced quotes of $15,000 per core, a stark increase from their initial acquisition costs. This incident exposed the lack of auto-renewal features or renewal reminders in the early Coretime design. To prevent indefinite outages, the Polkadot Fellowship submitted an emergency proposal that temporarily increased available cores from 62 to 66 and granted free coretime to both projects for two months. While this intervention maintained network stability, it raised questions about the integrity of decentralized market mechanisms.

The absence of a formal secondary market has also led to monopolistic behavior. A group known as coretimenegotiation acquired 73 of 94 cores on Kusama and 9 of 18 on Polkadot, initially paying minimal amounts like 9.2 DOT per core on Polkadot. They subsequently set resale prices as high as 3,500 DOT per core during negotiations, effectively controlling scarce resources through private agreements. This centralized control disrupts the intended openness of the Coretime market and forces projects into expensive off-market deals. Web3 Foundation researcher Jonas has proposed a redesign including auto-renewals and a proper secondary market to address these issues. Until those changes land, teams must manage manual renewals with extreme care. Does the current friction justify the move away from auctions? I believe the trade-off favors efficiency, provided the governance layer continues to patch these human-centric design flaws.