Starknet Teases Grinta Upgrade For September

In September, Starknet will activate Grinta (v0.14.0), its most ambitious upgrade to date. The release introduces multi-sequencer architecture, sub-second pre-confirmations, faster blocks, and a new fee market, pushing the Ethereum Layer-2 closer to full decentralization.

According to Starknet team, Grinta isn’t just an upgrade. It’s a turning point in L2 design, bringing Starknet’s infrastructure in line with its decentralization and UX goals, while giving builders a faster, fairer, and more programmable network.

What’s New in Grinta

Here’s what v0.14.0 brings to mainnet:

1. Multi-Sequencer Consensus (via Tendermint BFT)

Starknet now runs with three sequencers (StarkWare + partners like Juno), moving away from a centralized operator. It uses Tendermint BFT to agree on blocks, with a roadmap to expand to a permissionless set.

Block finality now depends on Top 100 STRK stakers, requiring ⅔ weighted signatures: a governance-linked security layer.

For fallback, a temporary “safety lever” lets StarkWare step in if liveness drops too low, overseen by a Security Council.

2. Sub-Second UX with Pre-Confirmations

Starknet introduces pre-confirmations in ~0.5 seconds, decoupling user experience from full finality. Transactions now feel instant, while full consensus continues in the background—huge for DeFi and gaming.

3. Block Times Cut to 4–6 Seconds

From 30s down to 4–6s, block times get a dramatic boost. Further reductions are planned, including native Cairo execution to scale throughput without sacrificing STARK-proof security.

4. Fee Market Overhaul (EIP-1559 Inspired)

Starknet ditches FIFO mempools and introduces:

• Base fee + tips model (starting at 3 gFRI)
• Tip-based prioritization for transactions
• v3 transactions only, requiring STRK for gas (paymasters like AVNU can abstract)

The base fee isn’t burned yet, but that’s coming later this year. For now, the model aims for predictable costs, spam resistance, and validator sustainability (~$1K/month run cost).

The Deployment Plan

• Approved by token holder vote (SNIP-30) in late August
• Tested on Goerli and Sepolia
• Mainnet pause: ~15 minutes during upgrade
• No breaking changes to RPC or dev tools
• Wallet support: Argent X, Braavos ready; v3 txs only

Validators and full node operators must upgrade clients immediately (e.g., Pathfinder v0.10.3+, Juno v0.9.1+).

Strategic Context: Not Just Ethereum Scaling

Grinta isn’t just about Ethereum. It aligns with Starknet’s broader roadmap, including:

• STRK staking integration (~430M STRK staked pre-upgrade)
• BTC interoperability via LBTC and ColliderVM
• Future upgrades (SNIP-33) to decentralize proving and expand sequencer participation

Starknet’s Grinta upgrade delivers on three fronts: decentralization, usability, and economic design. It’s the first step toward full permissionlessness—and a clear signal that the L2 race isn’t just about TPS anymore.