Solana activated SIMD-0286 on mainnet at epoch 1009 on July 29, 2026, raising the per-block compute unit (CU) ceiling from 60 million to 100 million — a 66% capacity expansion. The upgrade, authored by Lucas Bruder of Jito Labs and first proposed in May 2025, represents the single largest through...
"100M already merged. I love it here." — Brennan Watt, VP Core Engineering, Anza
Solana activated SIMD-0286 on mainnet at epoch 1009 on July 29, 2026, raising the per-block compute unit (CU) ceiling from 60 million to 100 million — a 66% capacity expansion. The upgrade, authored by Lucas Bruder of Jito Labs and first proposed in May 2025, represents the single largest throughput expansion the network has deployed in recent history. No breaking changes accompanied the activation. Applications, wallets, and on-chain programs required zero modifications.
The move addresses a documented capacity constraint: 11.2% of all blocks produced under the prior 60M regime hit or exceeded 56M CUs, primarily during market volatility and peak DeFi trading periods. An intermediate 80M step was evaluated but bypassed after core engineers determined that 12 months of validator software improvements — particularly the adoption of XDP kernel-bypass networking by over 70% of mainnet stake — justified a direct jump to 100M. Block times remain at 400 milliseconds. The network did not trade latency for throughput.
The upgrade lands amid a broader Solana infrastructure cycle that includes Firedancer validator client adoption at 26% of validators, the Alpenglow consensus overhaul targeting Q3–Q4 2026, and a fee architecture that generates approximately $10 million daily in ecosystem-level revenue across DeFi protocols.
SIMD-0286 modifies exactly one network parameter. The change is narrow by design:
| Parameter | Previous | New | |-----------|----------|-----| | Max Block Compute Units | 60M | 100M | | Max Writable Account CUs | 12M | 12M (unchanged) | | Max Block Accounts Data Size Delta | 100 MB | 100 MB (unchanged) | | Slot Time | 400 ms | 400 ms (unchanged) |
The feature gate identifier is P1BCUMpAC7V2GRBRiJCNUgpMyWZhoqt3LKo712ePqsz. Activation progressed through devnet (epoch 1100), testnet (epoch 983), and mainnet (epoch 1009) without incident. No indexing changes are required. RPC providers and indexers should verify their infrastructure can handle blocks at the new 100M CU ceiling, though the data format itself is unchanged.
The proposal was authored by Lucas Bruder of Jito Labs, publicly posted to GitHub in May 2025, and spent approximately 14 months in review, testing, and staged deployment before reaching mainnet.
The previous capacity increase, SIMD-0256, raised the limit from 50M to 60M in July 2025. SIMD-0286 is three times larger in absolute CU terms (adding 40M versus 10M).
The upgrade responds to empirical data, not speculative demand projections. Between the July 2025 activation of SIMD-0256 and the July 2026 activation of SIMD-0286, 11.2% of all blocks produced on mainnet reached or exceeded 56M CUs — within 93% of the 60M ceiling.
This congestion pattern clustered around specific conditions: high-volatility trading periods, DEX order-book settlement spikes, and concentrated DeFi activity windows. The network was encountering its capacity ceiling precisely during the moments when throughput mattered most — peak demand. Transactions that could not fit within a saturated block were deferred to subsequent slots, introducing variable latency under load.
The 11.2% figure may appear modest in aggregate, but it represents a structural constraint. A network that hits its ceiling more than one block in nine during peak usage faces compounding queuing effects, particularly for time-sensitive operations such as liquidations, arbitrage, and payment settlement.
The capacity increase was gated on a prerequisite: sufficient adoption of XDP (eXpress Data Path), a kernel-bypass networking mode that reduces packet processing overhead for validators handling larger blocks.
Prior to activation, over 70% of total mainnet stake had enabled XDP. This threshold was considered the minimum required for safe network-wide propagation of 100M CU blocks without risking slot misses or increased block propagation latency.
XDP is available in Agave 4.0.0+ and is on by default in the Firedancer validator client developed by Jump Crypto. It is scheduled to become the default setting in Agave 4.2. The dependency on XDP adoption explains why SIMD-0286 spent 14 months between proposal and activation — the infrastructure had to be ready before the parameter could move.
Current validator client distribution across mainnet stake, according to mid-2026 data:
The validator set has consolidated to approximately 840 nodes, down from a peak above 1,300. This reflects both hardware demands and economic concentration, not a direct consequence of SIMD-0286.
The unchanged 12M per-account writable CU cap creates a mechanical shift in block composition under the new 100M ceiling. Under the prior 60M regime, a single saturated account consumed 20% of total block capacity (12M / 60M). Under 100M, that same account consumes 12% (12M / 100M).
The practical implication: more room for parallel, unrelated transactions when one account is compute-heavy. This benefits use cases where multiple independent programs execute simultaneously — the core design thesis of Solana's Sealevel runtime. Order-book DEXs, parallel payment rails, and multi-protocol DeFi interactions are the primary beneficiaries.
This is not a speed increase for any single transaction. A program that consumes 10M CUs still consumes 10M CUs. The upgrade expands the container, not the contents. It reduces contention between unrelated programs competing for the same block.
Solana's real-world throughput as of mid-July 2026, per Chainspect and Solana Compass data:
For context, Solana processed 25.3 billion transactions in Q1 2026, according to Startup Fortune — more than any other Layer 1. Daily active addresses exceeded 1.7 million as of mid-July 2026, compared to approximately 420,000 for Ethereum mainnet over the same period.
Average transaction cost on Solana stands at approximately $0.002, versus $0.053 on Ethereum mainnet. The capacity expansion via SIMD-0286 is unlikely to alter fee dynamics materially in the near term, as the network is not currently fee-constrained at the base layer.
SIMD-0286 is one component of a multi-quarter infrastructure cycle. The remaining elements:
Firedancer Full Adoption. Jump Crypto's independent validator client launched on mainnet December 12, 2025. As of Q2 2026, 26% of validators run Firedancer or the Frankendancer hybrid. Early benchmarks suggest Firedancer can handle over 1 million TPS in controlled test environments. Full majority adoption is projected over the next 12–24 months. Network resilience improves with client diversity — a single-client failure cannot halt the chain.
Alpenglow Consensus Overhaul. The most significant architectural change in Solana's history. Alpenglow replaces both Tower BFT and Proof of History — the two mechanisms Solana has run since genesis — with a new protocol called Votor. According to Coindesk, co-founder Anatoly Yakovenko indicated in May 2026 that Alpenglow could arrive "next quarter." The upgrade passed validator governance with 98.27% approval and is currently in test-cluster phase.
If deployed, Alpenglow would reduce transaction finality from approximately 12.8 seconds to 100–150 milliseconds — roughly a 100x improvement. Votor finalizes in one round if 80% of staked weight votes, or two rounds if 60% votes. A Validator Admission Ticket (VAT) mechanism, proposed at 1.6 SOL per epoch under SIMD-0357, would restructure validator entry economics.
SIMD-0306. This pending proposal would raise the per-account writable CU cap from the current fixed 12M to 40% of the block limit — which under the new 100M ceiling would mean 40M CUs per account. If activated, this would substantially change the parallelization dynamics described above and could benefit compute-heavy single-account programs.
The capacity expansion occurs against mixed economic signals:
Fee Revenue. On July 18, 2026, Solana topped all blockchains in daily network revenue for the first time in nearly five months, according to Solana Compass. Pump.fun led with $2.04 million in 24-hour fees. Solana's base-layer transaction fees contributed $568,000. Jupiter generated $172,000. Weekly protocol revenue in mid-July: Pump.fun at $5.51 million, FOMO at $1.39 million (an all-time high), Jupiter at $915,000.
TVL. Solana DeFi TVL stood at $4.86 billion as of July 13, 2026 — up 3.39% over the trailing 30 days per DefiLlama, though down from $5.28 billion on June 2.
DEX Volume. Solana processed $1.14 billion in DEX volume over 24 hours ending July 13, representing 20.45% of the $5.59 billion in total on-chain DEX volume across all chains. However, DEX volume declined 44.67% over the trailing 30 days to July 13.
Stablecoins. USD-pegged stablecoin supply on Solana totaled $14.92 billion — more than triple the chain's DeFi TVL — indicating significant settlement and payment infrastructure usage beyond DeFi lending and trading.
The economic question SIMD-0286 poses is whether expanded capacity translates into expanded revenue, or merely reduces fee pressure on existing demand. The data is inconclusive at this stage; the upgrade has been live for less than 48 hours.
SIMD-0286 is an infrastructure upgrade, not a feature launch. It expands the execution container by 66% while preserving latency characteristics. The engineering is sound: XDP adoption was gated as a prerequisite, the upgrade was staged through devnet and testnet, and no breaking changes accompanied it.
The economic implications are less certain. Additional block space reduces contention under peak load but does not generate demand. Solana's fee model — approximately $0.002 per transaction at the base layer — means that even substantially higher throughput produces limited incremental protocol revenue unless accompanied by increased transaction volume.
What matters next is whether the capacity is consumed. The 11.2% saturation metric that justified the upgrade will now reset under the 100M ceiling. If Alpenglow ships in Q3–Q4 2026 as projected and finality drops to sub-200ms, the combination of expanded capacity and near-instant settlement could shift the economics of on-chain payments and high-frequency DeFi. That remains forward-looking.
For now, Solana has more room and the same speed. Whether it fills that room is a demand question, not an infrastructure one.