Right as epoch 1009 rolled over, Solana quietly flipped a switch. Blocks got bigger in a very specific way: more compute per block. If you were refreshing a trading UI, a DePIN dashboard, or
Right as epoch 1009 rolled over, Solana quietly flipped a switch. Blocks got bigger in a very specific way: more compute per block. If you were refreshing a trading UI, a DePIN dashboard, or an NFT mint page, you might have just felt things breathe a bit easier.
The change is simple on paper. The maximum block compute moved from 60 million to 100 million CUs. That is a 66% bump. It went live at the very start of epoch 1009 on July 29, 2026, after months of prep work and network checks Solana Foundation (Upgrades).
It sounds like a number tweak. In practice, it touches fees, failure rates, validator hardware pressure, and how MEV searchers race to fill blockspace. Let’s unpack it in plain English.
Solana had been bumping into its 60M-CU ceiling during spicy traffic. One in roughly nine blocks came close to the cap over the past year, which telegraphed a simple message: the chain needed more headroom or more aggressive scheduling Solana Foundation (Upgrades).
A 66% capacity lift does not guarantee lower fees at all times. It widens the pipe. Whether fees drop depends on demand, how localized congestion is, and how quickly validators can propagate heavier blocks.
Who cares about this? App teams that hate user-facing failures. Validators who worry about networking and disk pressure. Traders and MEV bots who thrive when there is a burst of blockspace to capture. And regular users who just want their swap to clear on the first try.
Solana meters work using compute units. Every instruction burns some amount of CU. The runtime schedules many transactions in parallel as long as they do not touch the same accounts. In calm periods, a 60M-CU limit is fine. In peak periods, the limit acts like a low ceiling in a crowded room. You can still stand, but you are hunched over.
That ceiling affects two things people notice: how often a transaction bumps into congestion and how much you must tip to get priority. Local fee markets help isolate hotspots, but when a chunk of the network is hot at the same time, the global cap still bites. That was the story during busy NFT mints, viral memecoins, or oracle-heavy DeFi windows.
Solana increased the per-block compute limit from 60M to 100M CUs via SIMD-0286, a proposal widely covered as authored by Lucas Bruder of Jito Labs Solana Compass. The feature gate was activated at the start of epoch 1009 on July 29, 2026 Solana Foundation (Upgrades).
The foundation’s upgrade notes point to one main prerequisite: propagation needed to keep up. XDP, a kernel bypass networking stack, had to be in wide use so nodes could sling heavier blocks across the network fast enough. More than 70% of mainnet stake had XDP enabled before the switch was flipped Solana Foundation (Upgrades).
Item Before After Notes Max block compute 60,000,000 CUs 100,000,000 CUs 66% increase Solana Foundation (Upgrades) Activation 60M-CU limit live since Jul 22, 2025 100M-CU live at epoch 1009 (Jul 29, 2026) Epoch-based feature activation Solana Foundation (Upgrades) Prerequisites Baseline networking stack XDP widely deployed >70% of stake with XDP before flip Solana Foundation (Upgrades) Proposal N/A SIMD-0286 Reported author: Lucas Bruder, Jito Labs Solana Compass
This was not change for change’s sake. Solana’s own data shows that between the 60M-CU activation in July 2025 and the 100M uplift, 11.2% of blocks hit 56M CUs or more. Roughly one out of nine blocks was scraping the ceiling Solana Foundation (Upgrades).
Hugging the cap is a recipe for jitter. Latency becomes uneven. Fees spike in pockets. Some wallets get more aggressive with prioritization fees, which helps those transactions but leaves others stuck. A higher ceiling does not remove contention entirely, but it absorbs more of those short bursts that used to cause visible pain.
When traffic surges, more compute per block should reduce the frequency of failed or re-sent transactions that lost the fee auction by a hair. Fee markets will still do their job, and demand can still outrun supply, but the failure cliff likely moves further right on the curve. If your app used to see a spike of 4299 errors at the top of the hour, you might now see a smaller spike that clears faster.
Arb bots and searchers live in micro-bursts. A fatter block lets more bundles make it in before the state changes again, which may shave some tail risk for latency-sensitive strategies. That said, the extra space can also attract more bots. Net effect on realized fees is a tug-of-war between added supply and added demand.
These verticals tend to push many small writes and occasional heavier updates. Extra headroom makes it easier to schedule noisy periods without starving other programs. Oracles that batch updates might spread less over time if they can fit more into a single block window.
Jito-powered searchers and builders will adapt fast. With more block compute, the viable set of bundles and sandwich windows widens. But propagation still rules everything. If heavier blocks take longer to get around, the advantage tilts further to nodes with the fastest networking stack, which is why the XDP requirement matters so much Solana Foundation (Upgrades).
Validators feel this upgrade in their bones. Heavier blocks mean more to verify, more to store, and more to ship to peers. The foundation flagged XDP kernel bypass networking as table stakes for this round. It is less about raw CPU and more about getting packets off the NIC and into consensus with as few context switches as possible Solana Foundation (Upgrades).
Teams running nodes in data centers with modern NICs and tuned kernels are likely fine. Operators at the edge, or on shared hosts, may discover that missed slots creep up if they do not keep pace with the stack others are running. Syncing and snapshot churn could grow too. The healthy outcome here is bifurcation: those who invest in the stack capture more rewards, those who do not get nudged to upgrade or delegate.
One more point that is easy to overlook. Bigger blocks can amplify the cost of small bugs. If a heavy block stalls propagation for a subset of validators, you might see pockets of temporary liveness hiccups even if the average looks stable. Monitoring becomes more important than ever.
A higher ceiling typically flattens the fee curve a bit in the near term, then rising demand chases the new ceiling. Expect a short window where priority fees soften, then a return to dynamic equilibrium. Local fee markets continue to be the main line of defense against global congestion.
Builders sometimes contort programs to fit inside tight CU budgets. With 100M per block, some of those constraints loosen. You might see more complex on-chain checks run inline instead of deferred. Still, compute is a budget, not a blank check. Reckless CU usage will just push everyone back to the same congestion point.
Some teams used to punt heavy logic off-chain during peak. With more room, a portion of that logic can come back on-chain without murdering latency. Hybrid designs probably remain the sweet spot, but the slider can move a notch toward trust-minimized paths.
Raising the cap to 100M CUs is not an endpoint. It is a checkpoint that de-risks the next wave of demand while the network keeps getting faster at the plumbing layer. If demand keeps growing, there will be fresh pressure to raise limits again. If propagation ends up the bottleneck, the focus will swing to more XDP-like wins, better block engineering, and smarter scheduling.
Two practical predictions. One, builder UX improves quickly. Fewer support tickets about failed sends is a huge win. Two, validator stratification intensifies. Those who treat their nodes like high-performance trading infra will capture more of the pie. The community will likely debate how far to push limits versus how much to focus on making the current limit cheaper to propagate.
Capacity increases are leverage. They make good days smoother and bad days riskier if the plumbing is not ready. XDP adoption and careful monitoring are the guardrails.
If you want one place tracking these protocol-level shifts without the fluff, Crypto Daily keeps a tight read on upgrade notes, validator chatter, and real user impact. The coverage is quick and plain English, which helps separate signal from noise Crypto Daily.
A compute unit is Solana’s way of metering how much work a transaction does. Each on-chain instruction consumes a set amount of CUs based on its complexity. Blocks have a maximum CU budget, so the network can keep performance predictable even when traffic is wild.
It activated at the start of epoch 1009 on July 29, 2026, after the network cleared readiness checks around block propagation and networking Solana Foundation (Upgrades).
The change is tracked as SIMD-0286. Coverage and reporting credit Lucas Bruder of Jito Labs as the proposal’s author Solana Compass.
Heavier blocks require faster networking or you risk slow propagation and missed slots. XDP is a kernel bypass technique that helps nodes move packets with less overhead. The foundation noted that more than 70% of staked validators had XDP enabled before flipping the feature gate Solana Foundation (Upgrades).
Not automatically. It increases supply of blockspace. If demand is steady, fees should soften. If demand grows to match the new ceiling, fees will normalize around a new equilibrium. Local congestion can still push fees up for certain programs.
Between July 2025 and the new activation, 11.2% of blocks used 56M CUs or more. That is a lot of blocks brushing the 60M limit, which suggested the chain was frequently operating near the ceiling Solana Foundation (Upgrades).
Disclaimer: This article is provided for informational purposes only. It is not offered or intended to be used as legal, tax, investment, financial, or other advice.
