The reputation of Solana as a high-speed, low-cost blockchain has always been based on its ability to process huge volumes of transactions with minimal friction. But even the fastest systems encounter limits.
Currently, each block on Solana is limited to 60 million units of computation (CUs). This ceiling applies uniformly, regardless of whether a validator uses a modest infrastructure or an enterprise-grade server.
Jump Crypto's Firedancer team believes these limits are no longer useful for the network. Their new proposal, SIMD-0370, aims to eliminate the cap per block altogether, allowing performance to grow based on hardware rather than governance decisions.
1/ SIMD-0370, by Jump's Firedancer team, proposes removing Solana's fixed compute unit block limit after Alpenglow. This would eliminate static caps on block limits and have validators skip blocks they can't process in time. Here's what changes 🧵 pic.twitter.com/xge1IViKnH
- Anza (@anza_xyz) September 27, 2025
If approved, the change would come after the long-awaited Alpenglow update, which introduces security mechanisms to make such an experiment feasible.
"Network throughput is constrained by an arbitrary number, not by what validators can actually do," explained Anza, a research company born from a spin-off of Solana Labs."SIMD-0370 corrects this mismatch."
Alpenglow as an enabler
The proposal relies heavily on Alpenglow, an update that Solana developers have described as the biggest protocol change in the history of the network.
Alpenglow introduces a mechanism that allows validators to skip blocks they fail to execute within the time slot. Instead of creating bottlenecks, the system continues without interruption. In parallel, it reduces the finality of transactions from 12.8 seconds to just 150 milliseconds, a leap that brings Solana closer to an 'internet-like' level of responsiveness.
This 'skip and go' dynamic is key. In this way, oversized blocks no longer pose a risk to the entire network. Validators unable to process them lose their rewards, while faster peers continue to produce. Block production thus becomes an adaptive system driven by hardware capacity rather than fixed limits.
"It creates a performance flywheel," added Anza."Block producers enter more transactions to earn commissions. Validators who skip blocks lose rewards and are incentivised to upgrade. Network performance grows progressively."
Supporters see a self-improving system
Advocates for the removal of the CU limit argue that the move could free up Solana's latent capacity. In their view, fixed constraints are a crutch of governance that prevents validators from competing on merit.
Three expected benefits:
- Dynamic scalability:during traffic peaks, block size naturally expands without waiting for planned parameter increases.
- Economic incentives:Providers maximise fees by entering more transactions, forcing lagging validators to improve or risk falling behind.
- Hardware innovation:Competition pushes validators to optimise clients and upgrade infrastructure, thus raising the average network performance threshold.
Some contributors see the proposal as continuing broader trends in Solana's validator ecosystem. Firedancer, launched on the mainnet in limited capacity in 2024, is a performance-focused client validator designed to diversify beyond Agave, the Solana Labs client. Inter-client competition has already improved efficiency, and the elimination of block limits could amplify this momentum.
Critics-warn-about-centralisation-and-complexity
Not everyone is convinced. On GitHub, engineers and community members have raised concerns about possible unintended consequences.
The most frequently cited risk is the pressure towards centralisation. Validators with greater economic resources could continually upgrade their systems to capture rewards, while smaller operators would risk being excluded.
"If larger validators continue to upgrade with increasingly expensive hardware, smaller ones could be forced out," warned one contributor.
Other risks include:
- Validator exclusion:Nodes too far behind could struggle to recover from snapshots or keep up with the chain.
- Strategic uncertainty:Leaders would have to guess the "safe" size of the block without risking jumps, a game that could favour the richest or most cautious.
- Future conflicts:The removal of the limits could complicate plans for multiple competing proposers or asynchronous execution, still under debate.
Persino Anatoly Yakovenko, co-founder of Solana, expressed scepticism. According to him, since today's blocks are not always saturated, removing the cap could have little impact on fees, latency or user experience in the short term. "There are a lot of vague arguments," he wrote, suggesting that the change could hinder future enhancements such as competing multi-proposer designs.
Economists also fear possible 'games' by validators: leaders who overload blocks with profitable transactions but risk being skipped, destabilising incentives. Unlike fixed limits, a market-driven approach lacks predictability, making it more difficult for smaller operators to plan sustainable investments.
What's next?
SIMD-0370 is still under review and would not be activated before Alpenglow's deployment. The developers expect extensive testing on testnet before any mainnet deployment.
For it to be adopted, a number of conditions must be met:
- Validator consensus:majority must accept the risks of centralisation and economic effects.
- Strong modelling:stress tests on peak demand, propagation errors and reward dynamics are needed.
- Future compatibility: designers must ensure that the change does not compromise other long-term scalability strategies.
If implemented, the proposal would mark a turning point in blockchain architecture: Solana would move from fixed ceilings to a market-driven performance regime, where hardware innovation directly determines network throughput.
It remains to be seen whether this will make Solana the fastest blockchain on a large scale, or push it towards a concentration of validators. What is certain is that Firedancer's proposal forces the network to tackle a crucial node of Web3: the balance between speed, decentralisation and resilience.
