After a 70% drop in nodes, this time Solana got anxious

Author: momo, chaincatcher

The number of Solana nodes has decreased by 70% from its historical peak. In early April, according to data from Solana Compass, the number of validators sharply dropped from 2,560 in March 2023 to about 756; during the same period, the Satoshi coefficient declined from 31 to 20, a 35% decrease, indicating a weakening of decentralization.

This change occurs precisely at a time when Solana is trying to tell a grander story—becoming the “On-Chain Nasdaq,” carrying the global capital markets. The sharp reduction in nodes versus expanding ambitions creates an unavoidable tension.

Regarding node and centralization issues, Solana has not been without responses in the past, but the results have been less than ideal. Recently, according to SolanaFloor, the Solana Foundation announced a new validator policy, which will take effect on May 1. What is the focus of this new policy? Can it change the current situation?

1. Why has the number of nodes decreased significantly?

Looking at the trend of Solana node numbers, the sharp decline in validators has not been sudden. Since early 2024, the number of nodes has been steadily decreasing, gradually falling below 1,000.

The large drop in node count earlier this year once caused community panic. In response, Solana founder Toly explained that the main reason was the end of subsidies.

For a long time, Solana has been criticized for insufficient nodes and over-centralization. To quickly expand the validator base, Solana launched an early Delegation Fund Program (SFDP), which provided support through staking matching, residual delegation, and voting fee subsidies.

Simply put, the foundation matched external stakes 1:1, with a maximum match of 100k SOL; the remaining staked SOL after matching was evenly distributed among qualified validators; and daily voting transaction fees were also subsidized. This mechanism was effective in the short term. A report from Helius in August 2024 showed that at its peak, over 70% of validators depended on this system to varying degrees.

However, problems soon emerged. While these subsidized nodes made up the majority in number, they only controlled about 19% of the total staked tokens; in contrast, around 420 non-subsidized nodes held over 80% of the stake, with the top 20 nodes accounting for more than one-third of the total stake.

Clearly, having more nodes does not equate to “decentralization.” Subsidies attracted many low-stake, low-performance “nominal nodes” that, although dispersed, lacked the capacity to participate in real staking competition; institutions and large holders with substantial SOL holdings preferred to allocate resources to large, reliable, operationally stable nodes.

This also explains why, despite the previous surge in node numbers, the Satoshi coefficient did not increase proportionally.

For Solana, maintaining a large number of underperforming, minimally contributing “nominal nodes” is less desirable than establishing a smaller, more professional validator set to ensure long-term network stability and security. Therefore, Solana began to actively reduce subsidies.

Starting in 2025, the foundation gradually adjusted delegation strategies, phasing out nodes that relied heavily on subsidies. The core principle is summarized as “one in, three out”: for each new subsidized node, three old nodes are eliminated based on two criteria—having received foundation delegation for at least 18 months and having external stakes below 1,000 SOL. At that time, it was estimated that about 51% of nodes met the criteria for removal, totaling approximately 686 nodes.

After subsidies ended, survival for small and medium nodes became more difficult. Analysis indicates that nodes need to have about 3,500 SOL staked and incur roughly $45k annually (mainly due to voting fees, about 400 SOL) to operate.

Meanwhile, competition within the network intensified, with top validators vying for delegation with nearly zero fees, further squeezing profit margins for small and medium nodes.

Additionally, hardware upgrade requirements from network upgrades like Alpenglow have led some older equipment to be phased out, raising the entry barrier for validators.

However, the cleanup of small nodes and the sharp decline in node numbers have raised community concerns about excessive centralization. A Twitter user commented: “Users use PoS chains for security, but the chain becomes more centralized in pursuit of security. What exactly are we supporting?”

2. What is the intention behind Solana’s new validator policy?

In this context, let’s review Solana’s latest validator delegation plan.

The most significant change focuses on strict constraints at the infrastructure layer.

No single ASN (which can be understood as a cloud provider or network service provider) can hold more than 25% of the total stake, and no single data center can account for more than 15%.

In other words, even if you run compliant, stable nodes, if “too many people are in the same place,” you might lose foundation support.

The logic behind this is straightforward. Currently, although Solana validators appear dispersed, physically they are highly concentrated among a few cloud providers and data centers. Helius reported that two hosting providers controlled over 40% of the total stake, mostly concentrated in Europe. Sources told ChainCatcher that the Solana Foundation is also interested in supporting nodes in Asia.

This new regulation is more like a “forced splitting,” aiming not to increase the number of nodes but to require nodes to migrate out of overly centralized infrastructure, dispersing the risks associated with stacking on a few nodes.

At the same time, the rules tighten validators’ operational freedom. These include requirements to complete transaction ordering within 50 milliseconds, process transactions according to predefined rules, enforce data sharding at a steady pace, and explicitly prohibit censorship or delays of transactions received by TPUs. These constraints directly address long-standing issues of MEV competition and opaque execution, essentially compressing validators’ “operational space” and replacing it with more standardized rules to improve network consistency.

From a strategic perspective, this is an upgrade of the “one in, three out” policy from last year, using rules to filter for more qualified nodes.

However, controversy has arisen. Chainflow, a node operator, raised concerns in public discussions.

On one hand, under the current rules, whether a validator continues to receive delegation depends less on operational quality and more on its “location.” If a cloud provider or data center has already hit the limit, then regardless of the node’s performance, deploying there might exclude it from foundation support. This could mean that some long-standing, stable small validators lose their support simply because they are “located in a crowded environment.”

On the other hand, the practical issue is migration. High-quality infrastructure resources are concentrated among a few large providers. Once small and medium nodes are forced to migrate, they often end up in less reliable, lower-performance data centers. This could lead to decreased node performance, lower block production rates, and reduced earnings, potentially accelerating their market exit.

In summary, Chainflow believes that for small and medium validators, the greatest uncertainty from the new rules is not technical but an “elimination mechanism unrelated to their own capabilities.” Therefore, Chainflow suggests that instead of imposing rigid limits on the overall network share, it would be better to refine restrictions to the proportion of subsidies allocated within each data center, maintaining high-quality infrastructure while achieving more granular decentralization.

The new policy has less than a month before implementation, and it may further squeeze some small validators, reducing node numbers. The ultimate effect will depend on data from Ghost after May 1 and the foundation’s enforcement details.

3. The “On-Chain Nasdaq” competition

Currently, public chains are entering a contest to become the “On-Chain Nasdaq” that carries the global capital markets.

For traditional financial capital, “speed” and “cost” are important, but only if “security” and “compliance” are guaranteed. This means that the long-standing criticism of Solana’s node centralization will be amplified when talking to institutional narratives.

According to RWA.xyz, Ethereum still dominates in RWA asset value, with over $16 billion deployed on-chain, accounting for over 55%; BNB Chain ranks second with $3.5 billion and a 12.13% share; Solana ranks third with about $1.9 billion and a 6.65% share. Most large tokenized government bonds and private credit platforms for institutions are still deployed on Ethereum.

In terms of RWA assets, Solana’s wallet count and active addresses have now surpassed Ethereum. The growth of on-chain RWA users mainly stems from the mid-2025 launch of tokenized xStock stocks. Thanks to its speed and low cost, Solana has carved out a niche among retail users.

In this competitive landscape, both Ethereum and Solana are undergoing key upgrades in 2026 to address their shortcomings. Ethereum’s main upgrades, Glamsterdam and Hegotá, focus on making the mainnet faster and more efficient—through parallel execution, increasing Gas limits, optimizing transaction ordering, and lowering validator entry barriers to encourage more participation.

Solana’s focus is on improving stability and resilience. Besides the aforementioned validator policies, it is upgrading its consensus mechanism to reduce final confirmation time from seconds to milliseconds, and introducing a second independent client to prevent the entire network from collapsing if one software fails.

These two paths are converging toward the same goal. Currently, when institutional funds and RWA assets begin large-scale on-chain deployment, the market will favor more mature, stable, and predictable infrastructure. For Solana, the key challenge remains whether it can resolve structural issues like centralization and turn “fast” into “trustworthy fast.”