Behind Solana's Summer Rally - Speed Claims, TVL Growth and Centralization Concerns

As ETH gas fees climbed alongside the NFT boom, traders began looking elsewhere for cheaper and faster execution — and Solana was one of the main beneficiaries. With $SOL rallying and the network pitching itself on low-cost, high-speed transactions, funds bridged from Ethereum via the Solana Wormhole rose 82% over a single week.

Solana's total value locked doubled in the first week of September, climbing from $3.53 billion to $7.78 billion, according to DappRadar data — though that growth tracked the rising SOL price far more than it reflected any real increase in on-chain activity.

Much of the inflow into Solana has been driven by the network's claimed technical edge over Ethereum. Founder Anatoly Yakovenko's whitepaper describes a design built around Proof of History, which streamlines Proof of Stake block confirmation by letting validators verify directly against the current leader, cutting out much of the validator-to-validator communication that traditional PoS networks require.

Solana advertises a theoretical ceiling of 50,000 transactions per second, though live figures on the Solana Beach dashboard run well below that. Over a recent one-hour sample, average block time sat at 0.56 seconds with throughput around 2,000 TPS — still far ahead of Ethereum's roughly 30 TPS. That said, most of that Solana volume is generated by the validator network itself, since the architecture requires all validator coordination traffic to run through mainnet.

Solana's Gulf Stream protocol eliminates the mempool entirely, instead assigning each transaction to a future block and forwarding it to validators before the prior block has even finished confirming. In principle, no mempool means no frontrunning — at least for now.

Ethereum and Solana are inevitably being measured against each other, since both are chasing the same underlying goal: higher sustainable throughput. Their paths there diverge sharply, though. Ethereum's approach leans on protocol-level upgrades and workarounds — ETH 2.0 and layer-2 networks — while Solana instead pushes the burden onto its validators, who are expected to continually upgrade their hardware as the network grows. The implication, per Moore's Law, is that Solana's throughput should organically increase over time as hardware improves.

Technical differentiation aside, the application layer built on top of Solana has so far looked familiar rather than novel — largely rushed clones of proven Ethereum projects, including NFT collections like "Degenerate Ape Academy," "Solana Monkey Business," and "SolPunks." Even FTX joined in, launching a cross-chain NFT marketplace to support Solana-based NFT trading — a move whose commercial logic isn't hard to guess at.

Centralization is arguably Solana's bigger vulnerability: roughly 48% of the total token supply was allocated to venture investors and insiders. Promises that the network will decentralize further over time carry less weight given how the token has been marketed so far. As Ryan Berckmans put it on Twitter:

Vitalik and other high-level people at the EF focus almost exclusively on the tech roadmap and dev programs, and not on biz dev with specific partners who, among other things, sell tokens to retail.

The throughput numbers themselves also invite scrutiny. As @fpieper wrote in the Lobster Chat Telegram group:

the majority of transactions are voting "transactions" (which are part of Solana's consensus). Basically, Solana is artificially pumping their numbers.

To bring some numbers - last 3 minutes based on solanabeach.io (total 371974 transactions):

Vote 311182 => 84%

Serum 37298

System 18219

SPL Token 4106

Memo 773

SPL 309

Swap 63

Stake 24

Overall, this means Solana shows 2000 TPS in their explorer, but 1680 TPS are consensus votes and not real transactions.

Only 300 TPS are real transactions.

Projecting that to [the network's theoretical capacity of] 65K would result in only 10K real TPS.

Also your network starts to be congested if you are reaching around 50% of capacity which would be 5K real TPS for Solana.

The picture that emerges is of a network whose headline TPS figures are dominated by internal consensus voting rather than genuine user activity, leaving real usable throughput far lower than the marketing suggests.

On decentralization specifically, @Hsaka has noted:

In its current form, Solana is just as centralised as BSC. The top 12 validators have the cumulative power to halt the network.

That framing didn't stop Binance Smart Chain from gaining significant traction despite similar objections from maximalists, suggesting centralization concerns alone may not be enough to slow adoption of a chain that's cheap and fast. At the moment, genuinely decentralized infrastructure still tends to come at a cost premium that puts it out of reach for many users — leaving an open question of just how much the market is actually willing to pay for decentralization when a faster, cheaper, more centralized alternative is sitting right there. With SOL's chart showing signs of cooling and attention increasingly drifting toward Arbitrum, where Solana's momentum goes from here remains to be seen.