Anatoly Yakovenko

Yakovenko’s endgame is a single Solana L1 so efficient that running an app on its own app chain becomes “economically irrational,” not a prediction that Ethereum disappears. He allows for Solana processing 90% of transactions while Ethereum retains 90% of TVL, because each occupies a different point on the scalability trade-off curve. Whether either architecture wins remains “up to the PMF gods.”

SOL’s value accrual ultimately depends on sustained demand for blockspace, not a burn mechanism. Yakovenko calls EIP-1559’s deflationary story a “nice meme”: higher fees can reward validators and increase staking demand regardless of whether tokens burn. The complication for SOL is that localized fee markets prevent one congested application from repricing the entire chain, while ever-cheaper hardware keeps expanding supply.

Firedancer’s largest contribution is client diversity and network safety, not dramatically cheaper validators. Even if its code is 10x more efficient, compute represents only about one-quarter of a validator’s total machine, bandwidth and power costs, so Yakovenko estimates perhaps a 20% overall reduction. More importantly, validators can fail over between clients, turning a potentially catastrophic safety failure into a recoverable liveness problem: “I will be able to sleep at night.”

A US stablecoin bill is Yakovenko’s clearest adoption catalyst, with a “massive dramatic shift” into crypto-based finance possible over the following two to five years. He argues that crypto payment APIs are easier: generating an address and private key is simpler than connecting to PayPal or Stripe. He expects bank accounts to accept USDC deposits across multiple networks. “If and when we get a stablecoin bill … I think it’ll explode.”

Solana’s outage response targets the specific hotspot design failure behind earlier incidents. NFT bots once drove 20–40 Gbps of traffic; QUIC added flow control, stake-weighted quality of service preserved access, and localized fee markets capped one hot account at 12 million compute units inside a 48 million-unit block. The system currently supports roughly four simultaneous hotspots and, Yakovenko argues, declining hardware costs could make that eight within two years.

The strongest application evidence remains promising but small. Phoenix’s volume-to-TVL ratio was said to be roughly 30x Uniswap’s, yet daily volume was only $1 million–$5 million—“tiny things, but … not zero”—while Ethereum retains capital that holders simply do not want to sell or bridge. Helium’s $5 mobile plan and Hivemapper’s mapping network may be cleaner consumer proofs than another trading venue.

The unresolved investment question is whether demand can outrun relentless blockspace expansion. Star Atlas was already generating more daily transactions than Ethereum L1 or Polygon, but Yakovenko distinguishes “superfluous usage” from activity that genuinely requires a blockchain. Solana’s unusually durable community helps sustain the experiment—“find your niche, your nerds”—but he openly says he does not yet know which use case creates exponential demand.

🔗 Original source & video: Solana’s End Game with Anatoly Yakovenko