Solana
Anatoly Yakovenko and the Solana Vision
The story of Solana begins not in a garage or a dorm room but in a Qualcomm engineer's late-night fever of insight. Anatoly Yakovenko, a senior staff engineer at Qualcomm who had spent years working on operating systems and distributed systems for mobile devices, woke up one night in 2017 with an idea he couldn't shake: what if you could use time itself as a trustless cryptographic primitive?
By the time the sun came up, he had sketched the outline of Proof of History — a mechanism for encoding the passage of time into a blockchain without requiring validators to trust each other's clocks. That sketch would eventually become Solana, a blockchain that by 2021 was processing more transactions per day than any other programmable blockchain in existence.
The Qualcomm Background
Yakovenko's decade at Qualcomm was unusually relevant preparation for building a high-performance blockchain. Qualcomm builds the processors that run most of the world's smartphones, and the engineering challenges of mobile systems — limited power budgets, real-time constraints, distributed coordination between hardware components — overlap significantly with the challenges of blockchain design.
At Qualcomm, Yakovenko worked on operating systems and firmware, the low-level software that controls how hardware resources are allocated. He developed an intuition for performance bottlenecks and a habit of thinking about systems at the hardware level — where is the true bottleneck, the memory bus, the CPU pipeline, the network? When he turned his attention to blockchains, he applied the same diagnostic lens.
His conclusion: the bottleneck in existing blockchains was not raw computation but coordination. Validators spent enormous amounts of time agreeing on the order of transactions before they could execute them. If you could remove that coordination overhead — or reduce it dramatically — you could achieve throughput that was orders of magnitude higher than Bitcoin or early Ethereum.
The Proof of History Insight
The fundamental problem with distributed systems is that independent nodes have no shared clock. Without synchronized time, a network of validators cannot agree on the order in which events occurred without exchanging messages about ordering — and those messages take time, creating latency and reducing throughput.
Yakovenko's insight was to create a verifiable, decentralized clock using a hash function. Specifically, Proof of History (PoH) works as follows: a designated validator (the leader) runs a cryptographic hash function in a continuous loop, feeding the output of each computation as the input to the next. The sequence of outputs constitutes a verifiable timeline: anyone can prove that a particular hash output appeared at a particular position in the sequence by verifying the chain of computations.
When transactions are submitted, they are stamped with a hash from the PoH sequence. This stamp is a cryptographic proof that the transaction was submitted at a particular moment in the leader's timeline — before certain hashes and after others. Validators don't need to discuss ordering; the order is embedded in the transactions themselves via their PoH stamps.
What PoH Actually Does
It's worth being precise about what PoH achieves and what it doesn't. PoH is not a consensus mechanism — it does not determine which blocks are valid or prevent forks. It is a clock: a mechanism for establishing temporal ordering of events in a way that everyone can verify cryptographically.
Solana's actual consensus uses a variant of Proof of Stake (Tower BFT, based on PBFT) to finalize blocks. PoH feeds into this consensus by providing pre-established ordering, allowing the consensus protocol to skip the ordering negotiation phase and jump directly to voting on correctness. This is where the throughput gain comes from: Tower BFT doesn't need to discuss "which transaction came first?" because PoH has already answered that question cryptographically.
The Founding Team
Yakovenko's first call was to Greg Fitzgerald, a former colleague from Qualcomm and longtime friend. Fitzgerald had left Qualcomm to work at a startup but was immediately interested in what Yakovenko was describing. He wrote the first implementation of the Solana prototype — producing a working proof of concept in a few weeks that demonstrated the PoH concept could be implemented efficiently.
Stephen Akridge, another Qualcomm colleague, joined to work on GPU-accelerated signature verification. This focus on hardware optimization was characteristic of the Solana founding team: rather than ignoring hardware constraints and hoping software could compensate, they designed the protocol to take advantage of modern hardware — multi-core CPUs, GPUs, and high-bandwidth SSDs.
Raj Gokal joined from a consumer health startup to lead operations and strategy. His background was less technical than the engineering founders' but brought crucial perspective on product development and fundraising. The team was rounded out by additional hires from Qualcomm and other engineering-focused companies.
The Whitepaper and Early Development
Yakovenko published the Solana whitepaper in November 2017, describing Proof of History and its implications for throughput and latency. The paper made bold claims: by eliminating the coordination overhead of consensus-before-ordering, Solana could theoretically achieve hundreds of thousands of transactions per second on commodity hardware.
These claims attracted skepticism. Hundreds of thousands of transactions per second was a benchmark that made even permissioned blockchains — systems run by a single company with full control over the validator set — look slow. For a public, permissionless blockchain, the numbers seemed implausible.
The team spent 2018 building toward a testnet, working through the significant engineering challenges of making PoH work at scale. They secured seed funding in 2018 and launched a public testnet in 2019. The initial benchmarks, while not matching the theoretical maximums, were genuinely impressive — the network was processing tens of thousands of transactions per second, faster than any comparable public blockchain.
Eight Key Innovations
The Solana whitepaper and subsequent technical documentation describe eight core innovations that, taken together, explain the performance:
Proof of History establishes verifiable time ordering. Tower BFT is the PoH-optimized consensus protocol. Turbine is a block propagation protocol inspired by BitTorrent that breaks blocks into packets and distributes them across the network efficiently. Gulf Stream is a mempool-less transaction forwarding protocol that pushes transactions to validators before the current block is finished. Sealevel is Solana's parallel smart contract runtime that executes non-conflicting transactions simultaneously using the GPU. Pipelining applies the hardware concept of pipeline processing to transaction validation — different stages (fetch, verify, execute) operate on different transactions simultaneously. Cloudbreak is Solana's horizontally-scaled accounts database optimized for concurrent reads and writes. Archivers distribute the historical ledger across a network of storage nodes.
The combination of these eight components is what allows Solana to achieve its claimed throughput. No single innovation is revolutionary in isolation — PoH is novel, but Sealevel's approach to parallel execution is similar to ideas in database systems, and Turbine borrows from BitTorrent. The innovation is the integration: a complete system co-designed from the hardware up for maximum throughput.
The 2021 Bull Run and Beyond
Solana's real public emergence came in 2021. As Ethereum's gas fees soared during the DeFi and NFT booms, users and developers began looking for alternatives. Solana offered dramatically lower fees and faster finality. A wave of DeFi protocols, NFT marketplaces, and gaming projects launched on Solana in 2021, and the SOL token rose from under $1 at the start of the year to over $200 by November.
The rapid growth also revealed challenges. The network experienced several significant outages in 2021 and 2022, primarily caused by bots flooding the transaction queue with spam. The bottleneck was not raw computational throughput but the transaction processing pipeline, which could be overwhelmed by low-quality transactions before it got to legitimate ones. Subsequent protocol upgrades addressed many of these issues with quality-of-service mechanisms and fee market improvements.
The collapse of FTX in November 2022 — whose founder Sam Bankman-Fried had been one of Solana's most prominent supporters — hit the Solana ecosystem hard. SOL fell by 95% from its peak. Yet the developer community remained active, and the network continued operating. The recovery through 2023 and 2024 demonstrated that the project had a user and developer base that extended beyond any single investor's backing.
Yakovenko's insight from that sleepless night in 2017 — that time itself could be a cryptographic primitive — produced one of the most technically distinctive blockchains ever built. Whether the performance trade-offs it makes are the right ones for long-term decentralized infrastructure remains actively debated, but the engineering achievement represented by a working, public blockchain sustaining tens of thousands of transactions per second is real and significant.