Solana processed 1.01 billion non-vote transactions during the week ending August 2, 2026, its highest weekly total on record. On August 4, the network set a single-day record of 169.9 million non-vote transactions. Both milestones arrived within days of SIMD-0286, a mainnet upgrade activated Jul...
"That's a big one." — Anatoly Yakovenko, Solana co-founder, reacting to Solana processing 44% of global blockchain transactions in March 2026
Solana processed 1.01 billion non-vote transactions during the week ending August 2, 2026, its highest weekly total on record. On August 4, the network set a single-day record of 169.9 million non-vote transactions. Both milestones arrived within days of SIMD-0286, a mainnet upgrade activated July 29 that raised the per-block compute ceiling from 60 million to 100 million compute units — a 66% capacity expansion.
The usage data tells one story. The revenue data tells another. Network fees fell 78% year-over-year in Q2 2026, to $51 million for the quarter. SOL trades near $73.67, down roughly two-thirds from its peak. TVL has contracted 56% from its August 2025 high to approximately $4.8 billion. The divergence between transaction volume and economic extraction is the defining tension in Solana's 2026 trajectory.
This report examines the SIMD-0286 upgrade, the transaction record that followed, the structural reasons revenue has decoupled from activity, and what the gap implies for the network's economic sustainability.
On July 29, 2026, at slot 435,888,000 (epoch 1009), Solana activated SIMD-0286, a proposal authored by Lucas Bruder of Jito. The change raised the maximum compute units per block from 60 million to 100 million. Block times remained at 400 milliseconds.
The upgrade became viable after more than 70% of staked SOL enabled XDP, a kernel-bypass networking mode that lets validators transfer data faster by bypassing operating system layers. The Solana Foundation's assessment determined that XDP coverage above 70% of stake provided sufficient margin to absorb the additional execution load.
According to the Solana Foundation, 11.2% of blocks produced between July 22, 2025 and the activation date consumed at least 56 million compute units, indicating regular pressure against the previous 60 million ceiling. The upgrade carried no breaking changes for developers or indexers.
What the upgrade did not change: the per-account writable state cap held at 12 million compute units, and the 100 MB limit on block account data size remained intact. Under the old 60M limit, one account's maximum allocation represented 20% of block capacity. Under the new 100M ceiling, the same 12M-CU cap represents only 12%.
The record arrived quickly. During the week of July 27 to August 2, 2026, Solana processed 1,012,226,009 non-vote transactions, according to Blockworks data. Non-vote transactions exclude the validator consensus messages used to maintain the network, making the figure a direct measure of user and application activity.
On August 4, the network set a single-day record of 169.9 million non-vote transactions, according to data cited by Cryptopolitan and CoinReporter. That figure represents a 154% increase from the 66.9 million single-day record set during the PENGU token launch in late 2025.
For context, Solana processes approximately 70 times more daily transactions than Ethereum's layer-1, which handles roughly 2.89 million transactions per day. Solana's average transaction cost sits at approximately $0.00025, compared to $0.50–$3.00 for a standard Ethereum L1 transfer. This fee differential explains much of the volume disparity but also underpins the revenue problem.
The first quarter of 2026 had already set records: 25.3 billion transactions processed, the highest quarterly figure in the network's history. June 2026 extended the streak to nearly 4 billion monthly transactions — another all-time high. Lifetime cumulative transactions have surpassed 119 billion.
Here is where the data diverges. In Q2 2026, Solana generated $51 million in total network fees, a 43% decline from Q1 and a 78% drop year-over-year, according to AMBCrypto. Daily network revenue dropped 79% to $314,700 by early March 2026, recovering only modestly since.
The collapse has a specific cause: the memecoin economy that drove Solana's 2025 fee surge evaporated. Pump.fun, the memecoin launchpad that became Solana's largest fee generator, saw daily protocol fees fall approximately 83%, from $4.8 million per day in January 2026 to $800,000–$875,000 per day by June 2026, according to data compiled by Spotted Crypto.
SOL's price reflects the revenue erosion. The token trades at approximately $73.67 as of August 8, down from $117 at the end of January 2026. In June, SOL briefly touched $61, its lowest level since 2024. Monthly active users have fallen to 34.1 million, a two-year low.
Yet application-layer revenue tells a different story. According to 21Shares research authored by Maximiliaan Michielsen, applications built on Solana earned approximately $148 million per month in H1 2026, an annualized pace of $1.8 billion. The network settled approximately $3 trillion in stablecoin volume in 2025, and its share of global DEX spot volume has reached approximately 50%, or $425 billion monthly. The economic activity exists. The question is where value accrues — to the base layer, or to the application layer sitting on top of it.
The transaction records also reflect infrastructure improvements beyond SIMD-0286. Firedancer, the independent validator client built by Jump Crypto, went live on mainnet in September 2025 and now runs approximately 14% of staked SOL, with an additional 26% on its Frankendancer hybrid variant. More than 20% of active validators use some form of the client.
Operator data from the deployment period shows measurable gains: Firedancer validators achieved 18–28 basis points improvement in skip rate reduction, 15% fewer missed voting credits, vote latency of approximately 1.002 slots, and fuller blocks averaging 47 million compute units versus 44.8 million under the original Agave client. Lab benchmarks for the full Firedancer implementation have demonstrated 1 million or more transactions per second, though real mainnet sustains far less. The operational value lies in headroom during traffic spikes, not peak theoretical throughput.
The dual-client architecture also provides resilience. A bug in one client implementation no longer threatens network-wide outages — a vulnerability that plagued Solana's reputation through multiple outages in 2022–2023.
The composition of Solana's transaction activity has shifted materially in 2026. While memecoin trading has declined, tokenized real-world assets have expanded.
Solana's tokenized gold market cap grew 689.1% in the twelve months through August 2026, averaging 18.8% month-over-month, according to Birdeye Data's H1 2026 Solana report. That pace exceeded BNB Chain by more than 2x and Ethereum by approximately 4.6x. Gold perpetuals on Solana peaked in May 2026 at $7.19 billion in monthly volume. The growth coincided with physical gold crossing $5,000 per ounce amid geopolitical instability.
Total RWA supply on Solana reached $3.4 billion, up approximately 230% year-over-year, with over 290,000 holders. The network processed its largest weekly equity volume of $1.36 billion and recorded a single-day RWA transfer of $1.49 billion. According to 21Shares data, Solana captures approximately 97% of on-chain equity trading volume.
Stablecoin supply on Solana stands at approximately $15.6 billion, near record levels, with a monthly dollar turnover rate of approximately 40x and over $500 billion in 30-day transfer volume. USDC holds 44.76% dominance on the network.
This shift matters economically. Memecoin trading generated high fees per transaction through priority fees and MEV. Stablecoin transfers and RWA settlement generate high volume at near-zero per-unit cost. The network's fee architecture, designed for volume over per-unit extraction, amplifies this dynamic.
SIMD-0286 expanded room for parallel, unrelated transactions. It provided no relief for transactions competing within the same account. As CryptoSlate reported, "transactions targeting the same writable state get no extra room" despite the overall capacity expansion.
This matters most during periods of market volatility, when traders submit competing transactions against the same liquidity pools or order books. The Solana Foundation's own upgrade materials offer no post-activation before-and-after measurements for transaction inclusion rates, fees, block propagation, or replay times.
A subsequent proposal, SIMD-0306, addresses this by replacing the hard 12M-CU per-account cap with a ceiling of 40% of the block's compute limit — raising the effective per-account cap to 40M compute units. This change is expected to provide significant relief for hot accounts like popular DEX pools during congestion events.
The transaction volume records and infrastructure upgrades are precursors to Solana's largest planned protocol change: Alpenglow, a complete replacement of the consensus layer. Approved by validators with 98.27% support in September 2025, Alpenglow replaces both Proof of History (PoH) and Tower BFT with two new systems: Votor (a voting protocol) and Rotor (a block propagation system).
The target: reduce transaction finality from 12.8 seconds to 100–150 milliseconds while maintaining consensus under conditions of 20% adversarial and 20% offline stake. Critically, Alpenglow eliminates on-chain vote transactions, which currently consume approximately 75% of block space.
If activated — mainnet deployment is targeted for late 2026 via the Agave 4.1 release, following community testing and security audits through Q4 — the effect on usable capacity would be substantial. Removing vote transaction overhead from blocks that now process 100 million compute units of mixed user and validator traffic would effectively multiply available user capacity.
At Consensus 2026, Yakovenko argued that the industry may be overestimating how much raw scale it ultimately needs, suggesting focus should shift from transaction volume toward "speed, reliability, and economic value."
Solana's transaction records are real. So is the revenue collapse. The two facts coexist because the network's activity composition has changed. Memecoin trading — high-fee, high-MEV, speculative — has been replaced in part by stablecoin settlement, RWA transfers, and tokenized commodity trading, all of which generate volume at negligible per-unit cost.
The network's infrastructure is measurably better than it was 12 months ago: Firedancer provides client diversity and performance gains, SIMD-0286 expands block capacity, and Alpenglow promises to recover 75% of block space currently consumed by validator votes. But infrastructure capacity is not revenue.
The central question for Solana in H2 2026 is whether the new activity composition — stablecoins, RWAs, on-chain equities — can generate sustainable fee revenue, or whether the network's economic model remains structurally dependent on speculative trading cycles. The data is inconclusive. Transaction records provide capacity validation, not economic validation.