Crypto spent the first half of 2026 getting a hard reality check. Binance Research said the market saw a broad on-chain contraction, not a rotation: DeFi TVL fell, major Layer 1 valuations sank, Layer 2 activity weakened, and security failures kept draining confidence and capital.
- DeFi TVL fell $43.4 billion, or 38%
- Major Layer 1 market caps dropped $246.5 billion
- Hacks surged to 207 incidents, with $972 million stolen
- RWAs and prediction markets were rare bright spots
That is not healthy “capital rotation.” That is money leaving, users pulling back, and a few pockets of real demand struggling to stand out in a generally ugly tape.
DeFi took one of the biggest hits. Binance Research said total value locked, or TVL, fell by $43.4 billion in the first half, a 38% decline. TVL is the amount of assets deposited into DeFi protocols, so it is often used as a rough gauge of how much trust and liquidity the sector still has. It is not a perfect metric. Prices can distort it. But a drop that large still says plenty about sentiment. For a broader look at how market stress has been hammering onchain finance, see the crypto contraction that erased $43.4B from DeFi in H1.
When users worry about exploits, mismanagement, or just better opportunities elsewhere, they stop supplying liquidity. That is how DeFi works when the music slows down. Capital gets cautious fast, and the whole stack feels thinner. Elegant theory, brutal execution.
The damage was not limited to DeFi. Binance Research said the combined market capitalization of Ethereum, BNB, Solana, Tron, Sui, and NEAR fell by $246.5 billion, or 42%. That is the kind of drawdown that does not look like a tidy sector rotation into “the next narrative.” It looks like broad repricing across the market’s most visible Layer 1 assets, as reflected in the broader Chain Rankings by TVL.
Security stayed ugly
The security picture was just as nasty. TRM Labs recorded 207 hacks in the first half of 2026 and $972 million stolen. That was more than double the 83 incidents seen in the same period a year earlier, even though total dollar losses were well below the $2.3 billion stolen in H1 2025.
That distinction matters. More hacks did not mean a repeat of last year’s giant thefts. Instead, the sector saw a record number of attacks without another monster loss dominating the scoreboard. That is not the same as “safer.” It just means the pain was spread across more incidents. A closer breakdown is available in H1 2026 Crypto Hacks Reach Record High as Losses Fall Below.
TRM said 125 of the 207 incidents were smart-contract exploits, meaning attackers targeted vulnerabilities in deployed onchain code. But the bigger problem, measured by dollars stolen, was not just bad code. TRM said infrastructure and operational compromises accounted for about 76% of stolen value, even though they made up only about 15% of incidents.
That is the part people should not gloss over. Crypto loves to talk about audits, but a slick audit does not save a protocol if the keys are handled carelessly, the operational setup is weak, or the surrounding infrastructure is a mess. The code is only one layer of the threat model. Humans remain the favorite attack surface, as usual.
April was the worst month, according to TRM’s breakdown, because two large North Korea-linked attacks on Drift Protocol and KelpDAO drove a sharp jump in losses. TRM said North Korea-linked activity accounted for about $643 million, or 66% of H1 stolen funds. And that figure covers hacks and exploits, not the whole universe of crypto crime, which also includes scams, phishing, fraud, and the usual parade of scammers in expensive hoodies.
Ethereum: more activity, less fee power
Ethereum sat at the center of several crosscurrents. The report said Ethereum ETF balances fell from more than 6 million ETH to 5.2 million ETH, while digital asset treasury companies increased their ETH holdings from 6 million to 7.7 million ETH, using figures sourced from SoSoValue and Blockworks as of July 1.
Those numbers do not necessarily overlap cleanly, so they should not be treated as one tidy net flow. They do, however, show that institutional ETH exposure is getting split across different wrappers and balance-sheet strategies. Some holders trimmed, others added. Crypto never seems to allow one simple story when two messy ones will do.
Ethereum’s fee structure also shifted. Binance Research said average gas prices fell 75% from 2025 after the gas limit was raised to about 60 million, while transaction count rose roughly 50%. In plain English: Ethereum got cheaper to use and busier at the same time. For the mechanics behind that, see the technical overview of Ethereum gas and fees.
That sounds great for users. It is less great for base-layer revenue. More throughput does not automatically mean more fees if each transaction costs less. Binance Research said chain revenue is “projected to fall 53% by year end if conditions persist.”
That is the awkward reality of scaling. Success can compress the very revenue you are trying to grow. Ethereum wants more activity, but not every form of growth is fee-maximizing. Useful, yes. Fat margins, not guaranteed.
Layer 2s had their own rough half. Total Layer 2 user operations fell about 77% between January and June, while Ethereum mainnet user operations declined only 9% over the same period. In June, Layer 2s collected around $15 million in fees but paid Ethereum just $66, 397 for data availability, the data layer Ethereum provides so L2s can verify and settle activity.
That is a pretty brutal mismatch. L2s are supposed to scale Ethereum, not magically make the base chain wealthy. They can absolutely expand access and lower costs, but if usage falls sharply, the economic model starts looking a lot less impressive. Cheap block space is useful. Empty block space is just vacant real estate with better branding.
Solana’s memecoin dependence got exposed
Solana had a different problem: activity cooled as memecoin trading weakened. Binance Research said Solana’s real economic value fell from $40 million in January to $14 million in June, a 64.5% drop. Real economic value, or REV, is a revenue-style measure that includes fees and tips captured by the network.
Pump.fun volume fell from $30 billion to $17 billion over the same period, and memecoins still made up 25% of Solana DEX volume in June. That is the double-edged Solana story in one snapshot. When speculative froth is hot, the chain looks unstoppable. When it cools, a big chunk of activity disappears with it. A useful comparison is how Solana closes the fee gap with Ethereum as trading volume fuels revenue surge when trading is actually roaring.
Memecoins can drive real usage, but they are also notoriously fickle. A quarter of DEX volume tied to joke tokens is not exactly the foundation of a majestic financial civilization. It is more like a very fast carnival with venture funding.
BNB and RWAs showed some actual resilience
BNB stood out as the only deflationary major Layer 1 in the report, with a 5.05% annualized burn rate. In simple terms, that means BNB’s supply is being reduced through token burns at an annualized pace. That matters because supply contraction can support token economics, especially when the broader market is weak.
BNB’s tokenized real-world asset, or RWA, market value rose 107% to $3.8 billion in the first half. RWAs are traditional assets represented onchain, such as stocks or other financial instruments. This is one of the cleaner crypto narratives because it is not built entirely on speculative reflexes. It tries to bring real assets into programmable, tradable form. One example of that push is Coinbase’s CUSHY tokenized credit fund, which aims squarely at institutions rather than meme traders with a caffeine problem.
The broader tokenized RWA market also grew, from about $22 billion in January to roughly $34 billion by mid-July. BNB Chain led tokenized equity DEX volume, and tokenized equities reached 4% of Solana DEX activity in June.
That is still small, but it is meaningful. RWAs are one of the few corners of crypto where the pitch is not just “number go up.” The harder question is whether secondary liquidity, user demand, and regulation can support something durable instead of another shiny pilot that gets forgotten once the press releases dry up. Even geopolitics is being dragged into the mix, as shown by Russia legalizing Bitcoin, Ethereum, and Solana with strict caps and harsh penalties.
Prediction markets kept gaining traction
Prediction markets were the other standout. Monthly notional volume reached $51.6 billion in June, up 86% from January. Notional volume is the total face value of trades, not profit or revenue, so it shows activity rather than earnings. Big difference, because crypto loves to confuse raw volume with actual usefulness.
Kalshi recorded about $33 billion in June, while Polymarket recorded about $14.5 billion. Together, the two platforms accounted for 92% of June volume. Non-sports activity across both rose 136%, and the World Cup pushed Polymarket past $5 billion in tournament volume.
That is a real sign of demand, but it also comes with a warning label. Event-driven spikes are not the same thing as sticky adoption. A World Cup can pull in a flood of users. It cannot prove they will stay when the tournament ends and the next attention cycle rolls in.
That is why the second half looks like a test. Binance Research said it will “test” whether tournament-driven users remain active and whether regulation and settlement systems can support continued growth. Fair enough. Crypto has a long history of mistaking temporary obsession for permanent product-market fit.
What this means going forward
The first half of 2026 does not read like a healthy sector rotation. It reads like a market that got smaller, more defensive, and more selective at the same time. DeFi lost liquidity, major L1s got repriced, L2 activity weakened, Solana leaned too hard on memecoin churn, and security failures kept scaring off capital.
At the same time, not everything was a wash. RWAs and prediction markets showed there is still real demand for a few crypto-native use cases that are not just speculative junk dressed up as innovation. That matters. It is the difference between a sector with a pulse and a sector with a press release budget.
The bigger question now is whether those stronger niches can keep growing once the easy money and event-driven hype fade. If they can, H2 could show something more durable than narrative churn. If they cannot, the market is still mostly running on speculation, fragile liquidity, and too many people pretending that a quarterly chart is a business model.
Key takeaways
-
Was this just capital rotating inside crypto?
No. Binance Research said the first half of 2026 showed a broad on-chain contraction, not a rotation. DeFi, major L1s, and L2 activity all weakened together. -
Did security get better?
Not really. TRM Labs tracked 207 hacks and $972 million stolen, with more incidents than a year earlier. The good news is that losses were more dispersed than in H1 2025, when a few mega-thefts did the heavy lifting. -
What got hit the hardest?
DeFi and speculative trading-heavy chains took the biggest bruises. DeFi TVL fell $43.4 billion, Solana’s REV dropped to $14 million in June, and Layer 2 usage slumped hard. -
Which sectors showed real strength?
Tokenized RWAs and prediction markets. RWAs grew from about $22 billion in January to roughly $34 billion by mid-July, while prediction market volume hit $51.6 billion in June. -
Is Ethereum’s scaling progress paying off?
Not cleanly. Lower gas fees and higher transaction counts are good for users, but they can squeeze base-layer revenue. Binance Research said chain revenue could fall 53% by year end if current conditions persist. -
Can Solana grow without memecoin froth?
That is the real test. Solana still has meaningful activity, but the drop in REV and Pump.fun volume shows how dependent a lot of the chain’s traffic still is on speculative trading cycles. -
What should H2 prove?
Whether prediction market users stick around after the big event cycle, whether RWAs build lasting liquidity, and whether better security can stop capital from bleeding out through hacks and operational blunders.
Please provide the HTML content you would like me to do the impossible with next time might be the most unintentionally perfect request in crypto, because apparently even the market itself needs formatting help.
