A bootstrapped team of 11 people, with zero venture capital, has built a protocol that generated $909 million in fees in 2025, processed $2.74 trillion in perpetual futures volume, executed $890 million in token buybacks, and overtook Coinbase — a $50 billion public company — in notional trading ...
"If we're going to build something that's really going to be a credibly neutral platform on which everyone else can build, then a really important principle is to sort of not have insiders." — Jeff Yan, Founder, Hyperliquid
A bootstrapped team of 11 people, with zero venture capital, has built a protocol that generated $909 million in fees in 2025, processed $2.74 trillion in perpetual futures volume, executed $890 million in token buybacks, and overtook Coinbase — a $50 billion public company — in notional trading volume. Hyperliquid is no longer an emerging DEX. It is an institutional-grade financial venue competing directly against the largest centralized exchanges on the planet.
This report delivers the definitive competitive analysis: Hyperliquid measured head-to-head against every major DEX (dYdX, GMX, Jupiter, Vertex, Drift) and every major CEX (Binance, Bybit, OKX, Coinbase) across six dimensions — volume, fees, buybacks, governance, market share trajectory, and technical architecture. The data reveals a protocol that has fundamentally altered the competitive landscape of crypto trading, while carrying material risks in decentralization, token unlock schedules, and regulatory exposure that the market has not fully priced.
Hyperliquid's volume trajectory over the past year defies every norm in exchange economics. A fully on-chain venue is now generating daily volumes that place it among the top five trading platforms globally — decentralized or centralized.
Full-Year 2025 Performance:
| Platform | 2025 Perps Volume | Daily Avg. (Recent) | Type | |----------|-------------------|---------------------|------| | Binance | ~$19.3T (est.) | $53B | CEX | | OKX | ~$6.8T (est.) | $18.6B | CEX | | Bybit | ~$5.2T (est.) | $14.2B | CEX | | Hyperliquid | $2.74T | $6.4B | DEX | | Coinbase | ~$1.4T | $3.8B | CEX | | dYdX | ~$450B (est.) | $2.8B | DEX | | Jupiter Perps | ~$294B (cumul.) | $1.1B | DEX | | GMX | ~$180B (est.) | $490M | DEX | | Vertex | ~$85B (est.) | $230M | DEX | | Drift | ~$42B (est.) | $115M | DEX |
The milestone worth highlighting: Hyperliquid processed nearly twice Coinbase's notional volume in 2025 — $2.74 trillion versus $1.4 trillion. For a protocol that launched its mainnet in late 2023, this represents a growth trajectory without precedent in exchange history. More striking, Hyperliquid now handles over 10% of Binance's perpetual volumes, up from negligible levels just 18 months ago.
In October 2025, Hyperliquid led all perpetual DEXs with $299 billion in monthly volume, capturing 22% of the combined DEX-CEX perpetual trade share. The broader DEX perpetuals market hit $903.56 billion that month — a tenfold increase year-on-year — with Hyperliquid accounting for the lion's share.
The platform's user growth has been equally aggressive: daily active addresses near 825,000, monthly active addresses at 3.65 million, and a total user base exceeding 1.4 million — all achieved without token incentive programs or liquidity mining campaigns.
Revenue is the metric that separates real businesses from speculative narratives. By this standard, Hyperliquid is one of the most capital-efficient financial operations in history.
Fee Comparison Matrix:
| Platform | 2025 Total Fees | Maker Fee | Taker Fee | Revenue Model | |----------|----------------|-----------|-----------|---------------| | Binance | ~$12.5B (est.) | 0.02% | 0.05% | Retained | | Coinbase | $7.2B | Variable | Variable | Retained | | Hyperliquid | $909M | 0.015% | 0.045% | 97% to buybacks | | Uniswap | ~$840M | 0.01-1% | 0.01-1% | LP fees + protocol | | dYdX | ~$215M (est.) | 0.01% | 0.05% | Retained | | Jupiter | ~$350M (est.) | Variable | Variable | Retained | | GMX | ~$120M (est.) | 0.05% | 0.07% | 30% to stakers |
Hyperliquid's fee structure undercuts most competitors. Base perpetual rates of 0.015% maker / 0.045% taker are lower than Binance's standard 0.02% / 0.05%, and significantly cheaper than most DEX alternatives. VIP tiers reduce fees further — down to 0.003% maker and 0.02% taker for the highest volume traders. Staking HYPE grants up to 40% fee discounts, exceeding Binance's 25% spot and 10% futures discounts via BNB holdings.
The HIP-3 growth mode for newly launched markets reduces taker fees by over 90% — from 0.045% down to 0.0045-0.009% — a mechanism designed to bootstrap liquidity on new pairs that no centralized exchange has replicated.
But the critical differentiator is not fee levels — it is fee allocation. While Binance and Coinbase retain the vast majority of fee revenue for corporate operations, Hyperliquid routes approximately 97% of all trading fees back into the protocol via the Assistance Fund, which executes continuous HYPE market buybacks. In Q4 2025, this translated to approximately $2.86 million in daily buyback spending. The platform does all of this with a team that recently expanded from 11 to 14 employees, yielding a revenue-per-employee figure of approximately $68 million — dwarfing every major exchange and most technology companies globally.
No protocol in DeFi history has returned capital to token holders at the scale of Hyperliquid. The Assistance Fund — the protocol-level mechanism that converts trading fees into HYPE purchases — has spent $890 million on buybacks, representing 46% of all token buyback spending across the entire crypto industry in 2025.
Buyback Metrics:
| Metric | Value | |--------|-------| | Total buyback spend (2025) | ~$890M | | HYPE tokens repurchased | 40M+ | | % of circulating supply repurchased | ~13% | | Average monthly buyback spend | $65.5M | | Peak monthly buyback (Aug 2025) | $110.6M | | Lowest monthly buyback (Mar 2025) | $39.1M | | % of fees directed to buybacks | ~97% | | Weekly buyback rate (Jan 2026) | ~$25M |
The mechanism is structurally elegant. Trading fees flow into the Assistance Fund. The fund automatically market-buys HYPE on Hyperliquid's own spot market. In December 2025, validators voted — with 85% support — to permanently burn the fund's accumulated holdings: approximately 37 million HYPE tokens valued at over $1 billion. This converted the Assistance Fund from a buyback-and-hold system to a buyback-and-burn system, making the deflationary pressure permanent.
Comparison with industry peers:
Uniswap's "fee switch" debate — whether UNI holders should receive protocol revenue — has dragged on for years without resolution. Maker's surplus buffer mechanism returns value through DAI burns at a fraction of Hyperliquid's scale. Curve distributes fees to veCRV holders. None of these mechanisms approach the sheer volume of capital returned: $890 million in a single year from a single protocol.
The self-reinforcing loop is clear: more trading volume → more fees → more buybacks → less supply → higher HYPE price → more attention → more users → more volume. Whether this flywheel is sustainable through bear markets is the question the market has yet to answer.
Hyperliquid's economic success masks a governance structure that is, by DeFi standards, remarkably centralized. This is the most important tension in the Hyperliquid thesis, and one that institutional investors must weigh carefully.
Validator Network:
The governance model operates through validator votes on Hyperliquid Improvement Proposals (HIPs). Two landmark votes have occurred:
Both votes demonstrated functional governance. Neither demonstrated decentralized governance. With the Hyper Foundation controlling 81% of staked shares, validator votes functionally reflect foundation preference. The community has raised legitimate concerns: closed-source code makes independent validation difficult; limited documentation forces validator operators to rely on centralized APIs; and the bridge architecture creates a single point of failure.
Comparison with competitors:
Hyperliquid's response to these criticisms has been pragmatic rather than ideological. The foundation argues that premature decentralization would compromise product velocity — and the results validate this claim. The platform has iterated faster than any competitor, shipping HyperEVM, expanding market listings, and implementing fee changes at a pace that fully decentralized governance would struggle to match. The question is whether this tradeoff remains acceptable as the platform's systemic importance grows.
Hyperliquid's market share story has two chapters. The first is dominance: peaking at 80% of DEX perpetual volume in August 2025, with consistent 70%+ share through mid-year. The second is fragmentation: declining to approximately 38% by November 2025 as credible competitors emerged.
Market Share Evolution (DEX Perpetuals):
| Period | Hyperliquid | dYdX | GMX | Jupiter | Aster | Lighter | Others | |--------|-------------|------|-----|---------|-------|---------|--------| | Q1 2025 | ~65% | ~12% | ~8% | ~7% | — | — | ~8% | | Q2 2025 | ~71% | ~8% | ~5% | ~6% | ~2% | ~1% | ~7% | | Q3 2025 | ~75% | ~5% | ~3% | ~5% | ~5% | ~3% | ~4% | | Q4 2025 | ~38% | ~4% | ~3% | ~5% | ~18% | ~12% | ~20% |
The newcomers are real. Aster DEX surged to $2.76 billion in daily volume, challenging Hyperliquid's $3.76 billion. Lighter reached $1.81 billion daily with $731 million in open interest. EdgeX emerged as a serious contender, earning S-tier rankings alongside Hyperliquid in industry analyst assessments.
Against CEXs, the picture is different. Hyperliquid is gaining, not losing. The DEX-to-CEX perpetuals ratio climbed from 2.1% in January 2023 to 11.7% by November 2025. Hyperliquid alone accounts for more than half of this on-chain migration. Binance's spot market share fell 3.13 percentage points to 25.0% in December 2025 — its largest monthly decline since July 2024.
The interpretation matters: Hyperliquid's DEX market share is declining because the category is expanding rapidly and attracting well-funded competitors. Its absolute volume and revenue continue to grow. The total perpetual DEX market has expanded from under $100 billion monthly to nearly $1 trillion monthly in the span of a year. Even at 38% share, Hyperliquid is processing more volume than it was at 80% share — the pie grew faster than its slice shrank.
Hyperliquid's performance advantage is not a product feature. It is an architectural decision that would require competitors years to replicate. The platform runs its own Layer 1 blockchain with a custom consensus mechanism, an on-chain order book, and a unified execution environment — a design philosophy fundamentally different from every other DEX.
HyperBFT Consensus: Derived from HotStuff and its academic successors, HyperBFT is a Byzantine Fault Tolerant consensus mechanism purpose-built for exchange workloads. Performance metrics: 0.1-second median latency, 0.5-second 99th percentile latency, and support for 200,000 orders per second. Every order, cancellation, trade, and liquidation settles on-chain with single-block finality.
Dual-Layer Architecture: State execution is split between HyperCore and HyperEVM:
How this compares:
| Feature | Hyperliquid | dYdX v4 | GMX | Jupiter | |---------|-------------|---------|-----|---------| | Architecture | Custom L1 | Cosmos appchain | Arbitrum L2 | Solana program | | Order book | On-chain CLOB | On-chain CLOB | AMM pools | Hybrid oracle | | Consensus | HyperBFT | CometBFT | Ethereum (via L2) | Solana PoH | | Latency | 0.1s median | ~1s | ~2-3s | ~0.4s | | Throughput | 200K orders/s | ~500 orders/s | N/A (AMM) | ~65K TPS (shared) | | Self-custody | Yes | Yes | Yes | Yes | | Own chain | Yes | Yes | No | No |
The critical insight: Hyperliquid is not a DEX running on someone else's blockchain. It is a blockchain built specifically to run an exchange. This allows the team to optimize every layer of the stack — consensus, networking, execution, matching — for a single use case. dYdX v4 made a similar architectural bet with its Cosmos migration, but Hyperliquid's custom consensus delivers meaningfully lower latency and higher throughput. GMX and Jupiter, bound to Arbitrum and Solana respectively, inherit the performance ceilings and congestion risks of their host chains.
The unified state model is perhaps the most underappreciated advantage. Because HyperCore and HyperEVM share consensus, developers can build applications that atomically interact with the order book — composability that no other exchange architecture offers at this level of performance.
Hyperliquid processed $2.74 trillion in perpetual volume in 2025, nearly 2x Coinbase's $1.4 trillion, making it the fifth-largest derivatives venue globally by notional volume — DEX or CEX.
The protocol generated $909 million in fees with a 14-person team — approximately $68 million in revenue per employee, a figure unmatched in financial services.
$890 million in HYPE buybacks (46% of all crypto buybacks in 2025) with 37 million tokens permanently burned after a December 2025 governance vote, removing ~13% of circulating supply.
Governance remains centralized in practice: 16 validators, 81% foundation-controlled stake, closed-source code. Functional, but far from the decentralization standard set by competitors like dYdX.
DEX market share declined from 80% to 38% as Aster, Lighter, and EdgeX emerged — but absolute volumes continued growing as the total category expanded tenfold.
Technical architecture is the durable moat: a custom L1, on-chain CLOB with 0.1s latency, 200K orders/second throughput, and unified HyperCore-HyperEVM state — years ahead of competitors architecturally.
Material risks include $12 billion in scheduled team token unlocks, regulatory exposure on permissionless 40x leverage products, and bridge security concerns with only 4 validators managing cross-chain assets.
Hyperliquid has accomplished something the crypto industry has attempted for a decade: building an on-chain trading venue that competes with centralized exchanges not on ideology, but on product quality, execution speed, and economic efficiency. The numbers — $2.74 trillion in volume, $909 million in fees, $890 million in buybacks, all without venture capital — describe a protocol that has found genuine product-market fit.
But the report card is not clean. A 16-validator network with 81% foundation-controlled stake is not decentralized by any honest measure. Closed-source node software contradicts the open-infrastructure thesis. A bridge secured by 4 validators is a systemic vulnerability for a platform processing billions daily. And $12 billion in team token unlocks scheduled through 2027 creates sustained selling pressure that even the Assistance Fund's prolific buyback machine may struggle to absorb.
The competitive matrix reveals a platform that is winning on execution and architecture, holding its own on fees, leading overwhelmingly on capital return, but lagging on governance and decentralization. For a protocol whose founder explicitly rejects the notion of insiders, the gap between philosophy and practice is the most important variable to watch.
Hyperliquid is not the future of trading. It is the present — processing $6.4 billion daily, burning $25 million weekly in buybacks, and forcing every major exchange to answer a question they would prefer to avoid: why does it take 3,700 employees to do what 14 people are doing on-chain?