Someone tried to swap 400,000 GRT on a decentralised exchange recently. Not an unusual thing to want to do, in principle. They received a price impact warning of 45.26%.
To put that in human terms: they would have handed over roughly half the value of their position simply for the privilege of exiting it. At a spot price already sitting somewhere between $0.013 and $0.025 — near the lowest GRT has ever traded — that is the kind of number that makes you close the browser tab and go for a walk.
The tempting interpretation is that GRT has hit a bottom so hard and so inviolable that the market itself refuses to go lower. The less romantic interpretation, and the one we think is closer to the truth, is that the market simply can't go lower right now, because no one with meaningful size can actually sell. That's not a floor built on conviction. That's a floor built on illiquidity, and they are very different things.
We're the Lodestar team. We run an open analytics dashboard for The Graph network, which means we spend an unreasonable amount of time looking at indexer economics, subgraph data, and protocol mechanics. Here's our honest read on what's happened, what's being built, and what we think comes next.
The Invisible Hand Goes on Holiday
To understand the 45.26% figure, you need to understand what market makers actually do — and what happens when they stop.
In a healthy market, professional arbitrageurs sit between centralised exchanges (CEXs) and decentralised exchanges (DEXs). When the price of GRT on Uniswap drifts above Binance, they buy on Binance and sell on Uniswap. When it drifts below, they do the reverse. This constant, boring, mechanical activity is what keeps prices aligned across venues and, crucially, what keeps DEX liquidity pools topped up. It's the invisible hand, and it works precisely because you never notice it.
Then Binance delisted the GRTBTC trading pair. On the surface, losing one trading pair on one exchange sounds like a footnote. In practice, GRTBTC was a primary corridor for market makers operating between Bitcoin-denominated books and GRT spot. When it disappeared, the arbitrage math stopped working, and market makers — rational creatures who follow profitability rather than protocol loyalty — quietly redirected their capital elsewhere.
Coinbase had already delisted GRT perpetual futures in March 2026. The combination was a one-two to the market structure.
DEX liquidity pools that had been continuously refreshed by arbitrage began to thin. Concentrated liquidity positions drifted out of range and became inert. The result is the situation we're in now: a token trading near historical lows where attempting to exit any meaningful size produces slippage figures that look like typos.
"Can't go lower if no one can sell" sounds almost like a bullish thesis. It isn't. It means the price discovery mechanism is broken. When market makers return — if GRTBTC is re-listed, or new CEX presence develops — the market will restabilise. Whether that happens at current prices or lower depends on the fundamentals. Which brings us to the uncomfortable part.
How Did We Get Here
The liquidity crisis is acute, but the pressure on GRT has been building for years. It deserves honest accounting.
GRT peaked at roughly $2.84 in February 2021. The circulating supply since then has grown approximately 7.6 times over. That didn't happen by accident — it's baked into the protocol.
The Graph mints roughly 3% of total supply annually as indexing rewards — approximately $16–17 million per year in GRT terms at recent prices. This pays Indexers to run the infrastructure that serves data queries. In principle, a reasonable subsidy for bootstrapping a network. In practice, Indexers are running real servers with real costs, and most need to liquidate a portion of their GRT rewards to cover them. That's rational. In aggregate, it's continuous sell pressure.
The question is whether that sell pressure is offset by genuine demand for GRT. Here the numbers are uncomfortable. Organic query fee revenue in Q4 2025 came in at approximately $98,700 USD. Against $16–17 million being minted annually, that is less than one percent. The protocol is, at present, paying out far more than it earns from actual usage.
Staking participation fell to roughly 20% of circulating supply in Q2 2025 — reducing the proportion of supply locked away from circulation. Rising supply, modest query demand, lower staking participation. It's a tokenomics story that serious investors read and set down carefully.
We say all of this not to be gloomy, but because the path forward only makes sense if you understand clearly where the structural problems lie. Vague optimism doesn't serve anyone.
What's Actually Being Built
It would be dishonest to paint only the difficult picture without acknowledging what's been happening on the protocol side — because the answer is: quite a lot.
The Horizon upgrade, which shipped in December 2025, fundamentally reframed what The Graph is. The network moved from single-purpose subgraph indexing into a modular data infrastructure layer. New services now running or in development include Substreams for high-throughput data streaming, Token API for standardised token metadata, Tycho for DEX liquidity tracking, and Amp — an SQL analytics layer aimed at institutional users. The network serves 90+ blockchains and hosts 15,000+ active subgraphs. That ecosystem depth doesn't appear overnight.
The AI integration story is newer but genuinely interesting. The Graph has implemented an x402-compliant gateway, allowing AI agents to pay per query without API keys or subscriptions. Autonomous AI agents are poor candidates for traditional SaaS billing — they can't fill out forms or maintain credit cards. Pay-per-query micropayments are a natural fit, and The Graph is better positioned for this than most. Whether AI agent demand materialises at meaningful scale is still speculative, but unlike most GRT bull cases, this one doesn't rely purely on sentiment — it relies on a real structural fit.
On the roadmap, Direct Indexing Payments (DIPs) are slated for Q3 2026, allowing users to pay Indexers directly rather than relying entirely on protocol-minted inflation as the incentive mechanism. Combined with the Rewards Eligibility Oracle (REO), which ties reward eligibility to delivered value rather than mere participation, the protocol is structurally oriented toward fixing the fee-to-inflation gap.
We should be clear about competition too. Envio has benchmarked at 63 times faster than The Graph for equivalent workloads. The Graph's moat is not speed — it's decentralisation, verifiability, and ecosystem depth built over several years. That moat is real, but it requires tending.
What the Lodestar Team Actually Thinks
We've watched this protocol closely for a long time. Five honest takes, with the optimism and pessimism roughly in proportion to what the evidence supports.
The floor is illiquidity, not conviction. The 45.26% swap impact is a symptom of broken market structure, not proof of a hard support level. When market makers return, the price will likely become volatile in both directions. Anyone interpreting current prices as suppressed value waiting to spring should also hold the possibility that current prices reflect the absence of sellers rather than the presence of buyers.
DIPs are the metric that matters most in 2026. The Horizon upgrade was architecturally significant, but the number we're watching is query fees relative to issuance. If Direct Indexing Payments ship on schedule in Q3 2026 and fee revenue begins meaningfully outpacing minted inflation, that would be the first genuine value accrual signal GRT has had since inception. Watch that number.
AI agents are the most credible new demand vector. The x402 pay-per-query model is a genuine architectural match for how autonomous software agents operate. This is the bull case worth taking seriously, precisely because it doesn't depend on sentiment — it depends on whether AI agent workloads need indexed on-chain data at scale. The setup is right. The scale is still to be demonstrated.
The fee-to-inflation gap must close, and time matters. The REO and DIPs are moves in the right direction. But competitors are not standing still. Every quarter the gap stays wide is a quarter where protocol-aligned investors face a headwind and builders have alternatives. The roadmap is credible. Execution is the variable.
Liquidity fragmentation deserves serious attention. The GRTBTC episode should be a case study, not a footnote. A single trading pair's removal stranded meaningful DEX liquidity and broke the arbitrage mechanism that keeps markets functional. The Graph Foundation and community need a coherent liquidity strategy — one that doesn't leave the token's market structure dependent on any single exchange's product decisions.
Closing Thoughts
GRT at $0.013–0.025 with a 45.26% swap impact is not a comfortable place to be. We're not going to pretend otherwise.
What we can say, having watched the protocol up close, is that the work being done at the infrastructure level is real. Horizon was a genuine architectural shift. The AI agent integration is thoughtful. The direction toward fee-based revenue is, in intention, correct. Whether execution arrives fast enough, and whether the market structure can recover enough to let the token reflect any of that, are genuinely open questions.
The honest answer is: we don't know. Nobody does. What we think is that the next twelve months — particularly Q3 2026 — will tell us more about GRT's long-term trajectory than the previous three years combined. The pieces for a different story are being assembled. Whether they come together in time is the question we'll all be watching.
This post reflects the opinions of the Lodestar team based on publicly available data and our own analysis of The Graph network. It is not financial advice.