“AI tokens are up, so adoption is up.”
Which paid workloads, repeat customers or accepted outputs changed in the same period? What evidence would contradict the claim?
Separate demand from subsidiesAI & Web3 market research: follow the evidence
Follow a small set of markets. Keep the evidence behind your view.
Eight deliberately selected assets, not the whole market. Selection is editorial, not a recommendation.
Price retrieval: fresh · 2026-09-27T07:02:11.312Z. 7/8 provider prices within 15 minutes. History: fresh.
Start with what changed
TAO +6.05% is the largest absolute 1-day sampled move in this selected set.
Price movement identifies a question to investigate. It does not explain the cause.
| Watch | Asset | Price / freshness | 1d ≈ | 7d ≈ | 7-day path | Investigate |
|---|---|---|---|---|---|---|
| BTCBitcoin | $84,489.86fresh · 07:00:40 UTC | +0.72% | +5.25% | |||
| ETHEthereum | $2,706.43fresh · 07:00:40 UTC | +0.79% | +5.16% | |||
| USDCUSD Coin | $0.999838fresh · 07:00:40 UTC | -0.01% | +0.01% | |||
| USDTTether | $0.999759fresh · 07:00:40 UTC | -0.01% | +0.02% | |||
| TAOBittensor | $328.97fresh · 07:00:40 UTC | +6.05% | +30.11% | |||
| RENDERRender | $2.03stale · 06:46:50 UTC | — | — | |||
| FETFetch.ai / FET | $0.244872fresh · 07:00:40 UTC | +1.58% | +44.41% | |||
| AKTAkash | $0.703447fresh · 06:51:50 UTC | +1.59% | +28.57% |
Data: DefiLlama current prices · Dated history samples. Missing or stale data never becomes a signal. JSON with exact comparison dates.
Official updates to investigate
fresh. Protocol-wide context, not an explanation of token-price moves. Dates below are publication dates.
- Pull request for mintlify/docs-update-1790356082356 (#3584)x402 Foundation repository · 2026-09-25 · Code commit, not a release
- feat(svm): add batch settlement scheme (#3164)x402 Foundation repository · 2026-09-25 · Code commit, not a release
- fix(python): cap solana below 0.40 in the svm extra (#3573)x402 Foundation repository · 2026-09-25 · Code commit, not a release
Choose a question worth researching
Select “Review” beside an asset for its source and evidence checklist.
Make a payment cost explicit
Put fixed fees, percentage fees, FX spread, retries and settlement cost on one card so a teammate can recalculate it.
Scenario math only. It uses your assumptions, not a live quote, trading signal or payment recommendation.
What changed since I last looked?
Tick the assets you follow, then save a dated baseline. On your next visit, compare new observations with that exact snapshot.
Saved only in this browser, for up to 30 days. No login, wallet connection or background alerts.
No saved baseline loaded.
Before you repeat a market claim
Use these editorial questions to prepare a discussion with evidence. They are not community posts or AI-generated market explanations.
“This compute network is cheaper.”
Is the workload comparable after failures, data transfer and setup? Can someone reproduce the calculation?
Compare accepted-output costs“A protocol announcement means it works.”
Was the change released? Which version is actually deployed? Is there a test result rather than only a roadmap?
Check the implemented versionA view you can revisit
Write your claim, then try to disprove it. Save a private note or export the evidence brief to discuss with someone else or an AI assistant.
Your text stays in this tab unless you choose Save or Export. Saving uses this browser only; we do not receive or publish the note. The date is a manual review target, not a scheduled notification.
Generate a proof of research
After you record sources, counterevidence and a review date, generate a small proof of research work. It only proves that this research commitment was generated; it is not a token, balance, investment return or “correct answer”.
Private notes never enter the link; it contains only the asset, date, commitment hash, nonce and verifiable SHA-256 result. Anyone can recalculate it.
How to read this research board
Prices refresh on request after five minutes; history after one hour. A source price older than 15 minutes is marked stale individually. The 1d and 7d columns compare to daily source samples within two hours of the target time; exact endpoints are included in the JSON. Sparklines use separate vertical scales and do not show intraday highs or lows. Missing data is never zero.
The coverage list is a small editorial research set. It omits market cap, order books, volume and most assets. It is not a sector index, price forecast or evidence that a headline caused a move. For broad market screening, Binance markets, OKX markets and DefiLlama are established alternatives.
Discuss elsewhere using Binance Square or OKX Orbit if useful. Review a claim's source before sharing. This site does not post on your behalf or aggregate community votes. The research proof is a small optional anti-spam receipt; links can be copied and are not identity-bound. It carries no monetary value and is not a crypto-asset.
Method reviewed 20 September 2026 by the AGI Scorecard team. Citation kit · AI / MCP methods
Calculate a payment or gas-cost impact
Put the quote in the context of your task
Enter a dated source price or your own settlement rate. We do not silently replace your inputs with a moving quote.
21,000 is a simple native-transfer example, not a token-transfer or contract-call estimate. Obtain the actual gas estimate from your transaction. This calculator runs in your tab.
What needs a decision?
Did the invoice actually settle?
A token price near $1 does not match a payment to an invoice. Check the chain, asset, amount and confirmation record.
Check a stablecoin paymentWill fees erase the saving?
Separate gas units, effective gas price and ETH/USD. Include proving and verification before comparing deployment choices.
Build a gas and proof budgetDoes a protocol change affect my agent?
Pin the implemented version, inspect the official change and rerun your evidence review.
Review an integration changeUpdated