FET Whale Report: 49 Wallets, $23.7M Volume — Smallest Pool, Highest Flow-to-Cap Ratio [Live 2026]

Fetch.ai is an AI and machine-learning infrastructure platform originally built around autonomous economic agents — software entities that negotiate, transact and optimise on behalf of users without requiring continuous human intervention. In 2024 the project merged with Ocean Protocol and SingularityNET to form the Artificial Superintelligence Alliance (ASI), and FET became the unified token of the combined ecosystem. On 20 September 2026 it traded near $0.176 with a market capitalisation of about $407 million, up 2.6% over seven days and 10.1% over thirty, and roughly 94.9% below its all-time high of $3.45 reached on 28 March 2024.

This is the sixth issue of our token whale report series, and it introduces something the prior five did not have: a structural event that changed the token’s identity during the measurement window. Every previous token — MORPHO, PROM, AAVE, ENA, DEXE — was the same instrument at the start and end of its window. FET was not. Its ATH was set under a different supply structure, its wallet base now includes holders from three formerly separate communities, and the dollar volumes partly reflect migration-related activity rather than purely market-driven positioning.

What the ASI Alliance means for reading this data

The ASI Alliance merged three projects — Fetch.ai, Ocean Protocol (OCEAN) and SingularityNET (AGIX) — under a single token. OCEAN and AGIX holders were offered conversion to FET at fixed ratios. The result was a change in circulating supply, holder base and market identity that a standard “token whale report” framework cannot capture from flow data alone.

Two specific implications for the figures below:

The ATH comparison is weaker than usual. FET’s $3.45 high was reached on 28 March 2024, before the merger completed. The token that hit $3.45 had different supply dynamics and a different holder composition from the token trading at $0.176 today. A 94.9% drawdown from that figure is arithmetically correct but not a comparison of the same instrument across two dates in the way that, say, AAVE’s ATH comparison is.

Some exchange flow may reflect conversion mechanics. Token conversions between OCEAN/AGIX and FET, handled through exchanges, would appear in the data as ordinary withdrawals and deposits. This does not invalidate the flow figures — the transfers happened and Deep Blue Alpha recorded them — but it means the accumulation share may partly describe migration logistics rather than market conviction. We cannot separate the two from transfer data alone and do not attempt to.

The numbers at a glance

At $23.7 million of tracked volume, FET is the smallest token in this series by raw whale throughput. DEXE carried $85.51 million; PROM carried $40.48 million; ENA carried $404.34 million; AAVE carried $555.27 million. But raw dollars miss the point when the market caps are this different.

Why the flow-to-mcap ratio matters more than the dollar figure

FET’s $23.7 million of whale volume against a $407 million market cap gives a flow-to-mcap ratio of about 5.8%. Its +$5.9 million net represents roughly 1.45% of market cap.

Read the rightmost column. FET’s $5.9 million net sounds modest against PEPE’s $28.1 million, but it represents more than double the share of market cap. DEXE sits highest at 5.01% — a reminder that raw dollar flow is a function of token size as much as whale interest.

This ratio is useful because it makes cross-token comparison possible. A $5 million net inflow into a $400 million token is a meaningfully different event from a $5 million net inflow into a $4 billion token, even though the dollar figure is identical. The ratio captures that difference; the dollar figure conceals it.

49 wallets: what a narrow base means

FET’s 49 tracked wallets is the smallest base in this series by a wide margin. The next smallest is PROM at 314. At the other end, AAVE has 1,615.

The moves-per-wallet column is notable. FET’s 7.6 is comparable to AAVE’s 7.1 — these wallets are not hyperactive traders like DEXE’s (14.6 moves per wallet). They move at a moderate pace, but there are far fewer of them.

A narrow wallet base has three practical consequences for reading the data:

Individual wallets carry more weight. If AAVE’s top wallet shifted direction, it would be one voice among 1,615. If FET’s top wallet shifted direction, it would be one voice among 49 — roughly thirty times more influence on the aggregate ratio.

The accumulation share is less stable. A 62.4% accumulation share from 49 wallets could shift to 50% or 75% with two or three wallets changing behaviour. The same ratio from 1,615 wallets is harder to move.

Absence of wallets is itself a data point. Deep Blue Alpha tracks 20,000+ Ethereum whale wallets. Only 49 of them had FET activity in this window. That is not a coverage gap — it is a measurement. The market’s tracked whale population, by and large, was not moving FET.

The $64,000 average move

FET’s 370 moves across $23.7 million gives an average move size of roughly $64,000. That is moderate — larger than DEXE’s $7,742 average but smaller than AAVE’s $48,318 or MORPHO’s six-figure block transfers.

A $64,000 average move against a $407 million market cap is unremarkable at the per-transaction level. The significance comes from the aggregate: 370 of them adding up to a net of +$5.9 million, which is 1.45% of market cap. No individual transfer is large enough to move the market; the cumulative lean is what the data shows.

The drawdown context

At 94.9% below its all-time high of $3.45, FET has the deepest drawdown of any token in this series. For comparison, ONDO sat roughly 58% off ATH, AAVE roughly 62%, and even DEXE at 95.3% was comparable. But FET’s drawdown carries an asterisk that DEXE’s does not.

DEXE’s ATH of $48.91 was reached in February 2021 and describes the same token across the whole period. FET’s $3.45 was reached in March 2024, months before the ASI Alliance merger altered its supply structure. The token that hit $3.45 had a different circulating supply and a different holder composition. The arithmetic says 94.9% down; the structural context says the comparison is weaker than the number implies.

What is not weakened is the observation that FET trades at a fraction of its historical peak while whale flow ran net positive. That combination — deep drawdown plus accumulation-side lean — is the same pattern that appeared in DEXE, and it does not resolve into a prediction either way. Both facts can be true simultaneously without one explaining the other.

62.4% accumulation share: comparable to ONDO, different base

FET’s 62.4% accumulation share is the proportion of its tracked whale volume that consisted of exchange withdrawals rather than deposits. It sits in the same range as ONDO’s 64.9%.

The number is nearly identical to ONDO’s, but the wallet counts — 49 against 493 — make it a fundamentally different observation. ONDO’s 64.9% emerged from nearly 500 wallets; reversing it would require a broad shift in behaviour. FET’s 62.4% emerged from 49 wallets; it could reverse with a handful of wallets changing course.

Neither reading is “better” or “worse.” But they are not the same observation even though the percentage is comparable. The wallet count is the denominator that gives the percentage its meaning, and omitting it would make FET look like ONDO when the two situations are structurally different.

The AI-token question

FET is routinely grouped with AI tokens in market commentary — alongside RENDER, TAO, NEAR (for its AI efforts), and the other ASI Alliance components. The grouping is not wrong; Fetch.ai’s core technology is AI agents and machine-learning infrastructure. But in the context of whale flow data, the label does not add information.

Nothing in exchange-transfer data distinguishes an AI token from a DeFi token or a meme token. The transfers are the same: withdrawals from exchanges, deposits to exchanges, denominated in dollars, measured across wallets. A withdrawal of $64,000 worth of FET looks identical in the data to a withdrawal of $64,000 worth of AAVE. The sector classification is metadata, not a measurement.

What AI-sector membership does provide is a narrative context in which exchange-flow data is interpreted by other market participants. When “AI tokens” run as a group, individual tokens within the group often move together regardless of their individual flow dynamics. FET’s 10.1% thirty-day gain may partly reflect sector rotation rather than token-specific demand, and the whale flow data cannot separate these.

This is worth saying explicitly because the temptation to read AI-sector significance into FET’s accumulation share is strong, and the data does not support it. The 62.4% figure describes what 49 tracked wallets did. It does not describe why they did it, and attaching a sector narrative to the “why” from the outside is exactly the kind of interpolation this series avoids.

Supply: 85.3% circulating

FET has a circulating supply of approximately 2.31 billion tokens against a maximum supply of 2.71 billion — 85.3% in circulation. This is a relatively high ratio, meaning dilution from future token releases is limited compared to tokens with 30-50% circulating.

For whale flow analysis, the supply ratio matters because it sets the denominator for concentration calculations. A whale accumulating $5.9 million worth of a token where 85% is already circulating is making a different bet than one accumulating the same dollar amount of a token where 30% is circulating and large unlocks are scheduled. In the latter case, upcoming supply expansion could dilute the position regardless of demand.

FET’s high circulating ratio means the supply overhang is relatively small. That does not make accumulation more or less significant in any directional sense — it simply means the token’s available float is close to its terminal supply, which is relevant context for anyone comparing FET’s flow dynamics against tokens with upcoming unlock schedules.

Exchange transfers, not trades

These are exchange transfers: withdrawals from exchanges labelled as accumulation-side, deposits to exchanges labelled as distribution-side. A withdrawal removes tokens from immediately sellable exchange inventory but is not itself a purchase. We label transfers and trades separately rather than merging them.

This composition is not specific to FET. Exchange flow dominates tracked activity across effectively all monitored tokens. When we report a 62.4% accumulation share, we mean 62.4% of tracked volume consisted of tokens leaving exchanges. The heuristic — coins leaving exchanges lean toward accumulation, coins arriving lean toward distribution — is stated rather than assumed, and the underlying data is exchange-transfer data throughout.

A method for small-cap, narrow-base tokens

1. Check the wallet count before the flow. 49 wallets and 1,615 wallets produce numbers with different statistical weight. Note the base before drawing conclusions.

2. Use flow-to-mcap ratio, not raw dollars. FET’s +$5.9 million is modest in isolation and meaningful at 1.45% of market cap. Raw dollars are a function of token size; the ratio normalises for it.

3. Calculate the average move size. $64,000 per move at FET’s price and market cap is moderate. Compare it against the token’s daily volume ($121 million) to gauge how much of the market’s throughput the tracked wallets represent.

4. Discount the ATH comparison after structural events. A 94.9% drawdown from a pre-merger ATH is not directly comparable to a drawdown from a same-instrument peak. Note it, but weight it less.

5. Ask what a comparable accumulation share means from a narrower base. 62.4% from 49 wallets is not the same observation as 64.9% from 493 wallets. The percentage is similar; the stability of the reading is not.

6. Resist attaching sector narratives to flow data. FET is an AI token. The flow data does not know this and cannot confirm or deny whether AI-sector momentum explains the accumulation lean.

What this data cannot tell you

Whether the wallets are merger-related. Some of the 49 wallets may hold FET as a result of converting from OCEAN or AGIX rather than from a new market decision. Exchange-transfer data does not distinguish conversion-driven activity from market-driven activity.

Why the wallet count is so low. 49 out of 20,000+ tracked wallets is conspicuously small. It could reflect genuine disinterest from the whale population, it could reflect the token’s relatively small market cap making it below the threshold of interest for most large wallets, or it could be an artefact of how wallets are discovered. The data records who moved; it does not explain who did not.

Whether AI-sector rotation drove the flow. FET’s 10.1% thirty-day gain occurred alongside broader AI-token momentum. The accumulation lean may reflect sector-level positioning rather than FET-specific conviction, and the transfer data cannot separate these.

Anything about what follows. Every figure here describes the past. A 62.4% accumulation share and a 94.9% drawdown from ATH are both retrospective measurements, and neither is a prediction that price will converge toward the flow direction or away from it.

Why this series now covers AI infrastructure

Six issues in, the series has covered DeFi governance (AAVE, MORPHO), layer-one infrastructure (PROM), social trading (DEXE), stablecoins/yield (ENA), and now AI infrastructure (FET). Each added something the prior issues did not measure.

FET’s contribution is the narrowest wallet base the series has seen, combined with a structural merger that complicates every historical comparison. These are not exotic conditions — many tokens outside the top 100 have tracked wallet counts in the low dozens, and token mergers, migrations and rebrandings happen routinely across crypto. The method needs to work here, not just on AAVE’s 1,615-wallet dataset.

The adjustments are the ones outlined in the method section: normalise by market cap instead of reading dollars alone, note the wallet count before interpreting the ratio, discount ATH comparisons after identity-changing events, and resist attaching narratives the data cannot confirm. None of these are FET-specific. They apply to any small-cap token with a narrow tracked base and a complicated history, and that describes a large fraction of the market outside the top twenty.

The bottom line

FET whale wallets tracked by Deep Blue Alpha moved $23.7 million across 370 whale moves, ending +$5.9 million net at a 62.4% accumulation share from 49 tracked wallets — the fewest in this series.

The number worth carrying is the flow-to-mcap ratio of 1.45%. FET’s raw dollar net is the smallest in this series, but its share of market cap exceeds PEPE’s. That is a direct consequence of FET’s smaller size, and it is the reason normalisation matters: a $5.9 million net into a $407 million token is a proportionally larger event than a $28 million net into a $4.5 billion token, even though the second looks more impressive in a headline.

The second number worth carrying is 49 wallets. It determines how much weight to give every other figure in this report. A 62.4% accumulation share is an observation, not a consensus, when it comes from fewer wallets than many tokens have in their top-ten holder list. The narrow base does not invalidate the data — it sets the confidence interval around it.

Methodology & limitations. Figures derive from Deep Blue Alpha’s tracked-wallet database covering 20,000+ Ethereum whale wallets, queried 20 September 2026. Flow is exchange withdrawals and deposits, labelled separately from decentralised exchange trades. Market data via CoinGecko, 20 September 2026. Comparison figures for MORPHO, PROM, AAVE, ENA, DEXE, ONDO and PEPE use comparable windows. Wallet identifiers are withheld in this public edition. This covers wallets Deep Blue Alpha tracks, a subset of all FET holders — not a complete census of the token’s ownership. The ASI Alliance merger means some exchange transfers may reflect token-conversion activity rather than market positioning; the data does not distinguish these.

Related reading

MORPHO Whale Report

Issue one: the low-frequency extreme — six wallets, eight transfers.

PROM Whale Report

Issue two: same concentration, opposite transaction behaviour.

AAVE Whale Report

Issue three: one wallet exceeded the market's net result by 3.2 times.

ENA Whale Report

Issue four: the largest inflow measured, with a minority of wallets buying.

DEXE Whale Report

Issue five: the highest transfer intensity in the series — 203 transfers per large mover.

DeFi Yield Whale Activity

Where large wallets positioned across lending and yield protocols.

Live FET whale flow → All tracked tokens → Whale wallet leaderboard → Daily whale reports → More token deep dives →