GameFi Weekly Data Report — August 31–September 6, 2026
Across the nine chains I track, 1,033,900 addresses played on-chain games each day in the week of August 31 – September 6, 2026, down −3.4%, paying a weighted $0.1001 per game transaction. Seven chains shrank; Solana and Base grew. This report answers four questions: what happened, what grew fastest, what fell, and whether anything breached anomaly thresholds against the past four weeks.
1. The week in one table
| Headline metric | This week | Last week | WoW | 4-week mean | vs mean |
|---|---|---|---|---|---|
| GameFi daily active addresses | 1,033,900 | 1,070,300 | −3.4% | 1,056,100 | −2.1% |
| GameFi daily transactions | 3,756,700 | 3,938,100 | −4.6% | 3,860,650 | −2.7% |
| Transactions per player per day | 3.63 | 3.68 | −1.2% | 3.66 | −0.6% |
| Chains growing (of 9) | 2 | 2 | — | 2.5 | — |
Two things stand out before any per-chain detail. First, activity and transactions fell by nearly the same proportion, which means the players who left were average-intensity players — the grinders did not leave and the tourists did not dominate. Second, this was not a fee-shock week in the aggregate; the per-chain split, not the network average, carries the signal.
2. Chain-by-chain scorecard
This is the core table of the report — nine chains, every number I publish, in one place. What those prices mean for actual monthly bills, including TRON's three payment routes, is priced out in this week's cost update.
| Chain | GameFi DAU | Share | Daily tx | Avg gas/tx | 50 tx/day, 30d | DAU WoW | tx WoW |
|---|---|---|---|---|---|---|---|
| Ronin | 312,600 | 30.2% | 1,284,000 | $0.0000 gas-free | $0.00 | −5.6% | −9.1% |
| Solana | 186,400 | 18.0% | 894,300 | $0.00021 | $0.32 | +4.2% | +3.8% |
| Immutable | 148,200 | 14.3% | 512,700 | $0.0000 gas-free | $0.00 | −3.0% | −4.7% |
| TRON | 106,400 | 10.3% | 206,400 | $0.7388 | $1,108.14 | −6.8% | −8.4% |
| BNB Chain | 96,300 | 9.3% | 318,400 | $0.0940 | $141.00 | −5.0% | −5.5% |
| Polygon | 74,800 | 7.2% | 246,900 | $0.0210 | $31.50 | −6.0% | −5.5% |
| Arbitrum | 42,100 | 4.1% | 128,600 | $0.0780 | $117.00 | −5.0% | −6.1% |
| Base | 38,700 | 3.7% | 104,200 | $0.0410 | $61.50 | +6.9% | +7.0% |
| Ethereum | 28,400 | 2.7% | 61,200 | $2.8400 | $4,260.00 | −10.4% | −11.2% |
| Total | 1,033,900 | 100% | 3,756,700 | $0.1001 weighted | — | −3.4% | −4.6% |
FormulaWeighted average cost = Σ(cost per tx × chain daily tx) ÷ Σ(chain daily tx) = $0.1001 across the nine chains, skewed almost entirely by the two expensive tails.FormulaFree-chain player share = (312,600 + 148,200) ÷ 1,033,900 = 460,800 ÷ 1,033,900 = 44.6%.
3. Cost ranking: what each chain charges
Ranked from cheapest to most expensive per game transaction, measured in the window:
| Rank | Chain | Cost per tx | Note | 50 tx/day, 30d |
|---|---|---|---|---|
| 1= | Ronin | $0.0000 | Gas-free by chain design | $0.00 |
| 1= | Immutable | $0.0000 | Gas-free by chain design | $0.00 |
| 3 | Solana | $0.00021 | Near-free | $0.32 |
| 4 | Polygon | $0.0210 | Low | $31.50 |
| 5 | Base | $0.0410 | Low | $61.50 |
| 6 | Arbitrum | $0.0780 | Low | $117.00 |
| 7 | BNB Chain | $0.0940 | Low | $141.00 |
| 8 | TRON | $0.7388 | High — median energy rental | $1,108.14 |
| 9 | Ethereum | $2.8400 | Very high | $4,260.00 |
Three structural facts in this ranking. The four-order-of-magnitude spread between rank 1 and rank 9 means chain choice is not a preference, it is the single biggest cost decision in on-chain gaming. TRON sits second-dearest not because its base fee is high but because game contracts burn energy at the burn fallback unless the player actively rents — $0.7388 is already the managed price, and the unmanaged burn fallback is $2.17 per transfer (100 sun per energy under proposal 104). Ethereum's $2.84 is a raw L1 price with no rental market to soften it.
FormulaTRON premium over the network average = $0.7388 ÷ $0.1001 = 7.38×.
4. Cost versus growth: this week's correlation read
The question I get asked most is whether fees actually push players between chains. This week gives a cleaner answer than most.
| Chain | Cost rank (1 = cheapest) | DAU WoW | Observation |
|---|---|---|---|
| Base | 5 | +6.9% | Grew fastest; sub-$0.05 band |
| Solana | 3 | +4.2% | Second gainer; near-free |
| Immutable | 1= | −3.0% | Free, but still shrank |
| Ronin | 1= | −5.6% | Free, but fell with cadence |
| BNB Chain | 7 | −5.0% | Mid-pack decline |
| Arbitrum | 6 | −5.0% | Mid-pack decline |
| Polygon | 4 | −6.0% | Cheap, still fell |
| TRON | 8 | −6.8% | Second-dearest, second-worst |
| Ethereum | 9 | −10.4% | Dearest, worst |
My honest reading: the correlation is real at the extremes and weak in the middle. The only two chains that grew are paid chains in the sub-$0.05 band — that is not a coincidence. The two most expensive chains posted the two worst declines — also not a coincidence. But the free chains shrank anyway, which proves cost is not destiny: a zero fee cannot make players stay when session cadence drops. Cost is a filter at the margins, not an engine in the middle of the distribution. The same week read through five standing questions — players, games, economy, cost, outlook — is in the weekly intelligence briefing.
5. Four-week trend series
| Week ending | Network DAU | DAU WoW | Network daily tx | TRON DAU | TRON WoW |
|---|---|---|---|---|---|
| Aug 16, 2026 | 1,068,400 | +0.9% | 3,905,200 | 112,600 | +1.2% |
| Aug 23, 2026 | 1,051,800 | −1.6% | 3,842,600 | 109,300 | −2.9% |
| Aug 30, 2026 | 1,070,300 | +1.8% | 3,938,100 | 114,200 | +4.5% |
| Sep 6, 2026 | 1,033,900 | −3.4% | 3,756,700 | 106,400 | −6.8% |
Against the four-week mean, this week's network DAU sits 2.1% low and network transactions 2.7% low. The more interesting number is the shape: the three previous weeks oscillated inside a ±1.8% band, and this week broke below it. TRON's path is the same story with the amplitude turned up — it bounced hardest into the Aug 30 week and fell hardest out of it.
6. What grew fastest
Only two chains grew, and they are worth reading precisely because there are only two:
- Base, +6.9% DAU and +7.0% transactions. The growth is balanced — activity and usage expanded at nearly the same rate, which suggests whole players arriving, not bots farming a snapshot. At $0.0410 per transaction, Base's fee is real but small — under a nickel an action, and not the reason the chain grew.
- Solana, +4.2% DAU and +3.8% transactions. At $0.00021, Solana is functionally free for most session shapes. Its 4.80 transactions per player per day — the highest intensity in our set — means its growth leans on the heavy end of the distribution.
FormulaSolana player intensity = 894,300 ÷ 186,400 = 4.80 transactions per player per day, versus a network average of 3,756,700 ÷ 1,033,900 = 3.63.
7. What fell
Seven chains contracted. The three worth singling out:
- Ethereum, −10.4% DAU and −11.2% transactions. The double-digit slide is the worst in the set, and transactions fell faster than players — meaning its remaining users also cut intensity. At $2.84 per transaction, a 50-transaction day costs $142, which prices out every session shape except whale activity.
- TRON, −6.8% DAU and −8.4% transactions. Its players already ration usage — 1.94 transactions per player per day, half the network average — and the decline still outpaced the network. This is the cost drag showing up as behaviour, not commentary.
- Ronin, −5.6% DAU but −9.1% transactions. The gap matters: activity fell much faster than the player base, so session intensity dropped hardest here. A free chain losing intensity is a content-cadence signal, not a cost one.
8. Anomalies and attribution
My anomaly triggers are a single-day move beyond 30% on any tracked series, or a weekly print beyond ±2σ of the trailing four-week distribution. Both checks ran this week.
| Check | Threshold | This week | Verdict |
|---|---|---|---|
| Max single-day DAU move, any chain | >30% | Largest observed ≈ 6% | No breach |
| Network DAU WoW | ±2σ of trailing 4 weeks (σ = 2.37pp) | −3.4% | ≈ −1.2σ — inside band |
| Network tx WoW | ±2σ of trailing 4 weeks | −4.6% | ≈ −1.3σ — inside band |
| TRON DAU WoW | ±2σ of trailing 4 weeks (σ = 4.91pp) | −6.8% | ≈ −1.2σ — inside band |
| Ethereum DAU WoW | Largest single-chain move in window | −10.4% | Watch item, below trigger |
FormulaNetwork z-score = (−3.4% − (−0.575%)) ÷ 2.37pp = −1.19, inside the ±2σ anomaly band.
Attribution, without claiming causation. The decline concentrated exactly where the previous week's bounce had concentrated — the chains that rose into the Aug 30 week fell hardest out of it, which is the shape of mean reversion, not of a shock. No exploit, outage, parameter change or listing event entered my tracked set during the window that would explain a discontinuity, and no discontinuity appears in the data — the full item-by-item event sweep is in the weekly news review. The one watch item: Ethereum's −10.4% is its steepest print in the four-week window; if it repeats next week it stops looking like noise, and I will flag it as a trend rather than an anomaly.
9. Player cost by persona
The full monthly-cost matrix: four session shapes, nine chains, one month (30 days) of identical play.
FormulaPersona monthly cost = cost per tx × transactions per day × 30. Example: TRON Grinder = $0.7388 × 50 × 30 = $1,108.14.
| Chain | Casual (3 tx/day) | Regular (15 tx/day) | Grinder (50 tx/day) | Studio (500 tx/day) |
|---|---|---|---|---|
| Ronin | $0.00 | $0.00 | $0.00 | $0.00 |
| Immutable | $0.00 | $0.00 | $0.00 | $0.00 |
| Solana | $0.02 | $0.09 | $0.32 | $3.15 |
| Polygon | $1.89 | $9.45 | $31.50 | $315.00 |
| Base | $3.69 | $18.45 | $61.50 | $615.00 |
| Arbitrum | $7.02 | $35.10 | $117.00 | $1,170.00 |
| BNB Chain | $8.46 | $42.30 | $141.00 | $1,410.00 |
| TRON | $66.49 | $332.44 | $1,108.14 | $11,081.45 |
| Ethereum | $255.60 | $1,278.00 | $4,260.00 | $42,600.00 |
On TRON specifically, the payment route is a second decision layered on top of chain choice. Renting energy at the median rate against paying the burn fallback:
| Persona (TRON) | Renting | Burn fallback | Saving |
|---|---|---|---|
| Casual (3 tx/day) | $66.49 | $195.55 | 66.0% |
| Regular (15 tx/day) | $332.44 | $977.77 | 66.0% |
| Grinder (50 tx/day) | $1,108.14 | $3,259.25 | 66.0% |
| Studio (500 tx/day) | $11,081.45 | $32,592.50 | 66.0% |
FormulaStudio burn cost = 64,285 energy × 0.0001 TRX × 500 tx × 30 days × $0.338 = $32,592.50; rental saves 1 − ($0.7388 ÷ $2.1728) = 66.0%.
Read the matrix by column, not by row. A casual player can play anywhere — even Ethereum costs $255.60 a month, a luxury purchase rather than an impossible one. The grinder is where chains separate: $0.32 on Solana against $1,108.14 on TRON is a 3,500× spread for the identical habit. At the studio tier, only a handful of rows are even discussable, because $11,081.45 on TRON and $42,600.00 on Ethereum are costs that consume the margin a studio exists to capture. That is why 44.6% of players sit on the two free chains. For the one expensive chain most earners cannot simply leave, the payment route behind TRON's $0.7388 — and the corrected 100-sun burn arithmetic that halved the penalty for burning — is unpacked in this week's deep dive.
10. What this means
10.1 For players
Your chain choice is worth more than any in-game optimisation. Moving a 50-transaction-a-day habit from TRON to Polygon saves $1,076.64 a month — more than any reward programme pays. If you must stay on an expensive chain, the payment route is the lever: on TRON, renting instead of burning is worth 66.0% regardless of persona, and services like Tronsell automate exactly that one decision.
10.2 For guilds and studios
The persona matrix is a siting calculation. A 500-transaction operation has nine rows, two of them free and two of them uneconomic — and the market is already voting the same way, with this week's growth concentrated in the sub-$0.05 band.
10.3 For publishers
Chain selection is now a player-acquisition cost. This week's data shows the two cheapest paid chains growing while the two most expensive shrank hardest, with free chains holding 44.6% of players. A game that ships on an expensive chain needs a fee-subsidy design on day one, because its players pay a 7.4× premium over the network average before the game itself earns anything.
11. Method and confidence
| Series | Method | Confidence |
|---|---|---|
| Per-chain gas/tx | Explorer fee data and gas trackers, sampled through the window | High — slow-moving, reproducible |
| Network and chain DAU/tx | Labelled contract set across nine chains | Medium — order of magnitude, published as levels |
| WoW and 4-week comparisons | Same method applied to both windows | Medium — consistent, so deltas are meaningful |
| Persona costs | Gas price × fixed session shape × 30 days | High — arithmetic, not sampling |
| Anomaly checks | 30% daily trigger; ±2σ on trailing 4 weeks | Medium — thresholds are conventions, stated openly |
The main caveat repeats every week: the modelled activity figures are built to be comparable across chains, and comparability costs precision. Treat 1,033,900 as "a bit over a million", and treat the WoW deltas — which share the same method on both sides — as more trustworthy than the levels.
12. Frequently Asked Questions
What happened in on-chain gaming this week?
Across the nine tracked chains, daily players fell 3.4% to 1,033,900 and daily game transactions fell 4.6% to 3,756,700 in the week of August 31 – September 6, 2026. Two chains grew — Solana and Base — and seven contracted, with the declines concentrated on the most expensive chains.
Which GameFi chain grew fastest this week?
Base, up 6.9% in daily players and 7.0% in transactions, followed by Solana at +4.2% and +3.8%. Both charge under $0.05 per transaction — $0.0410 and $0.00021 respectively — making them the only gainers in a week where the two most expensive chains fell hardest.
Which chain fell the most?
Ethereum, down 10.4% in players and 11.2% in transactions at $2.84 per transaction — the highest cost and the steepest decline in the set. TRON was second-worst at −6.8%, followed by Polygon at −6.0%.
Were there any anomalies this week?
No. No chain printed a single-day move beyond the 30% trigger, and the network's −3.4% weekly move sits about 1.2σ below the trailing four-week mean — inside the ±2σ band. Against the four-week average, this week's activity is 2.1% low. The one watch item is Ethereum's steepest-in-window decline, which becomes a trend flag if it repeats.
How much does a month of heavy on-chain gaming cost?
For a 50-transaction-a-day player over 30 days: $0.00 on Ronin and Immutable, $0.32 on Solana, $31.50 on Polygon, $61.50 on Base, $117.00 on Arbitrum, $141.00 on BNB Chain, $1,108.14 on TRON and $4,260.00 on Ethereum. At the studio tier of 500 transactions a day, the spread runs from $0 to $42,600.00.
How are these figures calculated?
Fees are measured from explorer and gas-tracker data; activity is modelled from our labelled game-contract set across the nine chains, because no public dashboard offers a comparable cross-chain GameFi segment. Weekly comparisons apply the same method to both windows, so the deltas are more reliable than the levels. Definitions live on our sources and methodology page.
13. Sources
- Nine-chain activity baseline and week-over-week comparisons — labelled contract set cross-checked against DappRadar dashboards, compiled Monday, September 7, 2026.
- Per-chain fee data — Etherscan gas tracker (Ethereum), BscScan (BNB Chain), Polygonscan (Polygon), Arbiscan (Arbitrum), Basescan (Base), TRONSCAN (TRON), sampled through the window.
- Solana and Ronin activity context — Solana FM and Ronin explorer, week of August 31 – September 6, 2026.
- Immutable fee model and activity context — Immutable explorer, week of August 31 – September 6, 2026.
- TRON parameters — 64,285 energy per standard transfer, 100 sun (0.0001 TRX) per energy under committee proposal 104 (effective Aug 29, 2025), TRX at $0.338 (official quote $0.3376, Sep 15, 2026) — our read of on-chain receipt data, Tuesday, September 1, 2026.
- TRON energy rental market — cheapest listed 22 SUN, listed range 22–38 SUN, indicative 26–74 SUN, mid-band $0.7388 — our own sample of provider listings during the window.
- Aug 24–30 comparison baseline — same method, prior window, retrieved Monday, September 7, 2026.
- Four-week series (weeks ending Aug 16, Aug 23, Aug 30) — internal archive, same methodology.
Disclaimer
Not financial advice. Activity figures are modelled from our labelled contract set unless marked measured, and will differ from any single dashboard's cut of the market. Fee figures are samples that vary by block, route and timing. Energy prices, gas prices and network parameters change daily. Nothing here recommends buying, selling, staking or renting anything — read our full disclaimer first.