Web3 Gaming's Forecasts Say $279 Billion. Its Wallet Count Says 4.66 Million.
Two research firms priced blockchain gaming in 2025 at $24 billion and $229 billion. Same market, same year, a tenfold gap — and neither figure counts players. The one number in this category that gets measured rather than modelled, daily active wallets, fell to 4.66 million by the third quarter of 2025, down from a peak of 7.3 million that January. If you buy game catalogues for a living, the interesting part is not what happened to web3. It is that the same measurement failure turns up in decks that have nothing to do with a blockchain.
I am not writing this to dunk on a category. Forestry Games carries web3 titles alongside everything else, and a game is a game. I am writing it because blockchain gaming is the cleanest worked example available of a market number detaching from the thing it claims to describe, and because the people who got hurt by it were mostly not crypto people. They were operators, publishers and portal owners who read a forecast, believed the unit on the x-axis, and bought accordingly.
📊 Same Category, Same Year, a Tenfold Gap
Start with the two numbers, because the rest follows from them. IMARC Group puts the global blockchain gaming market at $24.0 billion in 2025, reaching $1,600.9 billion by 2034 on a 59.46% CAGR. Fortune Business Insights, covering a category with the same name, puts 2025 at $229.15 billion, 2026 at $279.10 billion, and 2034 at $1,351.87 billion on a 21.80% CAGR.
Look at the shape of that. The two models start a factor of ten apart and finish within about 18% of each other. That is not two attempts to measure one thing that landed in different places. It is two different definitions of what "blockchain gaming" contains, wearing the same three words and pointing at roughly the same trillion-dollar horizon.
Neither firm is doing anything improper. Both sell reports, and inside those reports is a scope note explaining what is in the total. The problem is downstream: the scope note stays in the PDF and the number goes travelling. By the time a figure reaches a board deck, a term sheet or a LinkedIn post, it has been stripped of the only sentence that made it interpretable.
🌍 One of Those Numbers Is Bigger Than the Entire Games Industry
Here is the test that takes ten seconds and settles it. Newzoo estimated the whole global games market — every console, PC and mobile game sold on earth — at $201.6 billion in 2025, up 9.1% year on year.
Fortune Business Insights' figure for blockchain gaming alone in 2025 is $229.15 billion. The segment is larger than the industry that contains it.
A sub-category cannot exceed its parent unless it is counting something the parent does not count. What that something is, I cannot tell you from outside the paywall, and pretending otherwise would be the same sin I am describing. The plausible candidates are token and NFT transaction volume, on-chain infrastructure spend, and chain fees — turnover, in other words, not revenue in the sense a publisher uses the word. A game that sells $10 of currency which then changes hands nine times has produced $10 of revenue and $100 of volume. Which one is in the total changes the answer by an order of magnitude, and that is exactly the size of the gap we are looking at.
For contrast, look at what disciplined revision looks like. Newzoo's own in-year forecast for 2025 was reported at $188.8 billion before its full-year estimate landed at $201.6 billion — a move of roughly 7%. That is a model being corrected by someone who has to defend it to people who actually sell games.
👛 The Metric Underneath Was Never a Person
Now to the number that does get measured. DappRadar's unique active wallets (UAW) metric counts unique wallet addresses interacting with a dapp's smart contracts. It is an honest measure of exactly that, and DappRadar says so plainly in its own explainer: "Wallets and real people only sometimes match one-to-one. Therefore, we must consider the case that one person holds multiple wallets."
That caveat is in the source. It almost never survives the trip into a pitch. And it matters more than it sounds, because of the direction the error runs. People hold extra wallets for security, but in this category they mostly held extra wallets to qualify for more airdrops, more mints, more allocation. The incentive to split one person across many addresses is strongest precisely when rewards are richest — which is to say, during the steep part of the growth chart that got put in front of investors and licensing buyers.
The measured trend, from DappRadar's Q3 2025 report and the quarters around it: blockchain gaming averaged 4.66 million daily unique active wallets in Q3 2025, down 4.4% quarter on quarter. Q2 2025 was 4.8 million, down 17% on the quarter and the lowest since early 2023. January 2025 was the peak at 7.3 million.
Two other things from that report are true at the same time and worth holding together. Gaming grew from 20.1% to 25% of all dapp activity, and gaming activity fell. It became the largest category in a contracting industry. Any metric that lets both sentences be true is a share metric, not a demand metric, and share metrics are the easiest thing in the world to quote out of context.
💀 The Shakeout, in Figures That Also Disagree
The research firm Caladan analysed more than 3,200 web3 gaming projects launched since 2020 and found roughly 93% inactive or dead, as reported by DailyCoin in April 2026. The same report puts total deployed capital across the sector at $12 billion, venture funding down from about $4 billion in 2022 to $360 million in 2025, token valuations down roughly 95% from 2022 highs, and GameFi's share of crypto venture funding down from around 60% to single digits.
CoinDesk's April 2026 write-up of the same research framed it as a $15 billion boom. So call it $12–15 billion. Even the amount of money that went in is a range, in a report about where the money went.
DappRadar's Q2 2025 quarter gives the attrition a texture: more than 300 gaming dapps went inactive in a single quarter, about 8% of all gaming dapps it lists, and web3 gaming funding fell 93% year on year to $73 million. Per the Caladan coverage, capital did not evaporate so much as relocate — roughly $2.6 billion into layer-2 infrastructure, $2 billion into real-world asset tokenisation and $1.8 billion into AI. Animoca Brands, one of the category's largest backers, cut web3 gaming from 62% to 25% of its portfolio.
If you want one example that makes the metric problem physical, use Hamster Kombat. The Telegram game reported figures around 300 million users in August 2024. By November 2024 the reported figure was 41 million. A number that falls by more than 250 million in three months was never an audience. It was a queue for a token, and it dispersed the moment the token priced.
🚪 The Distribution Rails Priced This In Years Ago
While forecasts were compounding at 59% a year, the platforms that actually control distribution were doing the opposite, and their positions have barely moved.
Steam banned applications built on blockchain technology that issue or allow the exchange of cryptocurrencies or NFTs in October 2021, and the rule still stands as of August 2026. Gabe Newell's stated reason, in a February 2022 interview, was that when Steam accepted cryptocurrency roughly 50% of those transactions were fraudulent.
Google Play went the other way in July 2023, permitting tokenised digital assets in apps and games — with conditions that matter more than the permission. Developers must declare blockchain elements in Play Console, may not promote or glamorise any potential earnings from playing or trading, and cannot sell a loot-box-style item for a chance at an NFT without meeting gambling eligibility requirements.
Apple permits NFTs but routes purchases and unlocks through in-app purchase, and prohibits on-device mining and distributing token rewards for completing tasks. Following the 2025 US court ruling in the Epic case, Apple's US guidelines allow links and buttons out to secondary marketplaces and third-party payment systems; the global guidelines remain tighter.
Read those three together and the operator translation is straightforward: the compliance surface is the earning language, not the chain. A game can hold assets on-chain and ship fine. A store listing that promises returns is what gets pulled. That is the same rule that governs any Android title you licence and publish — the copy, the rating and the disclosures are where policy bites, not the tech stack.
🧩 What Any of This Changes If You Licence Games for a Living
Strip the token layer off and a web3 game is a game. That is the practical position, and it survives whichever way the category goes:
- The title and the token are separate purchases. A web3 game minus its economy is an HTML5 or Android build with an extra SDK and an extra disclosure obligation. Evaluate it the way you evaluate anything else in that part of a catalogue: load time, session length, retention, device floor, localisation, and whether the licence lets you ship it where you operate.
- Never accept a chain metric in place of an audience metric. If a licensor quotes wallets, ask for daily active users, median session length and D1/D7 retention. If those do not exist, the game has not been played enough to have them, and the wallet number is measuring interest in a token.
- Assume the "earn" framing is a liability you inherit. You are the publisher of record on your own store listing. The licensor's marketing language is not a defence.
- The 2026 direction is abstracting the wallet away. Trade coverage this year consistently describes a gameplay-first turn in which chain mechanics are hidden from the player entirely. If that completes, the thing you are licensing looks like an ordinary game to an ordinary player, which is both the good news and the reason the category's own metrics will get harder to read, not easier.
🔍 Four Questions That Deflate a Market Number
None of this is really about blockchains. Portal operators get sent category forecasts constantly — for cloud gaming, for instant games, for regional markets, for anything with a growth story. Four questions do most of the work:
- Revenue or turnover? Ask whether the total counts money earned by publishers or money that moved. In categories with resale, secondary markets or in-game currency, those differ by an order of magnitude, and the bigger one is always the one in the headline.
- What is the unit? Wallets, installs, devices, accounts, registrations and humans are six different populations. Any of them can be called "users" in a chart title. Find out which one you are looking at before you divide anything by it.
- Is the base year measured or modelled? A forecast built on a modelled base year compounds an estimate. When two firms disagree tenfold on the base and converge on the horizon, the horizon is the assumption and the base is the guess.
- Does the segment fit inside its parent? The fastest sanity check there is. Put the sub-category next to the industry total from a source that has to survive contact with publishers. If the child is bigger than the parent, stop reading and go find the scope note.
🧭 What to Do With the Next Forecast You Are Sent
Invert the order you are used to. Before you open anyone's market report, write down the three numbers that would have to be true for the purchase in front of you to work: plays per month, net revenue per thousand plays, and the licence cost amortised over the term you can actually hold the games. Those numbers come from your traffic and your licence terms, not from a category CAGR. A forecast can then only do one of two things — support a case you already built, or fail to. It can no longer create one.
Then apply the four questions above to the report itself, and keep the answers with the file. Six months from now, when someone asks why you did or did not move on a category, the scope note is the part you will wish you had written down. The $24 billion and the $229 billion are both still sitting there, in public, describing the same year.
🌲 Where This Fits With Forestry Games
Forestry Games has licensed HTML5 and Android games since 2017 and now carries 1,049 titles, web3 games among them. The reason the category sits in the catalogue rather than in a separate pitch is the argument above: these are games, evaluated on the same criteria as everything else on the shelf. Operators sizing a portal are welcome to browse the full catalogue and judge titles on play data and licence scope. That is a duller conversation than a trillion-dollar forecast, and it is the one that survives a budget review.
