By Gowtham Raj, Director at TartLabs, who oversees mobile application development work for enterprise, manufacturing and retail clients.
The Answer in Brief
Blockchain changed mobile app development in three ways, and none of them is the one the 2021-era pitch decks promised.
First, it stopped being an app category and became a payments dependency. The blockchain feature that actually ships in 2026 is a stablecoin settlement rail inside an otherwise conventional app — a remittance app, a payouts flow, a treasury function. Nobody downloads it for the blockchain.
Second, the binding constraint moved from cryptography to policy. Apple's own guidelines name cryptocurrency wallets as a prohibited way to unlock app features. That single sentence invalidates most "token-gated app" designs outright, before a line of code is written.
Third, the tooling market consolidated hard. Of six well-known embedded-wallet vendors, five were acquired between June 2025 and July 2026. Choosing a stack this quarter is now as much a vendor-survival question as a technical one.
The ceiling is regulatory, not technical. The EU's transition window for unlicensed crypto services closed on 1 July 2026, and the trigger for licensing is control, not blockchain — a distinction that determines whether your feature is a weekend's work or an eighteen-month licensing project.
Key Takeaways
- Apple's Guideline 3.1.1 lists "cryptocurrencies and cryptocurrency wallets" among the mechanisms apps may not use to unlock features. Token-gating is not a grey area on iOS; it is named and prohibited.
- Apple also requires that NFT ownership "does not unlock features or functionality within the app." You may let users view their NFTs. You may not let the NFT do anything.
- A US court order rewrote Apple's crypto rules. Apple's own 1 May 2025 developer notice says the guidelines were "updated for compliance with a United States court decision." The resulting carve-out applies only to the United States storefront.
- Google Play explicitly exempts self-custody: "Non-custodial wallets are out of scope of the Cryptocurrency Exchanges and Software Wallets policy." Take custody and, in the US, you must be FinCEN-registered with a state money-transmitter licence or a chartered bank; in the EU, authorised as a crypto-asset service provider (CASP) under MiCA.
- The EU grandfathering window is gone. ESMA stated that the MiCA transitional period "will officially expire across the EU on 1 July 2026" — that was nearly three months ago.
- MiCA blocks the white-label custody shortcut. ESMA notes MiCA "prohibits CASPs from outsourcing or delegating certain services, namely custody, to entities not authorised as CASPs themselves." Being a B2B supplier does not transfer the obligation.
- Gartner's 2017 forecast set a checkpoint we can now grade. It projected a $176bn blockchain market by 2025. The two firms publishing a 2025 figure put it at $31.18bn (Fortune Business Insights) and $57.7bn (Grand View Research) — an overshoot of roughly 3x to 5.6x.
- The four firms selling forecasts of this market cannot agree on the present, let alone the future. Their 2026 figures range from $47.96bn to $108.3bn, a 2.3x spread for a year already three-quarters elapsed.
- The "560 million crypto owners" statistic is a 2024 estimate still being quoted as current — Triple-A's page says "As of 2024," and carries no 2025 or 2026 edition.
- India taxes this at 30% plus cess — 31.2%, before any surcharge — with no expense deduction and no loss set-off, and FIU-IND requires a live walkthrough of your KYC, transaction-monitoring, blockchain-analytics and sanctions-screening systems before registration.
Every figure above is sourced below, with the source named and the date it was read.
What "Blockchain in a Mobile App" Actually Means Now
Strip away the vocabulary and there are only four things a mobile app does with a blockchain: hold a key, show a balance, move value, or prove something was recorded. Everything else is one of those four wearing a costume.
That matters because the four are not equally hard. Holding a key and showing a balance are ordinary client-side engineering. Moving value is a regulated financial activity in most of the world. Proving a record exists is usually solvable with a database and a signature, and the honest answer is that most "blockchain for provenance" briefs we see do not survive the question "who, other than you, needs to verify this without trusting you?"
The useful way to categorise a blockchain feature in 2026 is therefore not by chain or by SDK. It is by whether your app can move a user's assets without the user's key. That one question sets your app store review path, your licensing exposure and the bulk of your compliance cost. We will come back to it.
The Rule That Kills Most Web3 App Ideas
Start with the constraint that no amount of clever architecture routes around.
Apple's App Review Guidelines, Guideline 3.1.1, state:
"If you want to unlock features or functionality within your app, (by way of example: subscriptions, in-game currencies, game levels, access to premium content, or unlocking a full version), you must use in-app purchase. Apps may not use their own mechanisms to unlock content or functionality, such as license keys, augmented reality markers, QR codes, cryptocurrencies and cryptocurrency wallets, etc."
(Apple App Review Guidelines, read 23 September 2026.)
Read that list again. Cryptocurrencies and cryptocurrency wallets sit in the same enumerated category as licence keys and QR codes — mechanisms Apple considers attempts to bypass in-app purchase. The token-gated community app, the "hold our NFT to access the premium tier" model, the wallet-as-login-that-also-grants-entitlements pattern: on iOS these are not risky, they are named.
The NFT clause in the same guideline closes the obvious workaround:
"Apps may use in-app purchase to sell and sell services related to non-fungible tokens (NFTs), such as minting, listing, and transferring. Apps may allow users to view their own NFTs, provided that NFT ownership does not unlock features or functionality within the app. Apps may allow users to browse NFT collections owned by others, provided that, except for apps on the United States storefront, the apps may not include buttons, external links, or other calls to action that direct customers to purchasing mechanisms other than in-app purchase."
(Apple App Review Guidelines, Guideline 3.1.1; the duplicated "sell and sell" is Apple's own typo, present in the live text as read on 23 September 2026.)
So: users may look at their NFTs. The NFT may not do anything. That is the whole of the permitted design space.
The US carve-out exists because of a court, not a change of heart
The "except for apps on the United States storefront" clause above has a specific cause. Apple's developer notice of 1 May 2025 says:
"The App Review Guidelines have been updated for compliance with a United States court decision regarding buttons, external links, and other calls to action in apps. These changes affect apps distributed on the United States storefront of the App Store, and are reflected in updates to Guidelines 3.1.1, 3.1.1(a), 3.1.3, and 3.1.3(a)."
(Apple Developer News, 1 May 2025, read 23 September 2026.)
That is Apple stating in its own words that the rule changed because a court required it. The underlying contempt ruling in Epic Games v. Apple was affirmed by the Ninth Circuit on 11 December 2025 (opinion, case 25-2935), and Apple's applications to stay it were denied — the Supreme Court docket shows an application for a stay "denied by Justice Kagan" in May 2026 (Supreme Court docket 25A1213, read 23 September 2026). The Court subsequently granted certiorari "limited to Question 1 presented by the petition" (docket 25-1311, read 23 September 2026) — a question about the standard for civil contempt, not about the steering rules themselves.
The practical reading for a product team: the external-link freedom is real today, it is US-only, and it rests on litigation that is still live. Do not architect a business model whose viability depends on it surviving 2027.
One more Apple change most coverage missed
In its 13 November 2025 guideline update, Apple added crypto exchanges to the list of "highly regulated fields" under Guideline 5.1.1(ix), which requires submission "by a legal entity that provides the services, and not by an individual developer" (Apple Developer News, 13 November 2025, read 23 September 2026).
Guideline 3.1.5 has long required the same organisational enrolment for wallets specifically:
"(i) Wallets: Apps may facilitate virtual currency storage, provided they are offered by developers enrolled as an organization."
"(iii) Exchanges: Apps may facilitate transactions or transmissions of cryptocurrency on an approved exchange, provided they are offered only in countries or regions where the app has appropriate licensing and permissions to provide a cryptocurrency exchange."
(Apple App Review Guidelines, read 23 September 2026.)
Same guidelines also prohibit on-device mining and restrict ICO-adjacent trading apps to established financial institutions. None of this is new in substance. What is new is that "crypto exchange" is now an explicitly enumerated regulated category, which in practice means a slower, more documentation-heavy review.
Google Play Draws the Line in a More Useful Place
Google's policy is narrower and, for engineers, considerably clearer. The Play policy page on cryptocurrency exchanges and software wallets contains the sentence that should shape your architecture:
"Non-custodial wallets are out of scope of the Cryptocurrency Exchanges and Software Wallets policy."
If you are in scope — that is, custodial — the requirements are jurisdictional and specific:
"The developer must be either (a) registered with FinCEN as a Money Services Business and with a state as a money transmitter or (b) a federal or state chartered bank entity."
"The developer must be authorised as a crypto-asset service provider (CASP) under the markets in crypto-assets (MiCA) regulation by a relevant national competent authority."
(Google Play — Understanding the Cryptocurrency Exchanges and Software Wallets Policy, read 23 September 2026.)
The listed jurisdictions each carry a named regulator: Bahrain, Canada, Hong Kong, Indonesia, Israel, Japan, the Philippines, South Africa, South Korea, Thailand, the UAE, the UK, the United States, and the EU member states — with Liechtenstein, Iceland and Norway added, per the same page, from July 2026.
Read the escape clause carefully, because it is widely misreported. Google does not say you are barred elsewhere:
"If your targeted location is not on the list, you may continue to publish cryptocurrency exchanges and software wallets. However, due to the rapidly evolving regulatory landscape worldwide, developers are expected to obtain any additional licensure requirements per local laws. Google Play may also request you to provide further information regarding your compliance in a jurisdiction not listed below."
India is not on the list. That does not mean an Indian-market wallet is un-shippable on Play — it means Play imposes no named licence requirement there, while still expecting local compliance and reserving the right to ask you to evidence it. Your Indian obligations come from FIU-IND and the PMLA, not from Google. We return to those below.
Separately, Play's blockchain-based content policy requires apps selling or awarding tokenised digital assets to declare this through the Financial features declaration in Play Console and to flag the in-app product as a tokenised digital asset. NFTs bought in games are expected to be used in the game rather than functioning as pure speculation, and paying for a chance at an NFT of unknown value is treated as gambling (Google Play — Blockchain-based Content, read 23 September 2026).
One honest caveat: neither Play policy page displays a "last updated" date. Widely-reported announcement and enforcement dates for the wallet licensing rules (August and October 2025) circulate in the trade press but do not appear on Google's own policy pages, and the Play policy-deadlines page carries no crypto entry. Treat the requirements as current and verifiable; treat the dates as reported rather than official.
The Only Question That Matters: Who Holds the Key?
Here is the distinction that governs everything downstream. Both app stores, and all three major regulatory regimes, draw the same line in almost the same place.
| What your app does | Regulated? | Where that comes from |
|---|---|---|
| Renders a UI; the user holds their own keys | Generally no | MiCA Recital 22; Google Play's non-custodial exclusion; GENIUS Act §2(7)(B) |
| Operates protocol code or a self-custodial interface | Expressly excluded (US) | GENIUS Act §2(7)(B) |
| Can move user funds without the user's key | Yes | MiCA Title V; Google Play FinCEN/CASP requirement; FIU-IND |
| Swaps assets or converts to fiat in-app | Yes | MiCA Title V; Apple 3.1.5(iii); FIU-IND |
| Issues a payment stablecoin to US users | Yes, from the GENIUS Act's expected effective date | GENIUS Act §3(a) |
MiCA's Recital 22 states that where crypto-asset services "are provided in a fully decentralised manner without any intermediary, they should not fall within the scope of this Regulation" (Regulation (EU) 2023/1114, EUR-Lex, read 23 September 2026). The US GENIUS Act, now Public Law 119-27, carries a parallel exclusion: its definitions carve out those merely developing or operating distributed ledger protocols, self-custodial software interfaces, or validating transactions (Public Law 119-27, govinfo, read 23 September 2026).
The rule, stated plainly: the regulatory trigger is control, not blockchain. If the user holds the keys and you only render an interface, three of the world's largest regimes and both app stores currently place you outside the perimeter. The moment your backend can move assets without the user's signature, you are in licensed-activity territory in the EU, the US and India simultaneously.
For teams building custom software with a value-transfer component, this is the single most consequential architectural decision in the project, and it is usually made implicitly — by whoever decides where the private key lives — rather than deliberately.
The EU Compliance Window Has Already Closed
If your roadmap assumed there was still time to sort out EU licensing, there is not.
ESMA's statement on the end of MiCA transitional periods is unambiguous:
"The MiCA transitional period will officially expire across the EU on 1 July 2026. After this date, any entity providing crypto-asset services to EU clients without a MiCA licence will be in breach of EU law and must cease offering such services."
"By 1 July 2026, when the transitional period ends across the EU, any unauthorised CASP must have implemented its wind-down plan."
(ESMA75-113276571-1679, 17 April 2026, read 23 September 2026.)
Member states had run different grandfathering windows — six months for the Netherlands, Poland, Finland and others; twelve for Germany, Ireland and Austria; eighteen for France, Italy, Spain and the remainder — all measured from 30 December 2024 (ESMA list of Article 143(3) grandfathering periods, read 23 September 2026). Every one of them has now expired.
And there is a clause in the same ESMA statement that development agencies and white-label vendors specifically need to read:
"…entities established outside the EU are, outside the narrow exception of reverse solicitation, not permitted to provide crypto-asset services that qualify as MiCA services to EU investors or to solicit EU clients with a view to provide MiCA services to them. For clarity, this also applies in a business-to-business context, as MiCA specifically prohibits CASPs from outsourcing or delegating certain services, namely custody, to entities not authorised as CASPs themselves."
Being the backend supplier does not launder the obligation. A non-EU vendor cannot provide custody as white-label infrastructure to an EU-facing app unless that vendor is itself an authorised CASP. If you are scoping a project where custody is quietly outsourced to a partner, that partner's authorisation status is a gating question, not a procurement detail.
The United States: One Law Passed, One Did Not
Two things are commonly conflated here, so be precise.
The GENIUS Act is law. Public Law 119-27, signed 18 July 2025. Its core prohibition — "It shall be unlawful for any person other than a permitted payment stablecoin issuer to issue a payment stablecoin in the United States" — is not yet operative. The statute's effective date is the earlier of 18 January 2027 or 120 days after final regulator rules. Treasury's own August 2026 announcement of proposed rulemaking refers to "the expected effective date," and rulemaking was still at proposal stage then (US Treasury press release, 17 August 2026, read 23 September 2026). The Act also exempts peer-to-peer transfers between individuals without an intermediary and transactions using personal software or hardware wallets.
The CLARITY Act is not law. The broader market-structure bill passed the House but failed to reach the 60-vote cloture threshold in the Senate on 15 September 2026 — eight days before this article was written. It is not enacted, and no post claiming otherwise should be trusted.
Meanwhile the SEC's enforcement posture has shifted markedly. The Commission dismissed its action against Coinbase with prejudice in February 2025. Commissioner Peirce wrote: "Today the Commission settled its case against Coinbase by dismissing it with prejudice," adding that "the current Commission has now issued such a directive" — a direction that policy staff, rather than enforcement, should lead — "with the formation of the Crypto Task Force: It is the policy staff who will take the lead in engaging with the public to build a regulatory framework that serves the American public" (SEC statement, 27 February 2025, read 23 September 2026). An enforcement retreat is not a safe harbour. The statutes are unchanged, the clarifying legislation failed, and postures reverse with administrations.
What This Looks Like From India
TartLabs builds from India, and a large share of our client work is either for Indian businesses or for international clients whose apps reach Indian users — much of it consumer commerce, where payment rails matter most, as we covered in how mobile apps are transforming e-commerce in India. The Indian position on blockchain is genuinely distinctive: not accepted, not banned, comprehensively taxed.
The Parliamentary Standing Committee on Finance concluded a year of hearings on virtual digital assets in September 2026. Its chairperson, Bhartruhari Mahtab, summarised the position bluntly on 16 September 2026: the government "is not accepting virtual digital assets, doesn't want to regulate it, but not regulating it also leaves greater scope for different types of indulgences" — describing these as "the grey areas which need to be tackled," with a report pending the government's formal reply (as reported by ANI and The Tribune, 16 September 2026, read 23 September 2026). This is live: re-check it before making commitments.
What is settled is the tax and AML treatment, and both are punishing.
Section 115BBH of the Income-tax Act taxes income from transfer of a virtual digital asset at "the rate of thirty per cent," allows "no deduction in respect of any expenditure (other than cost of acquisition, if any)," and permits "no set off of loss from transfer of the virtual digital asset" (Income Tax Department, Section 115BBH, read 23 September 2026). Adding the 4% cess takes that to 31.2%; any applicable surcharge pushes the effective rate higher still. Section 194S imposes 1% tax deducted at source (TDS) on consideration for a VDA transfer to a resident, above thresholds of ₹50,000 for specified persons and ₹10,000 otherwise (Section 194S, read 23 September 2026).
The product consequence is the part teams miss: if your app is the party paying consideration, your app is the TDS deductor. That is a withholding, filing and reconciliation workload sitting inside your payments stack — Form 26Q/26QE/26QF obligations, not a footnote for the finance team.
Then there is registration. The 15 September 2025 circular from the Financial Intelligence Unit – India (FIU-IND), the country's financial-intelligence and anti-money-laundering authority, lists five trigger activities, including limb (iv): "safekeeping or administration of virtual digital assets or instruments enabling control over virtual digital assets." Note the phrasing — instruments enabling control is broader than custody, and it is the clause a product team must argue against if shipping anything key-adjacent in India. The circular is explicit that "Registration with FIU-IND is pre-requisite," and failure to register "is in itself a non-compliance with the provisions of PMLA" — the Prevention of Money Laundering Act.
The onboarding gate is where the engineering bill of materials becomes concrete:
"At the time of in-person meeting, it is also required to make live demonstration/walk-through of AML/CFT compliance processes and systems i.e. live demonstration of tools/systems pertaining to KYC, Transaction Monitoring, Blockchain Analysis, Travel Rule Sanction Screening etc."
(FIU-IND circular, 3rd revision, 15 September 2025, read 23 September 2026.)
India's regulator expects a working demonstration of four named subsystems: KYC, transaction monitoring, blockchain analytics, and Travel Rule sanctions screening. That is not a compliance line item. That is a second product.
What Genuinely Got Easier
It would be unfair to present this as all constraint. Three specific changes did materially lower the engineering cost of a blockchain feature, and all three are verifiable against ethereum.org.
Dencun (13 March 2024) introduced blob-carrying transactions via EIP-4844, which collapsed Layer 2 data costs and took typical L2 transaction fees from the tens of cents to fractions of a cent (Ethereum Foundation, read 23 September 2026).
Pectra (7 May 2025) shipped EIP-7702, which lets an ordinary externally-owned account temporarily behave as a smart contract (Ethereum Foundation, read 23 September 2026). This is the upgrade that matters most for mobile UX, because it enables gas sponsorship and transaction batching without migrating the user to a new address. Before it, giving a user a smart account meant giving them a different wallet.
Fusaka (3 December 2025) added PeerDAS, and the subsequent BPO forks raised blob target/max from 6/9 to 10/15 and then 14/21 (Ethereum Foundation, read 23 September 2026), so those cheap fees survive contention rather than spiking under load.
The through-line: Dencun made the data cheap, Pectra made sponsored and batched UX possible without address migration, Fusaka kept the cheap data cheap. Together they removed the two objections that killed consumer blockchain features in 2021 — it costs a dollar to tap a button and the user must write down twelve words.
Adoption reflects it. As read on 23 September 2026, BundleBear's dashboards show ERC-4337 smart accounts with at least one user operation at 67.4 million, against 1.29 billion total user operations; and EIP-7702 at 59.5 million unique delegated externally-owned accounts against 252 million authorisations (BundleBear ERC-4337 and EIP-7702 overviews, retrieved 23 September 2026). For scale: in the first week after Pectra, roughly 11,000 EIP-7702 authorisations had been recorded (The Block, read 23 September 2026). Sixteen months later it is tens of millions.
A note on those numbers, because it matters: these are live-dashboard reads on our access date, not a fixed-period statistic, and they count on-chain events rather than people. Circulating listicle figures such as "200 million smart wallets" have no published methodology and should not be repeated.
The current reference implementation is EntryPoint v0.9.0, released 16 November 2025 (eth-infinitism/account-abstraction releases, read 23 September 2026). A number of 2026 explainers still present v0.8.0 as current; they are out of date.
The Vendor Layer Consolidated Underneath Everyone
The embedded-wallet vendor layer is now owned almost entirely by payments companies, exchanges and chains — which should change how you run a stack-selection exercise, because the shortlist you evaluate on features is really a shortlist of other companies' strategic interests.
Of six embedded-wallet vendors we would expect to see on a 2026 shortlist — a selected set, not a census of the category — five were acquired inside fourteen months. Web3Auth went to Consensys and is now MetaMask Embedded Wallets. Privy went to Stripe. Dynamic went to Fireblocks. Sequence went to Polygon Labs. Magic Labs sold its wallet business to Payward, Kraken's parent — an asset sale, with wallet customers transitioning to Payward Services while Magic itself rebranded as Newton Labs. Turnkey is the one still standing alone.
None of these are failures, and most of the products continue. But the selection question has changed. It is no longer which embedded wallet SDK is technically best; it is which of these will still be independently roadmapped in eighteen months, and whose strategic interest now sits upstream of mine? An embedded wallet owned by a payments company, an exchange, or a chain has an owner with opinions about where your transactions settle.
There is a naming hazard here too. WalletConnect the company is now Reown; the dApp-side SDK is AppKit; the wallet-side SDK was WalletKit and has been folded back into something called the WalletConnect SDK; and the network and token sit under a separate WalletConnect Foundation. Four names, one lineage. AppKit is actively shipping — version 1.8.24 on npm published 16 September 2026, with Reown Swift 2.4.0 and Reown Kotlin BOM 1.6.17 both published 14 September 2026 (npm, GitHub releases, read 23 September 2026).
For anyone auditing an inherited codebase: WalletConnect v1 has been dead since 28 June 2023, when its bridge servers went offline. Any tutorial or dependency still referencing v1 is a reliable signal the material is stale.
The hardware bet is a more mixed picture. Solana's first phone, the Saga, was discontinued roughly two years after launch at around 20,000 units against a 50,000 target, leaving owners without further security updates (as reported by Decrypt, read 23 September 2026). Its successor, the Seeker, actually shipped — from August 2025, to more than 50 countries, against roughly 150,000 pre-orders at $450–500 (as reported by The Block, read 23 September 2026). Both are trade-press figures rather than audited disclosures, and we could find no unit total published by Solana Mobile itself. Taken together, the category is neither dead nor proven.
The Numbers You Should Not Repeat
House rule here: if a statistic cannot be traced to a primary source, it does not go in the post. Applying that rule to this topic eliminates most of what circulates.
The forecasts do not agree about the present
Four firms sell forecasts of the same market. For 2026 — a year already three-quarters gone — they give $47.96bn, $54.08bn, $62.91bn and $108.3bn. That is a 2.3x spread on a measurement, not a projection.
Their terminal forecasts diverge further still: Fortune Business Insights at $577bn by 2034, MarketsandMarkets at $611bn by 2031, Precedence Research at $2,380bn by 2035, and Grand View Research at $9,056bn by 2033. Grand View's figure is roughly 15x Fortune's, and it arrives a year earlier. When four vendors modelling one market differ by more than an order of magnitude, the disagreement is about definitions, not about the market, and quoting any single figure as "the blockchain market" is a category error.
A specific trap: Grand View's most-linked press release still says $1,431.54bn by 2030 — and it is dated February 2022. Its current report page says $9,055.5bn by 2033. Anyone citing the first number in 2026 is quoting a release the firm itself has superseded. We were unable to load grandviewresearch.com directly (it returned HTTP 403 on repeated attempts), so the Grand View figures here are reported as per its published report page rather than as independently verified.
Gartner's forecast has a checkpoint we can now grade
Gartner's widely-quoted "$3.1 trillion in blockchain business value by 2030" comes from a forecast published in March 2017. That same forecast contained a nearer milestone: $176 billion by 2025.
2025 is behind us. Of the four firms, two publish a 2025 figure: Fortune Business Insights at $31.18bn and Grand View Research at $57.7bn (the Grand View leg per its published report page, which we could not load directly). Against either, the forecast overshot its first checkpoint by roughly 3x to 5.6x.
A projection whose first testable milestone missed by that margin has no remaining claim on anyone's 2030 planning. We could find no formal withdrawal of the $3.1 trillion figure, and equally no reaffirmation of it in the nine years since — and we could find no dedicated Hype Cycle for Web3 and Blockchain later than the July 2024 edition. Note which blockchain-adjacent item did survive into Gartner's 2025 general emerging-technologies list: programmable money. Which is, in practice, stablecoins.
The ownership statistic everyone quotes is two years old
"Over 560 million crypto owners worldwide" is the most-repeated number in this field. It comes from Triple-A, whose page states: "As of 2024, we estimated global digital currency ownership at an average of 6.8%, with over 560 million digital currencies owners worldwide." As read on 23 September 2026, that page still headlines the 2024 estimate, and no 2025 or 2026 edition is referenced (Triple-A, read 23 September 2026). The ownership rate also appears inconsistently across Triple-A's own material as both 6.8% and 6.9%.
So the field's favourite statistic is a two-year-old estimate whose publisher has not refreshed it, quoted daily as though it described today.
The most current alternative is Crypto.com's H1 2026 market sizing, which reports owners "rising from 741 million in December 2025 to 774 million in June 2026" (Crypto.com Research, 10 August 2026, read 23 September 2026). Its published methodology is, in full, "based on our on-chain estimates" — no sampling frame, no de-duplication method, no confidence interval, published by an exchange with a commercial interest in the number rising. Cite it as Crypto.com estimates, never as fact.
Chainalysis is the most methodologically defensible source here and deliberately does not produce the number people want: its Global Adoption Index is a relative ranking of 151 countries weighted by population and purchasing power, not an ownership headcount (Chainalysis, read 23 September 2026). Anyone citing "Chainalysis says X million people own crypto" has misread it.
And the enterprise-adoption percentages? The most recent edition of Deloitte's Global Blockchain Survey we could find is 2021, and the most recent Gartner survey figures we could trace are from 2019. We found no credible measurement of enterprise blockchain adoption more recent than that, which means any "X% of enterprises use blockchain in 2026" figure is almost certainly recycled from those or invented outright.
Where the Real Numbers Are: Stablecoin Settlement
One area does have defensible measurement, and it happens to be the one area where blockchain is genuinely inside mainstream mobile apps.
Stablecoin volume is real and large, but it is reported two incompatible ways and the gap is enormous.
Raw volume for full-year 2025 was about $33 trillion, up 72% year on year, per Artemis data reported by Bloomberg (8 January 2026). Adjusted volume was about $10.2 trillion for the twelve months to June 2026. That is Visa Onchain Analytics' measure, which strips out high-frequency trading, bots, bridge routing and internal exchange operations. It counts only addresses staying under 1,000 transactions and $10 million of transfers in any 30-day window.
Visa's own documentation includes a worked example in which $3.9 trillion of raw 30-day volume reduces to $817.5 billion adjusted. That is roughly an 80% haircut. Both numbers are correct; they measure different things. Quoting either without the qualifier is the single most common error in stablecoin coverage.
Two figures circulating in 2026 that we could not verify and would not use: claims of "$7.2 trillion adjusted in February 2026, surpassing ACH" (roughly 4x Visa's adjusted monthly figures for the same period, source unclear) and an "~$26 trillion per year adjusted" attribution to Artemis that appears on no Artemis page we could find.
Where Stablecoin Settlement Actually Ships: Western Union's USDPT
The concrete example: on 4 May 2026 Western Union announced USDPT, a US dollar payment stablecoin issued by Anchorage Digital Bank N.A. and built on Solana, describing it as "an always-on settlement asset… that eliminates the latency and fragmentation of traditional correspondent banking rails" (Western Union investor relations, 4 May 2026, read 23 September 2026).
Read the release carefully, though, because the sequencing is the interesting part. Western Union lists four services it "is developing" around USDPT: global exchange support, a Digital Asset Network connecting exchanges and custodians to its payout infrastructure, treasury and agent settlement, and — separately — "Stable by Western Union, a consumer-facing spend capability launching in 2026 in 40+ countries."
In other words, the settlement layer is the part that went live. The consumer-facing piece is a stated plan. Even the flagship example of blockchain in mainstream payments is not yet a blockchain feature inside a consumer mobile app; it is a blockchain rail underneath one, with the app layer scheduled to follow. MoneyGram and Remitly have moved along similar lines, adding stablecoin settlement behind conventional app interfaces. The pattern that works is blockchain as invisible infrastructure beneath a normal app, chosen because it settles faster and cheaper across borders than correspondent banking; the pattern that does not is blockchain as a visible feature the user is asked to care about. For how we weigh feature-level investment decisions more generally, see our analysis of the ROI of mobile app development.
What We Would Actually Build in 2026
Condensing all of the above into the advice we give on live projects:
Start from the custody question, not the chain question. Decide first whether your backend can move user assets without the user's key. That decision sets your app store path, your licensing exposure and most of your compliance cost. Chain selection is a detail by comparison.
Default to non-custodial unless custody is the business. Both stores treat it more permissively, MiCA Recital 22 and the GENIUS Act both carve it out, and FIU-IND's trigger list is aimed at control over assets. If you need custody, the realistic route on a normal budget is a licensed partner holding both the licence and the keys.
Never put entitlements behind a token on iOS. Guideline 3.1.1 names the mechanism. Use server-side entitlements and treat the wallet, if present, as an identity signal only.
Assume your wallet SDK vendor will be acquired. Five of six were, in fourteen months. Keep the integration behind an interface you control, and avoid vendor-specific key formats that make migration a user-visible event. This is an ordinary portability discipline, and it interacts with your framework choice — the trade-offs we weigh there are covered in our comparison of native, cross-platform and hybrid mobile app development.
Use EIP-7702-era account abstraction to hide the chain entirely. Gas sponsorship and batching without address migration is what makes a blockchain feature survive usability testing. If your users see a seed phrase, the design is four years out of date.
Budget the compliance subsystems as product. If India is in scope for anything custodial, FIU-IND expects a live demonstration of KYC, transaction monitoring, blockchain analytics and sanctions screening. Those are four systems, not four checkboxes — and they need people who have built them before, which is usually an argument for hiring a dedicated team rather than retraining an existing one mid-project.
Quote no market forecast. As shown above, they disagree by 2.3x about a year already in progress. If a business case rests on a market-size projection, the business case is decoration.
If you want a second opinion on whether a blockchain component belongs in your roadmap at all — including the answer "it does not" — get in touch. We would rather scope it out early than bill for it twice.
The Bottom Line
Blockchain did change mobile app development — by ceasing to be a thing users interact with.
The durable pattern is settlement infrastructure underneath an ordinary app: stablecoin rails that move value across borders faster and cheaper than correspondent banking, invisible to the person tapping the button. The pattern that did not survive is blockchain as a feature users are asked to notice, and the app stores wrote that outcome into policy. Apple names cryptocurrency wallets as a prohibited unlock mechanism. Google exempts non-custodial wallets and gates everything else behind money-transmission licensing.
What that leaves engineering leaders is a much simpler decision than the discourse suggests. Ask who holds the key. If the user does, you are building ordinary software with an unusual settlement layer, and the recent Ethereum upgrades have made that genuinely pleasant to build. If you do, you are building a regulated financial institution with an app attached, and the EU's window for doing that without a licence closed on 1 July 2026.
Everything else — the market forecasts that disagree by 2.3x about the present, the ownership estimate from 2024 still quoted as though it were current, the 2017 projection that missed its first checkpoint by 3x to 5.6x — is noise you can safely ignore when the business case lands on your desk.
If you are weighing a blockchain or stablecoin component for a mobile product, or you have inherited one and want an honest assessment of what it will cost to maintain, get in touch. Our mobile application development team scopes these the same way we have described here: control first, chain last.
This article describes regulatory requirements as we read them on 23 September 2026. It is general information, not legal advice. MiCA authorisation, FIU-IND registration and US money-transmission analysis are all questions for qualified counsel in the relevant jurisdiction, and the position in India in particular was still moving at the time of writing.
Every figure here names its source; unless another date appears inline, linked pages were consulted on 23 September 2026 and may change without notice. Where a widely repeated number could not be traced to a primary source, it is labelled as untraceable rather than repeated — and where a primary page could not be loaded at all, as with Grand View Research, that is stated at the point of use. Two circulating stablecoin figures are marked explicitly as claims we could not verify.
