In August 2025 the Bangko Sentral ng Pilipinas gave every institution it supervises two days to delete a link. The memorandum is short enough to read in a minute, and I did, because the primary document is the only part of this story that cannot be briefed at you second hand. Memorandum No. M-2025-029, Office of the Governor, subject line “Suspension of In-app Gambling Access in Mobile Payment Apps and Websites”. The operative sentence: “all BSIs are hereby instructed to remove links providing in-app gambling access in BSIs’ mobile payment apps and websites within 48 hours upon issuance of this Memorandum.”
The links came down. GCash and Maya pulled their in-app gaming tiles inside the window, the compliance box was ticked, and the measure was widely reported as done. Fourteen months later, CoinDesk reports that the largest crypto exchange in the Philippines is facing punitive action from the same central bank over flows from the same industry, serious enough, according to one person quoted, to cost the twelve-year-old company its licence.
Nobody checked where the money went. That is not a moral complaint. It is a measurement complaint, and it is the ninety-fifth unasked half of fairness.
TL;DR
- BSP Memorandum M-2025-029 (August 2025) ordered every supervised institution to remove in-app gambling links within 48 hours. It says in its own text that the suspension runs “pending the issuance and implementation of the corresponding policy”.
- That policy is still, as of today, an exposure draft on the BSP’s own site headed “CIRCULAR NO. ____, Series of 2025”, with the resolution number and the date blank.
- CoinDesk reports that gambling entities migrated to Coins.ph after the e-wallet order, and that the exchange now faces regulatory action. Coins.ph “firmly denies any claim that it knowingly serves offshore gaming or gambling operators”.
- Every control in the draft circular is denominated in a named person’s account: 20% of average daily balance, a six-hour window, one account per eligible owner, facial biometrics. None of it survives a conversion into a bearer token.
- I measured the one on-chain peso rail the central bank has actually blessed. PHPC has 190,672 units outstanding on Ronin and 21,031 on Polygon, 67 holders, and a median transfer of 1.67 pesos. About 2.7 US cents.
- Satoshie’s draw is provable. Satoshie’s funding rail is somebody else’s licensed institution, and this story is about the rail.
What the memorandum actually instructed
Read the memo as a piece of engineering and it is precise about a very narrow thing. “In-app gambling access includes any product/service that redirects an account holder to a gaming/gambling site.” The regulated object is a redirect. The instruction is to remove it. Compliance is verifiable by opening the app and looking.
That precision is a virtue, not a flaw. Forty-eight hours is a realistic deadline for a UI change and a wildly unrealistic one for anything else, and the memorandum knows it. The sentence immediately before says the suspension was approved “pending the issuance and implementation of the corresponding policy that provides standards and expectations for BSIs in the provision of online gambling payment services”. In plain terms: we are switching this off now, and the real rules are coming.
So the memorandum is an interim measure with a stated dependency. The question that follows is not rhetorical, it is arithmetic. How long has the interim measure been the whole regime?
I went looking for the corresponding policy. What is on the BSP’s exposure drafts page is a document headed “CIRCULAR NO. ____, Series of 2025”, approved by “the Monetary Board, in its Resolution No. ___ dated ____________ 2025”, signed by nobody, dated “_______ 2025”. Nine pages of real substance with every identifying field still blank. Its glossary refers to the “Philippine Gaming and Amusement Corporation”, which is not PAGCOR’s name; the corporation is the Philippine Amusement and Gaming Corporation. That is the kind of error a document carries while it is still a draft and loses when it is signed. I could not find a final issuance of it. If one exists I would genuinely like to be corrected, because the entire shape of this post changes if the standards landed and I missed them.
Call it fourteen months between the 48-hour deadline and today, with the measure that was explicitly temporary still doing all the work.
Where the flow is reported to have gone
CoinDesk’s account, sourced to two people with knowledge of the situation and a third familiar with central bank processes, is this. After the e-wallet order, several gambling entities migrated to Coins.ph. The move had an added advantage, according to one of the sources: access to crypto and stablecoins for paying offshore partners. At the start of October 2026 the BSP imposed restrictions on deposits made through the exchange’s main wallet firm, DCPay Philippines. The restriction was not time-bound, unlike the temporary suspensions normally used, and no reason was given. Earlier this year its chief compliance officer, its head of AML transaction monitoring and its money laundering reporting officer all departed.
Coins.ph has responded, and the response belongs here in full force rather than in a footnote. The company says users’ funds are safe, that it continues to operate as a BSP licensed institution, that it expects to restore affected services soon, and that it “firmly denies any claim that it knowingly serves offshore gaming or gambling operators. We hold ourselves to the requirements and standards of the BSP, AMLC and other regulatory authorities, and full compliance sits at the core of how we operate.”
Here is the part I want to be careful about, because it is also the part that makes the argument worth making. Nothing in this post depends on the allegation being true. Coins.ph may be entirely vindicated. The structural point survives either way, because it is a point about what the rules can see, and a rule that cannot see something cannot see it regardless of whether anything is there.
The scale matters for a different reason. The exchange has over 18 million customers, roughly $50 million of daily spot volume, and controls something like 75% to 80% of peso-to-stablecoin liquidity in a country CoinDesk puts at around $43 billion of annual remittances. Whatever the outcome, an enforcement action at that institution is not a scalpel. It lands on a rail that millions of people use to send money home, and none of those people are party to anything.
Every control in the draft is denominated in the account
The draft circular is the most interesting document in this story and almost nobody has read it, so here is what it actually proposes. Payment service providers would need prior authority from the BSP, a supervisory rating no lower than 3, and “a minimum capitalization of ₱300,000,000”. Each eligible customer would get a separate Online Gambling Transaction Account, one per person, funded only by on-us transfers. Opening it would require “mandatory facial biometric verification for account opening and periodic facial biometric re-verifications”.
Then the limits. Transfers into that account are capped at “twenty percent (20%) of the average daily balance of the eligible owner’s transaction account”. The provider must set a transaction window that “should not exceed six (6) hours per day”. Heavy usage triggers a 24-hour cooling-off period. Opening the account disables every lending option on the same platform. Monthly reporting to the central bank would include, per user, a unique identifier, volume and value, “age, economic class, total number and amount of top-up, starting and ending balance of transaction account”.
Note the clocks, too. The interim memorandum gave institutions 48 hours to remove a link. The draft that would actually govern the payments gives providers already in the business “a period of six (6) months from the effectivity of this Circular” to come into compliance, and takes effect fifteen days after publication. The fast deadline was attached to the cosmetic change and the slow one to the substantive change, which is the correct way round and is exactly why the gap between them needed watching.
The sanctions section tells the same story. Violations draw a monetary penalty “not to exceed ₱100,000 per calendar day” for continuing breaches, or a maximum of ₱1,000,000 per transactional violation, with suspension of the authority on a first offence and revocation on a second. Every one of those penalties lands on the payment service provider. Not one of them is denominated in the flow. You can price a provider’s failure precisely and still have no number at all for the thing that moved.
This is a thoughtful set of rules. It is also, without exception, denominated in a named person’s account at a supervised institution. Twenty per cent of what balance? The one the institution holds. Six hours of whose day? The one the institution clocks. One account per eligible owner, verified by whose camera? The institution’s. Age and economic class reported by whom? The institution that knows.
Every single control assumes the regulator is looking at a ledger entry with a face attached to it. That assumption is load-bearing, it is never stated, and it is the thing that stops being true the moment a peso becomes a bearer token. There is no average daily balance of a stablecoin in a self-custodied wallet. There is no six-hour window on a chain that runs all of them. The draft is not wrong about any of its limits. It is silent about the conversion, and the conversion is the only step that matters.
We have written before about rules that were never in force and the behaviour they changed anyway, in the eighty-seventh instalment on FinCEN’s withdrawn proposals. This is the mirror image and the more awkward case. Here a rule was in force, it was obeyed inside 48 hours, and the effect it produced was a change of address.
So I went and measured the visible rail
If the attraction of a crypto exchange to a gambling operator is “access to crypto and stablecoins”, the natural question for anyone who writes about on-chain anything is: which stablecoin? There is an obvious candidate. Coins.ph issues PHPC, the first peso-pegged stablecoin, launched with central bank blessing. It is on-chain. It is auditable by anyone. If the peso flows that regulators want to see had moved onto a chain, this is the one instrument where everybody could see them.
So I measured it, this morning, from the chains themselves rather than from a dashboard.
PHPC’s main deployment is on Ronin, the chain Sky Mavis built for Axie Infinity. The contract at 0x63C6E9F027947BE84D390cFA7B2332D13B529353 is verified, compiled with Solidity 0.8.9, and named StableCoinsOfPeso. Its header credits an author at coins.ph. Total supply: 190,672 PHPC. Holders: 67. Lifetime transfers: 8,315. The Polygon deployment holds another 21,031 PHPC, a figure two independent RPC endpoints agree on.
Then the activity. I pulled the most recent 1,500 transfers on Ronin, which reach back to 22 September 2026, and sorted them. The median transfer is 1.67 PHPC. At today’s rate of about ₱62.80 to the dollar, that is 2.7 US cents. The largest transfer in those three weeks was 180 PHPC, about $2.87. There were zero mints and zero burns across the entire window.
Put the two numbers side by side. The complete outstanding supply of the Philippines’ first regulated peso stablecoin, across both of its chains, is 211,703 units. About $3,371. Its issuer trades roughly $50 million a day. The entire supply of the regulated, visible, peso-denominated on-chain rail is worth about six seconds of its own issuer’s volume.
Whatever moved, it did not move through PHPC. It could not have. The instrument is a pilot with sixty-seven holders passing three cents around a gaming chain.
That last phrase is not a jab. It is the most quietly strange fact in the whole story. The only peso stablecoin a central bank has blessed has its principal home on Ronin, a chain built to settle an on-chain game, and the transfer sizes look exactly like what that implies: dust moving between players. We have written about Ronin before, when it spent four years and $625 million learning what a general-purpose chain gives you for free. A regulated national-currency instrument living there, barely used, while the flows everyone is arguing about run through dollar-denominated tokens nobody in Manila issues, is a reasonable summary of where crypto’s visible surface and its actual plumbing have ended up.
One more detail from the contract, because it is the sort of thing you only find by reading the source. PHPC inherits from Pausable and BlackList, with an owner-only destroyBlackFunds that burns a blacklisted address’s entire balance. A comment in the file explains that the getters exist “to allow the same blacklist to be used also by other contracts (including upgraded Tether)”. The regulated peso instrument is, in its lineage, a copy of Tether’s contract. Its answer to misuse is owner-controlled confiscation. That is a coherent design choice for a supervised issuer, and it is the exact opposite of the property we build on, which is that nobody can reach into the mechanism after the fact.
Why this is a fairness question and not a compliance one
Gambling regulation almost never touches the game. It touches the venue and it touches the payment rail, because those are the two things a regulator can reach. We made the same observation about the CFTC’s raffle exclusion three days ago: the tests ask who holds a licence, never who picks the winner.
Follow that through and you get the uncomfortable structure of this whole story. A prohibition is enforced at a chokepoint. Its success is measured by absence at the chokepoint, which is cheap to verify, instead of by presence somewhere else, which is expensive and which nobody is tasked with. The links were gone in 48 hours, and that is a true statement about links. Fourteen months later the question of where the payments went is being answered by an enforcement action rather than by a report, which tells you the measurement never existed.
And the cost of that enforcement is borne in a pattern nobody designed. A non-time-bound deposit restriction with no stated reason does not land on an offshore gambling operator. It lands on the next person trying to receive money from a relative abroad through the country’s dominant peso-to-stablecoin rail.
What Satoshie can claim here, and what it cannot
The narrow claim first, because it is the only one I can prove. On Satoshie, entry stakes are escrowed on entry. The VRF coordinator is named in deployed code you can read before you commit anything. Chainlink’s proof is verified on-chain before our callback is allowed to run. The winner is computed inside that callback from a formula published in advance, and the payout happens in the same transaction. Ticket counts are readable contract state, so your odds are arithmetic over a public number before you enter. None of that depends on us being trustworthy, and none of it can be relocated, because the resolution logic does not live at a venue.
Now the limits, and this is the part of the post I would cut if I were selling something.
One: our players arrive through exactly these chokepoints. Every person who has ever entered one of our draws converted money somewhere, and that somewhere was a licensed institution with a compliance department and a regulator. Our permissionlessness begins at the contract and ends at the on-ramp. If a central bank restricted deposits at the exchange our players use tomorrow, our provable fairness would do precisely nothing for them. A provable game sitting downstream of an unprovable rail is still downstream.
Two: a permissionless draw cannot implement an eligibility rule, and that is a cost. The draft circular’s definition of an eligible account owner runs to eight conditions: at least 21, not a student, not a government official connected with operations, not a member of the armed forces or police, not a beneficiary of the 4Ps social welfare programme, not on PAGCOR’s restricted persons database. Whatever you think of each item, the list exists because somebody decided certain people should be protected from this. Satoshie cannot enforce a single line of it. We do not know who you are, which is the feature, and it means we cannot keep anyone out, which is the price. I would rather state that plainly than pretend permissionlessness is free.
Three: we published the optimistic version of this story and it has not aged well. In June 2026 we argued that the crackdown on anonymous crypto gambling would push players towards platforms that can actually prove their claims. The reported behaviour here runs the other way. When the e-wallet route closed, the operators are said to have moved to the largest licensed institution in the country, not to a verifiable one, because what they needed was banking access rather than credibility. Demand for provability and demand for a working rail are different demands, and we conflated them.
Three questions worth asking
When a payments regulator switches something off, what is the mechanism by which anyone finds out where it went, and who owns that measurement?
If every limit in a rulebook is expressed as a share of an account balance or a number of hours in a day, what does that rulebook do about an instrument that has neither?
If the only on-chain peso rail your central bank has approved has 67 holders and a median transfer of three cents, is the visible instrument the one being used, or the one being pointed at?
The closer
A 48-hour deadline is a wonderful thing to have. It is unambiguous, it is cheap to verify, and it produces a clean before and after. It is also, in this case, the only part of the regime that ever shipped. The standards it was pending have spent fourteen months as a document with a blank where the number goes.
Money does not wait for the number to be filled in. It goes to the next institution that will take it, and if that institution happens to sell the ability to turn a supervised peso into a bearer dollar, then the thing the rules measure and the thing that actually moved have quietly stopped being the same object. The links came down on schedule. The flow was never followed.
Our draws are provable because the proof is in the transaction, not in an attestation that something was switched off. That is a small claim about one mechanism, and it is the only kind worth making.
Sources read first-hand: BSP Memorandum No. M-2025-029; the BSP exposure draft “Regulations on Online Gambling Payment Services”; CoinDesk, 13 October 2026; the verified PHPC contracts on Ronin and Polygon, read via public RPC and the Ronin block explorer on 14 October 2026.
📷 Photo by Matt Benson on Unsplash
