Alan Carter sat down with Paul Harris at the 2026 Precious Metal Summit in Beaver Creek and said the quiet part in a calm voice. Cabral Gold has poured gold at Cuiú Cuiú, in Pará, Brazil. The first pour came about two months ahead of the old schedule. A second pour has followed. He expects to declare commercial production in November, in the fourth quarter, in line with the plan. For the first twelve months he is thinking about 25,000 ounces. He will not give a full 2027 guide until early January, because the plant is still ramping.
That is the idea, and it is the only one. The ounce Cabral can sell this year is an oxide ounce. About three quarters of the 1.2 million ounces he cited still sits in hard rock the phase-one plant does not treat. The loan that built the plant is repaid in gold, so a higher price raises the cash and the bill together. A boulder that averages 91 grams is not a resource. Commercial production is the event. Everything else is a name. This piece is not a recommendation to buy or sell Cabral or any other security.
What a pour is, and what it is not
Harris asked the question that separates a headline from a mine. What happens between the first pour and commercial production? Carter's answer was the right size. The first pour is part of commissioning. Commissioning means every piece of equipment has to run as specified. Cabral has a dry circuit, with conveyors and a lot of moving parts. The wet circuit is more complicated. The company hired a commissioning team out of Perth to watch the gear against the specs. There are always teething problems. Throughput has to be optimized. It usually takes several months. He says they are on that schedule.
A pour is a test that gold can leave the circuit. It is not a proof that the circuit will do it every day at the rate in the study. Commercial production is the later mark, when the plant is stable enough to count. Carter put that mark in November. Until that word is in a filing, the ounces are commissioning ounces. They can be sold. They should not be modeled as a year.
The company has already sold metal. A late-September report said more than 2,400 ounces went out the door at an average net price above US$4,200, after a first pour of about 1,130 ounces on September 10. At Beaver Creek, Carter told Harris the cash balance before gold sales was about C$55 million, and that sales would push it up. The till is moving. A moving till is not the same thing as a declared commercial mine.
The study was written at $2,500
Carter was clear about the old model, once the tape is cleaned up. The July 2025 pre-feasibility study, done with Ausenco, used a gold price of US$2,500 an ounce. He did not say $25,000. That figure is a transcription slip, and it would make the rest of his math nonsense. At Beaver Creek he put the spot price near US$4,300. The study's all-in sustaining cost was US$1,210 an ounce. On the company's own page, stage-one capital is US$37.7 million. After-tax IRR is 78 percent at $2,500. NPV at a 5 percent discount is US$74 million. Payback is 10 months. Initial production is 25,000 ounces a year. A later account of the same study puts life-of-mine output near 113,000 ounces over about 6.2 years, with a strip ratio of 0.78.
His cash line for next year was about US$75 million of free cash, on roughly 25,000 ounces and today's price. The arithmetic is easy to see and easy to over-trust. Twenty-five thousand ounces times a margin of about $3,000 is $75 million. That margin is a study cost subtracted from a spot price. It is not a cost the plant has posted for a year. Power, reagents, rain, recoveries, and a slow ramp can eat it. He also said production should grow after year one, and that he will not lock 2027 until January. Treat $75 million as his illustration. Do not treat it as a budget.
The price gap is still the point of the hour. A study written at $2,500, with costs near $1,210, is a different business at $4,200 or $4,300. The extra dollars fall to the owner only after the plant runs, the loan is served, and the oxide lasts. Carter knows the first condition is not finished. He is in commissioning. The second condition is a gold loan. The third is geology.
They borrowed gold, and they repay gold
The plant was funded with a US$45 million gold loan from Precious Metals Yield Fund, an affiliate of Phoenix Gold Fund, which the company describes as its largest institutional shareholder. The deal was announced in October 2025 and closed in November. Interest is 10 percent, which Carter called competitive. The term, as he stated it, is 39 months. They borrowed gold, turned most of it into cash, and kept some as gold. They pay the fund back in bullion, not in a dollar cheque.
He is still finishing the mechanics. The first interest payment is due at the end of December. Another interest payment and a principal payment fall in March. A published note on the loan says principal is repaid in kilograms of gold starting March 31, 2027. He told Harris he expects to pay the loan down faster than the term. At $2,500, the study's payback on the gold loan was 10 months. He said they borrowed a bit more than the study's capital. At $3,500, he said, a capital bill he rounded to $38 million would pay back in six months. He does not plan to hedge the price just to soften the loan. He thinks the extra revenue outruns the extra dollar value of the ounces owed.
That last sentence is the risk, and it is also the logic. A gold loan does not get cheaper in ounces when the price rises. The quantity stays put. The dollar value of what you owe goes up. So does the dollar value of what you mine. Both sides win, he said, because the mine's extra revenue is larger than the loan's extra value. That is true only if the ounces are actually produced. If the ramp slips, the price rally is a bigger bill and a smaller till. If the price falls, the dollar bill shrinks and the margin shrinks with it. The loan is matched to the product. It is not matched to a delay.
Cash of about C$55 million, before the new sales, is a cushion. It is not a reason to forget the December interest or the March principal. A cushion spent on drills, before the plant is commercial, is a choice to fund the district with money that was raised to build the oxide mine. He says there is now enough cash flow to do the targets justice. The order matters. Oxide first. Then the names.
Three quarters of the gold is not in this plant
Carter put the current resource at Cuiú Cuiú at about 1.2 million ounces, indicated and inferred together. The last full update of all the deposits was in September 2022. The company has since published oxide figures in later reports. On its own summary, primary indicated material is about 450,000 ounces and primary inferred about 455,000. Oxide indicated is about 216,000 ounces. Oxide inferred is about 71,000. Add them and you are near 1.2 million. Oxide is roughly a quarter. Hard rock is roughly three quarters. That matches what he told Harris. About 75 percent of the 1.2 million ounces is below the saprolite.
Phase one mines the saprolite. He said that weathered layer is about 60 metres thick. The plant now running is for that soft material. Phase two is the hard rock underneath. Most of the ounces in the old global number are phase two ounces. They are not in the circuit that just poured. An investor who buys "1.2 million ounces in production" has bought a sentence with the wrong verb. About a quarter of those ounces are the kind this plant can see. The rest are a reason to do a study.
He expects a new global resource by the end of 2026. In 2022 the model had three deposits. The update is modeling six. Jerimum Cima, one of the newer discoveries, is in that set and was not in the old base. He thinks the number grows, and then keeps growing, and that a producer should put out a resource every year. Growth in a resource is not growth in the plant. If the new ounces are oxide, they can feed the expansion he wants. If they are hard rock, they feed phase two, which does not exist as a mine yet. Read the update for the split, not for the headline total.
Doubling the soft ore is the near catalyst
He has already said, in recent weeks, that Cabral plans to expand phase one. Harris asked if that means 3,000 tonnes a day to 6,000. Carter said that is the objective. Guidance on how, and on cost and timing, is due in the next few weeks. He hopes that in about a year the phase-one capacity has doubled. He also said a preliminary economic assessment on phase two, the hard-rock mine, is likely to start, and that he would like to be talking about those results in roughly twelve months.
Put those in order, because he listed them as catalysts and the order is the discipline. This year: commercial production, the resource update, and guidance on the phase-one expansion. Drilling continues, with six rigs. Next year's exploration budget is not set. It will be at least six rigs. It may be more. In 2027: the doubled oxide rate, and the phase-two assessment. Exploration results will run alongside all of it. Harris warned him, kindly, that fifty targets and a new pour can blur the message. Carter agreed. The mine is the bread and butter. Exploration is upside. He said he will keep trying to put releases into a few plain points. The district is large. The names are hard to say. Clarity is now an operating job, not a slogan.
A double from 3,000 to 6,000 tonnes a day is not free because gold is $4,300. It needs ore, pads or plant, water, people, and a permit path he did not itemize in the interview. It also needs the oxide to be there in the quantity the new pits assume. He said drilling since the study has given more confidence on both oxide and hard rock. Confidence is not a reserve. The weeks-away guidance is the document that turns "objective" into a cost. Until it arrives, 6,000 tonnes a day is a direction.
The boulders are not the mine
Carter's most dangerous sentence, for a careless reader, was also his most honest. Cabral controls the district. He counts more than fifty targets outside the six deposits in the model. These are not geophysical guesses. Gold has been seen in drill holes, trenches, or boulder fields. One boulder field averages about 91 grams per tonne. Another, about five kilometres away, averages about 75 grams. Gold, he said, not silver. They have not had the money to explore them properly. Cash flow changes that. He has hired Elton Pereira, a Brazilian exploration geologist he rates highly, who has found gold and nickel deposits. Six rigs are turning. The regional work, starved while the plant was being built, can now be fed.
A boulder is a rock on the ground. A field of rich boulders says gold moved through that place. It does not say how much is still in the ground, at what depth, or whether it can be mined at $1,210 an ounce. Ninety-one grams in a boulder and half a gram in the oxide resource are both "gold at Cuiú Cuiú." They are not the same asset. The oxide indicated grade on the company's summary is about half a gram per tonne. That is the material this plant is built to treat, in bulk, by heap leach. The boulder is a reason to drill. The half-gram oxide is a reason the plant exists.
Harris was right to call the news flow a champagne problem. Fifty names, six rigs, a resource update, an expansion guide, and a phase-two study will fill a year of headlines. A shareholder who cannot say which headline feeds the plant will own a story. The plant does not eat stories. It eats saprolite. Carter said the focus stays on Cuiú Cuiú, and that other assets may be looked at and may be declined. That is the right sentence. A small producer with a district does not need a second country. It needs the first plant to stay boring and full.
What you can underwrite
You can underwrite a commissioning plant that is early, not a mine that is finished. Two months ahead on the first pour is a real mark. Commercial production in November is still a sentence until the company declares it. Between those dates the Perth team is hunting teething problems. Model a ramp. Do not model a switch.
You can underwrite year one as about 25,000 ounces of oxide, if he is right, at a study cost near US$1,210 that has not yet been proven in this plant. The margin to $4,200 or $4,300 is wide on paper. Wide on paper is how people skip the ramp. Free cash of US$75 million is his sketch of that margin. It is not a covenant. It also has to cover, or sit beside, a gold loan whose ounces do not shrink when the price rises.
You can underwrite the loan as ounces, not as a dollar that got cheaper. Ten percent was a fair rate in his telling. The December interest and the March principal are dates. Paying faster is his hope, and the cash balance is why the hope is plausible. It stays plausible only if gold keeps leaving the circuit. A higher price helps the hope. It does not replace the circuit.
You can underwrite the resource only after you split it. About 1.2 million ounces is the old global scale. About a quarter is oxide. About three quarters is hard rock. Six deposits will be in the year-end model, against three in 2022. Jerimum Cima is one of the additions. Ask, when the update lands, how many of the new ounces are soft enough for phase one. Those are the ounces that can justify 6,000 tonnes a day. The hard ounces justify a study, a PEA he hopes to discuss in about a year, and patience.
You cannot underwrite 91 grams. You cannot underwrite fifty targets. You cannot underwrite a district as if it were a second plant. You cannot underwrite phase two as if the oxide cash had already built it. Carter was plain that phase two is the larger mine coming behind, and that the PEA is still ahead. A PEA is a sketch. It is not a mill.
What would make this reading wrong
The reading is wrong, on the plant, if November comes and goes without commercial production, or if the ramp reaches it and then cannot hold 25,000 ounces over the first year. Early pours have disappointed before. A Perth team and a dry circuit do not repeal gravity, rain, or a bad recovery. If the year comes in near his ounce line and near the study cost, the margin is no longer a sketch. It is an operation.
The reading is wrong, on the loan, if the gold price falls hard while the ounces owed stay fixed, and the extra revenue he counts on is not there. It is also wrong, in his favour, if he pays the loan down as fast as he hopes and the December and March dates become a footnote. Watch the payments. They are the audit of the "both sides win" line.
The reading is wrong, on the geology, if the year-end resource shows a large rise in oxide, not just in hard rock, and the expansion guide prices a path from 3,000 to 6,000 tonnes a day that the new oxide can fill. Then phase one is bigger than the 2025 study, and the "quarter of the ounces" line is stale. If the growth is almost all hard rock, the line stands, and the PEA is the next honest document. Boulder grades will not decide it. Drill holes with widths will.
The idea, once
At Beaver Creek, Alan Carter told Paul Harris that Cabral Gold's Cuiú Cuiú plant in Pará is in commissioning, about two months ahead, with commercial production expected in November. Year one is about 25,000 ounces. The July 2025 study used US$2,500 gold, US$1,210 all-in sustaining cost, and US$37.7 million of capital, with a 10-month payback. He sees free cash near US$75 million at today's price, and a faster paydown of a US$45 million gold loan that is repaid in bullion at 10 percent. Cash before those sales was about C$55 million. The global resource he cites is about 1.2 million ounces, and about three quarters of it is hard rock below a saprolite layer about 60 metres thick. Phase one treats the soft ore. He wants to double it from 3,000 to 6,000 tonnes a day and to show a phase-two study in about a year. More than fifty targets sit outside six deposits, including boulder fields he says average 91 and 75 grams. Those are rocks on the ground.
Sell the story only after you know which ounce is in the plant. The oxide ounce is the mine. The hard rock is a study. The boulder is a reason to drill. The loan is a stack of ounces that do not care what the headline says.
A note on sources and limits
Carter's comments on timing, commissioning, year-one ounces, free cash, the gold-loan mechanics, the resource split, Jerimum, the tonne-per-day objective, and the boulder fields are from his interview with Paul Harris of Kitco at the 2026 Precious Metal Summit in Beaver Creek. The study capital, IRR, NPV, cost, payback, and the oxide and primary ounce counts are from Cabral's published project summary of the July 2025 pre-feasibility work and later resource notes. The loan holder, the 10 percent rate, and the closing window are from the company's description of the US$45 million gold loan. The first pour date, the first sale, and the price above US$4,200 are from a late-September 2026 account of Carter's remarks, not from the Beaver Creek tape alone. The "$25,000" figure in a raw transcript is a garble of US$2,500. Names garbled on the tape are corrected here, including Cabral, Cuiú Cuiú, Pará, Ausenco, and Jerimum Cima.
Nothing here is investment advice or a solicitation to buy or sell any security. Resources are not reserves. Inferred ounces are the least certain. A pre-feasibility study is not a promise of cost or recovery. Boulder samples are not a deposit. Gold loans can hurt a ramp that misses. Readers should read the technical report, the loan terms, and the filings, and should speak with a licensed adviser before any decision.