The AI buildout’s copper problem is usually framed as a permitting problem. Find a deposit, spend a decade in hearings, hope the price still works. On August 12, Vale Base Metals took a different route and got a meaningful tonnage for a fraction of the money and a fraction of the time: it approved a Coarse Particle Flotation project at the Salobo Copper Complex in Brazil that adds up to roughly 30,000 tonnes of annual copper production by rebuilding how the existing plant separates rock [Vale Base Metals via PR Newswire, August 12, 2026].
No new pit. No new orebody. A new flotation circuit inside a plant that already exists.
What’s happening
- VBM is proceeding with CPF at Salobo, adding 6 million tonnes to the plant’s annual ore processing capacity and taking total processing capacity to 42 million tonnes per year. Incremental copper production is projected at up to approximately 30,000 tonnes per year contained in concentrate, carrying approximately 15,000 ounces of gold as a byproduct [Vale Base Metals via PR Newswire, August 12, 2026].
- Capex was cut to approximately $215 million from an initially projected $225 million to $275 million. Wheaton Precious Metals is contributing $40 million in two $20 million installments tied to project milestones, in lieu of future milestone payments under the existing Salobo streaming agreement. VBM’s own spend is expected at approximately $175 million, on an assumed BRL:USD of 5.20 [Vale Base Metals via PR Newswire, August 12, 2026].
- The company puts capital intensity at $5,000 to $6,000 per copper equivalent tonne and internal rate of return above 50 percent. Those are the company’s own figures, not an independent estimate, and VBM describes the capital intensity as well below the industry average [Vale Base Metals via PR Newswire, August 12, 2026].
- First commercial production is guided to the first half of 2028, approximately one year earlier than initially projected. VBM received the construction licence from IBAMA ahead of schedule [Vale Base Metals via PR Newswire, August 12, 2026].
- The mechanism matters. CPF, underwritten by the Eriez HydroFloat process, adds a flotation circuit that removes waste rock earlier in the processing stream, which reduces grinding requirements, relieves the grinding bottleneck, increases metallurgical recoveries and enables higher throughput. VBM says it does this with lower average energy and water usage compared to other methods [Vale Base Metals via PR Newswire, August 12, 2026].
The price backdrop is what makes 30,000 tonnes worth writing about. LME copper hit a record US$14,455 per tonne on August 6 and was US$14,290 on August 10, having first surpassed US$14,000 in May. The drivers cited are constrained supply, heavy investment in power grids, uncertainty around US tariffs and rising electrification demand, with data centres and power grids named as the primary driver [IndexBox, August 11, 2026, citing CNBC and Barchart’s William Osnato].
Brazil angle
Salobo is the largest copper operation in Brazil, and CPF is the second acceleration VBM has announced in weeks. CEO Shaun Usmar said the company announced in late July the early delivery of Bacaba, the first project in its copper pipeline, by six to nine months, and framed the one year acceleration at Salobo as the second [Vale Base Metals via PR Newswire, August 12, 2026].
Two things are worth separating here. The tonnage is modest against a global deficit. The method is not. VBM began CPF research and development for Salobo more than two years ago, with lab testing at its Sheridan Park Research Facility in Canada and field testing at a pilot plant in Brazil’s Carajás region, and it says it is now exploring additional CPF opportunities at its other copper projects in Carajás and across the portfolio, including at polymetallic orebodies [Vale Base Metals via PR Newswire, August 12, 2026]. If that transfers, Carajás becomes a district where incremental copper can arrive on a two year clock rather than a ten year one, sitting on a stated endowment of more than 53 million tonnes of contained copper in mineral reserves and resources including inferred resources [Vale Base Metals via PR Newswire, August 12, 2026].
The IBAMA construction licence arriving ahead of schedule is the part Brasília should notice. Brazilian permitting is normally the friction in this story. This time it was not.
US angle
This desk’s read is that the American conversation about the copper gap runs on two levers: opening new domestic supply, and adjusting what imported metal costs. Tariff uncertainty is itself one of the named drivers of the current copper price [IndexBox, August 11, 2026]. Neither lever produces a tonne quickly.
The Salobo decision points at a third lever that gets less attention: metallurgical retrofit of permitted, operating plants. Note that this is not permit free. VBM still needed a construction licence from IBAMA, and got it ahead of schedule [Vale Base Metals via PR Newswire, August 12, 2026]. But a recovery and throughput upgrade inside an operating footprint is a far shorter approval path than a new mine, and it reuses grid connection, water infrastructure and workforce that are already in place. For US operators with mature assets and grinding constraints, the barrier to copying this is technical capability and capital allocation, not geology. Worth noting where the capability was built: VBM did the lab work in Canada, at its Sheridan Park Research Facility, not in the United States [Vale Base Metals via PR Newswire, August 12, 2026].
China angle
China’s problem is the mirror image. Its smelting capacity has outrun global concentrate supply, pushing spot treatment and refining charges into negative territory and leaving smelters reliant on sulphuric acid byproduct revenue to stay viable. The China Smelters Purchase Team, a 16 member consortium, committed to cutting 2026 copper output by approximately 10 percent. Chinese refined copper output instead grew 7.4 percent in January to April 2026 against the same period of 2025, per China’s National Bureau of Statistics, and the group has withheld quarterly TC/RC guidance for six consecutive quarters [Discovery Alert, June 18, 2026].
Salobo produces concentrate, not cathode. Every incremental tonne of concentrate lands in the market where Chinese smelters currently hold the least leverage they have had in years.
What it means
The desk’s TAI-M sub-index, the pure materials read, stood at 101.8, with TAI at 105.9 and SDX at 98.3 at the last weekly recompute [Tantalum indexes.json, as of August 14, 2026]. Those values predate the August 12 announcement and should be read as such. Vale is an SDX constituent, and a capital light, high return brownfield add is structurally supportive of the SDX thesis, which is that the alternative supply base outside the China and DRC nexus can absorb Western capital and deliver. It does not follow that the index level moves on this news.
The broader read: the copper response to AI demand will not arrive as a wave of new mines inside this decade. New mine supply from the Democratic Republic of Congo, South America and Central Asia is generally placed on a two to four year horizon before it eases concentrate tightness [Discovery Alert, June 18, 2026]. Between now and then, the tonnes that actually show up come from recovery, throughput and grinding economics inside plants that already hold their licences.
What to watch
- VBM’s next named CPF site in Carajás. The company said it is exploring the technology at its other copper projects and at polymetallic orebodies. A second named site would confirm this is a district strategy rather than a single asset fix [Vale Base Metals via PR Newswire, August 12, 2026].
- Bacaba commissioning against the accelerated schedule. VBM claimed six to nine months of early delivery in late July. Watch whether the Q3 and Q4 2026 disclosures hold that date.
- The next TC/RC benchmark round. If Chinese smelters again fail to convert announced discipline into actual output cuts, concentrate producers keep pricing power, and Brazilian concentrate keeps commanding it [Discovery Alert, June 18, 2026].