EIC Accelerator Winner Hypermine and the Ore Sorting Race

Whether rock goes to the mill or the waste dump is usually decided from block model averages of 3,000 to 10,000 tonnes, according to Hypermine Global. The company, a 2025 spin-off of Finland's VTT Technical Research Centre, says it can make that call for each truck load of about 100 to 400 tonnes, by lighting the rock with a broadband laser and reading the reflected spectrum. Both claims come from the company's seed round release with VTT of 18 June 2025, which tech.eu reported the same day.

The EIC Accelerator picked the project, Minext, at its 1 October 2025 cut-off. The EIC's list of selected companies names Hypermine Global Oy/Ltd with a recommendation for blended finance, grant plus equity. CORDIS records a grant of €2,497,767.50, signed by the Commission on 5 June 2026, for a project running from 1 July 2026 to 30 June 2028. Four months in, there is little to check yet beyond what the project promised and what the company had done before it.

A laser instead of daylight

Hypermine's sensor reads minerals from the spectrum of light the rock reflects, and it brings its own light source rather than relying on daylight, which the company says lets it work in all ambient conditions. The CORDIS summary describes a supercontinuum laser-based active hyperspectral sensor paired with machine-learning software that classifies the minerals. A Maintworld profile of November 2025 describes the same sequence: a broadband laser hits the rock, a sensor measures what returns, and algorithms work out the composition in under a minute.

The company website lists scanning of rock faces underground, muck piles in open pits, excavator and truck loads, and conveyor belts after the primary crusher, at a range of up to 50 metres and one to two minutes per scan. It says the hardware needs no on-site calibration and integrates with fleet management systems.

Hypermine is not alone in trying to shrink the decision unit. MineSense, a Vancouver company, puts X-ray fluorescence sensors in the brow of a shovel bucket; its 2022 AusIMM conference abstract describes cutting the selective mining unit from a block of about 1,500 cubic metres to a bucket of 45 cubic metres or less. Closer to Hypermine's own method, US patent 12,523,529, granted on 13 January 2026 to CGG Services and now assigned to Sercel, covers active hyperspectral imaging with supercontinuum light for mining and field use.

An intellectual property database lists two pending patent applications under Hypermine Global Oy, in the United States and Canada, titled "Active hyperspectral mineralogy sensing for mining" and filed on 29 February 2024, a year before the company was registered. The same listing shows the HYPERMINE trademark registered at the EU Intellectual Property Office on 4 March 2026. Who filed the 2024 applications and how they passed to the company is not shown there; the seed release says VTT invested intellectual property in the company.

Years of research before the company

The sensor work runs back at least to GoldenEye, a Horizon 2020 project on satellite, drone and in-situ sensing for mines that VTT coordinated from May 2020 to October 2023. A conference paper by Kamen Bogdanov, Mikhail Mekhrengin, Francisco Senna Vieira and others, presented in December 2020 and acknowledging GoldenEye funding, reported active hyperspectral sensing of altered rock from gold and copper deposits and said the team was extending the light source's coverage to 2,500 nanometres. Hypermine Global Oy itself appears in the Finnish trade register data from February 2025, with no earlier name.

The seed round followed in June 2025: private equity from Stephen Industries and Sage Enterprises and an intellectual property investment from VTT, per the VTT release and tech.eu. Neither gives an amount, and no figure could be found elsewhere.

The leadership has changed since. The June 2025 release names Markku Makkonen as CEO and Mikhail Mekhrengin as CTO. The Maintworld profile of 30 November 2025 and the company website both give Mekhrengin as CEO, and the website lists Jake Greenberg, the Sage Enterprises managing director quoted in the release, as chairman. No announcement of the change was found.

Pilots, paid work and no named mine

The June 2025 release says the technology was co-developed and validated with a leading mining company, and that pilots were running in the United Kingdom, South Africa, Brazil and Canada; a VTT portfolio roundup of May 2026 repeats the four countries. The CORDIS summary says the sensor was "validated through paid pilots (TRL 6) with a TOP5 mining company". Maintworld quotes the company as saying its first customer is one of South America's largest mines, and that more than 95% of over ten large mining firms it spoke to wanted to pilot immediately. None of these sources names a mine or a customer.

The first accounts point to small but real income. Asiakastieto's summary of the 2025 accounts shows revenue of €238,000, three staff and an equity ratio of 96%. Fonecta gives €239,000 and three staff, a rounding difference. Asiakastieto's table shows an operating loss of €140,000, while its own stated operating margin of minus 43.5% implies a loss nearer €104,000; the full income statement is not freely available, so which figure is right could not be checked.

What Minext has to deliver

The project's own description is brief. The EIC Accelerator, it says, "will fund the transition to TRL 9 via scale-up, certification and large-scale pilots", from the technology readiness level 6 reached in the paid pilots. CORDIS gives no total project cost and, as the project only started on 1 July 2026, no report or deliverable. The selection list recommended blended finance, but CORDIS records only the grant, and no EIC Fund investment in Hypermine has been announced in the sources searched on 4 November 2026.

1. Validated sensor at TRL 6Company claimPaid pilots with a top-five miner per CORDIS; the mine is not named.
2. Scale-up to TRL 9Too earlyThe project runs to 30 June 2028.
3. CertificationToo earlyNo certification is named in the project summary or on the website.
4. Large-scale pilotsToo earlyEarlier pilots in four countries, per the June 2025 release; none reported under Minext yet.
5. Measured benefits at a mineNot publishedThe 3% to 10% profitability and 15% water figures are claims, with no site data in public.

Mining efficiency and Europe's raw materials plan

The EU's Critical Raw Materials Act, Regulation (EU) 2024/1252, in force since 23 May 2024, sets 2030 benchmarks for strategic raw materials: domestic extraction of at least 10% of annual EU consumption, processing of 40%, recycling of 25%, and no more than 65% from any single non-EU country, which the IEA describes as voluntary. The extraction benchmark is the part of EU policy closest to Hypermine's market. The VTT release frames the company's case through falling ore grades, citing a 30% decline in Chile's average copper grade over 15 years. The pilot countries the company has given so far are all outside the EU: the UK, South Africa, Brazil and Canada; the CORDIS summary does not say where the large-scale Minext pilots will run.

The cut-off was a crowded one. The EIC said on 17 October 2025 that 923 companies had filed full applications asking for nearly €7 billion, and on 17 February 2026 that 121 had been interviewed and 61 selected, with €467 million of proposed funding and 85% on blended finance. Hypermine's grant agreement was signed within four months of that announcement, according to the CORDIS signature date.

Hypermine was not contacted for this article. The reporting is based on publicly available sources, listed below.