He spent a career finding things that were meant to stay hidden. Now he builds machines to find copper buried under a mile of rock.
The architect who decided rock was just another thing to redesign.
Yair Frastai runs a company that doesn't own a single mine, yet spends its days finding them. VerAI Discoveries points artificial intelligence at the one place geologists struggle most: the deposits nobody can see.
The premise is almost absurd in its ambition. Most of the world's easy mineral deposits - the ones that outcrop, that glint at the surface, that centuries of prospectors already stumbled onto - are gone. What remains is buried under gravel, forest, ice, and time. Frastai's bet is that the signal is still there, in the raw geophysics rippling up from deep underground, and that a machine trained to read it can find what human intuition misses.
So VerAI feeds its models magnetic, gravimetric, electromagnetic, and seismic data - the earth's deep murmurs - and asks a blunt question: where is the economic deposit, right now, under this covered ground? Not where are the surface clues that hint at one. The deposit itself. "We deploy a novel proprietary AI technology to search directly for those concealed economic mineral deposits," Frastai has said, "skipping the underperforming proxies analysis."
It is a strange place for a former intelligence officer with a degree in architecture to end up. It is also, when you look closely, exactly where someone like that would end up.
There are no longer 'easy' discoveries when it comes to mining. Most of the underexplored parts of known mining jurisdictions are under covered terrain.- Yair Frastai, Co-Founder & CEO, VerAI Discoveries
Frastai likes to state the problem in one merciless number. Today, only 1 in 1,000 exploration projects ever becomes a mine. That is not an industry with a quality problem. That is an industry playing a lottery, one drill hole at a time, spending real capital on odds that would embarrass a casino.
His argument is that the math is broken because the method is old. For a century, exploration has leaned on proxies - the faint, indirect hints near the surface that a deposit might lie below. Proxies are cheap to read and easy to misread. VerAI's pitch is to skip them and aim the model straight at the target, using data that comes from the depths rather than the topsoil.
The claimed payoff is not subtle: VerAI describes results that are roughly 100x more accurate, 20x faster, and 20x cheaper than the traditional approach. Whether the industry agrees is exactly what the company's growing portfolio is built to prove.
Figures as described by VerAI. Relative bar lengths are illustrative.
Three careers that shouldn't connect. Frastai connected them.
He trained as an architect at Bezalel Academy of Art and Design in Jerusalem, earning a B.Arch. Architecture is the discipline of imagining structure you cannot yet see and making it hold. It turns out to be surprisingly good preparation for modeling what sits beneath solid ground.
He served in the Israeli Intelligence special operations forces, where he pushed through what colleagues describe as far-reaching paradigm shifts in planning and operations. For that work he was awarded the Israel Defense Prize. The through-line: finding signal in noise, under pressure, when the target does not want to be found.
For more than a decade he worked as a mining-sector executive across the Americas, leading exploration R&D and field operations - including as head of R&D at QPX in Chile and the US. He collaborated with research institutes from MIT to CSIRO to Imperial College. This is where the problem found him.
In 2020 he and Amitai Axelrod - a partner of more than twenty years - founded VerAI. Two people who had spent careers running multidisciplinary teams against impossible discovery problems, now aimed at the hardest one in mining: what's hidden under cover.
Serves in Israeli Intelligence special operations; drives paradigm shifts in planning and operations. Awarded the Israel Defense Prize.
Becomes a mining-sector executive, leading exploration R&D and operations across North and South America; head of R&D at QPX (Chile & US).
Co-founds VerAI Discoveries with Amitai Axelrod to find concealed mineral deposits with AI.
VerAI raises a $12M Series A led by T. Rowe Price Associates, alongside Orion, Chrysalix, and Blumberg Capital.
VerAI secures $24M Series B led by Insight Partners; Jeff Horing joins the board.
Here is the twist that makes VerAI more than a science project: it isn't really a mining company. Frastai built a project generator. VerAI finds the high-probability targets, then partners with explorers and miners who bring the capital and the drills. VerAI keeps skin in the game through earn-ins, equity exposure, and royalty streams.
"Our business model is to be a scalable project generator rather than a sole explorer or miner," Frastai explains. "This is a capital-light approach." In a sector where a single deep drill program can swallow millions, refusing to be the one holding the shovel is not timidity. It is strategy. The software scales; the shovels don't.
The portfolio spans copper, gold, silver, nickel, and cobalt across Chile, Peru, Ontario, and Arizona - including a target roughly 18 km south of Freeport-McMoRan's Sierrita mine. More than 60 projects, generated by a company whose core asset is a way of seeing.
Our business model is to be a scalable project generator rather than a sole explorer. This is a capital-light approach.- Yair Frastai
"Today, only 1 in 1,000 exploration projects become a mine."
"We are feeding our machines with the right data for the problem: geophysics, signals coming from the depth of the subsurface."
"We are unlocking a multibillion-dollar opportunity by discovering under-cover deposits."
"We see a US$100 billion opportunity to disrupt the industry by providing high-probability targets in underexplored covered areas."
Strip away the algorithms and VerAI is a bet on a supply crisis. The energy transition runs on copper, nickel, cobalt - metals that the world will need in quantities the current pipeline of discoveries cannot supply. The mines that will feed the 2030s and 2040s mostly haven't been found yet. And the reason they haven't been found is precisely the reason Frastai built VerAI: they're hidden under cover.
It is a tidy piece of logic. The most valuable deposits are the hardest to find. The hardest to find are the ones under covered terrain. And covered terrain is exactly what surface-clue exploration is worst at. Frastai's move is to change the tool, not the target - to make discovery systematic where it has always been a matter of luck and stubbornness.
Whether AI can truly out-see a century of geological instinct is the open question of his career. The $24 million from Insight Partners is a vote that it might.
His degree is in architecture - from an art academy - not geology or computer science.
He holds the Israel Defense Prize, among the country's top defense recognitions.
He and co-founder Amitai Axelrod worked together for over twenty years before VerAI.
VerAI finds deposits but lets partners dig, keeping equity and royalties instead.
The models eat magnetic, gravimetric, electromagnetic, and seismic signals - not surface samples.
One Arizona target sits ~18 km from Freeport-McMoRan's Sierrita mine.
This new round of funding endorses our proprietary and validated AI-driven discovery platform - one that has generated a portfolio of high-probability mineral projects.- Yair Frastai, on VerAI's 2025 Series B
Yair Frastai is the co-founder and CEO of VerAI Discoveries, a Boston-based company using proprietary AI to find concealed mineral deposits buried under covered terrain. A former Israeli Intelligence special operations officer and Israel Defense Prize recipient turned mining-sector executive, Frastai built VerAI in 2020 with longtime partner Amitai Axelrod to attack what he calls mining's biggest problem: only 1 in 1,000 exploration projects ever becomes a mine. VerAI reads raw geophysical signals from deep underground with machine learning, and has raised roughly $39 million to date, including a $24M Series B led by Insight Partners in 2025.
Last updated: