Get Your Weekly Digest

Receive the best podcast summaries and insights delivered to your inbox every week.

We respect your privacy. Unsubscribe anytime.

TED Talks Daily

The emerging science of finding critical metals | Mfikeyi Makayi

with Mfakeyi Makai

Published September 24, 2025
View Show Notes

About This Episode

Host Elise Hu introduces a TED Talk by mining innovator Mfakeyi Makai about how the world's transition to electrification and a circular economy requires a massive increase in critical metals like copper, lithium, cobalt, and nickel. Makai explains that while ore deposits are abundant, the mining industry has underinvested in exploration and still relies on outdated methods, so her team at Kobold is using AI and machine learning to model subsurface geology, quantify uncertainty, and design more efficient, safer, and environmentally sustainable mines. She illustrates how their approach guides where to explore, when to stop drilling, and how to plan operations, highlighting the Mingamba project in Zambia as a prototype for the mine of the future.

Topics Covered

Artificial intelligence Renewable energy and decarbonization Sustainability Mining technology Critical minerals Electric vehicles Circular economy Machine learning Zambia Resource exploration

Disclaimer: We provide independent summaries of podcasts and are not affiliated with or endorsed in any way by any podcast or creator. All podcast names and content are the property of their respective owners. The views and opinions expressed within the podcasts belong solely to the original hosts and guests and do not reflect the views or positions of Summapod.

Quick Takeaways

  • The global shift to electrification and a circular economy will require more than 400 new mines for critical metals like lithium, copper, cobalt, and nickel by 2040.
  • Despite its importance, the mining industry spends less than a penny on exploration for every dollar returned to shareholders, leading to stagnating discovery rates and outdated technology.
  • Most ore deposits still exist but are harder to find because they are deeper and not visible at the surface, making intelligent prediction essential.
  • Makai's team at Kobold uses machines and machine learning to predict subsurface rock properties and rigorously quantify uncertainty instead of relying on a single geological model.
  • By simulating many possible ore body geometries and mine designs, their approach helps decide where to collect data, when to stop drilling, and how to minimize waste and water use.
  • Better predictions can make mines safer, more environmentally sustainable, and more resilient for local communities across commodity price cycles.
  • The Mingamba project in Zambia is being designed as a "mine of the future" using these AI-driven methods and diverse global and African talent.
  • Handling uncertainty explicitly, rather than ignoring it, is central to building responsible and efficient mining operations.

Podcast Notes

Show introduction and framing of the mining and critical metals topic

Host introduces TED Talks Daily and the day's idea

Elise Hu welcomes listeners to TED Talks Daily and notes the show brings new ideas to spark curiosity every day[2:26]
She identifies herself as the host, Elise Hu[2:34]

Framing the rise of electric vehicles and mineral demand

Electric vehicles are described as exploding in popularity and use around the world[2:35]
Elise notes that EVs require rare earth minerals in order to function[2:41]
She points out that those minerals must be mined, and mining is not known for sustainable practices[2:48]

Introducing the speaker and core idea of the talk

Host introduces mining innovator Mfakeyi Makai as the speaker[2:52]
Elise summarizes that Makai will share how her team is working to build the sustainable mine of the future[2:56]
The approach involves radical new AI-aided technology[3:01]
The goal of the technology is to maximize resource recovery while minimizing environmental impact[3:04]

Speaker background and why mining matters for modern life

Personal background in Zambia and mining

Makai states that she was born and raised in Zambia[3:09]
She describes Zambia as a country known for its rich copper mining history[3:18]
She says the "alignment of the stars" meant that by birth and by science she became a miner[3:20]

Everything is either grown or mined

Makai asserts that everything we build and use was either grown or mined[3:26]
• She lists examples: walls, windows, tables, and chairs
• She includes phones and computers as products dependent on mined or grown materials
• She notes that even the stage she is standing on is an example
• She points to her copper earrings and possibly the audience's jewelry as mined materials

Future circular economy and growing demand for critical metals

Vision of a circular, electrified economy

Makai says that when we talk about building a circular economy today, we mean we need to electrify everything[3:40]
She envisions future economies with cars and trucks powered by batteries[3:46]
• She adds robots, drones, and aircraft as examples of battery-powered systems in this future
She stresses that children will need computers in all schools for technical access[3:53]
She imagines data centers full of advanced chips to bring AI to us[3:58]
She notes that all of this will be sourced by abundant sources of renewable energy[4:02]

Metals required and scale of new mining needed

Makai says the raw materials needed will be recyclable so society can become clean and circular[4:09]
She states this means a lot more lithium, copper, cobalt, nickel and other materials are required[4:13]
She shares that more than 400 new mines must be built by 2040 for society to become circular[4:22]
She reminds that before building a mine, the raw materials have to be found[4:24]

Problems with current mining exploration and underinvestment in technology

Mining industry's limited contribution to quality of life through innovation

Makai argues that today's mining industry leaders are doing too little to advance our qualities of life[4:31]

Comparison of R&D investment with other industries

She compares mining to other discovery-driven industries like pharmaceuticals and technology[4:39]
In those industries, she explains, for every dollar returned to shareholders, about a dollar is spent in R&D[4:49]
In mining, by contrast, for every dollar returned to shareholders, less than a penny is spent in exploration[4:52]
She characterizes this as underinvestment and says it should not be surprising that exploration and mining technology has barely advanced[4:58]

Declining discovery effectiveness

Makai states that the industry has gotten 10 times worse in the last 30 years at making ore body discoveries[5:03]

Abundance of deposits but lack of information

Most ore deposits still exist but are harder to find

She offers a positive note: the vast majority of ore deposits are still out there waiting to be found[5:13]
They are more difficult to find because past mines were easy-they were poking out of the surface or near the surface[5:20]
She says that now we need to look deeper to find new deposits[5:25]

Challenging the idea that materials will run out

Makai notes that, controversially, people have been taught that these materials will run out[5:27]
She argues that we do not lack ore body deposits, but rather we lack information about where they lie[5:33]

Using AI and machine learning to predict subsurface geology

From crystal ball metaphor to prediction mindset

Makai uses a metaphor: with a crystal ball, one could look into the ground and dig out the best rocks that generate the least waste[5:41]
She emphasizes that since we do not have a crystal ball, the proper response is to make predictions of where materials lie[5:47]

Kobold's goals: predict everything and quantify the unknowns

Makai says her colleagues and she at Kobold are doing what the industry has neglected to do[5:52]
Their aim is to predict everything and quantify what they do not know[5:55]
They also aim to collect information efficiently[6:00]

Audience exercise: imagining subsurface rock properties

Makai asks the audience to predict the concentration of copper 1,000 meters below their feet where they are sitting[6:09]
She further asks them to predict how hard the rock is, how fractured it is, and its density[6:13]
She says her team aims to predict all these properties and more[6:19]
They are developing machines and machine learning technologies to help make these predictions and rigorously quantify uncertainties[6:23]

How exploration data is collected and why it is ambiguous

Aerial geophysical surveys and their limitations

Makai explains that when exploring for mines, teams often fly aircraft thousands of kilometers across the Earth[6:37]
These aircraft collect information such as the Earth's magnetism and gravitational field[6:39]
Such measurements tell geologists something about the rocks beneath the surface[6:45]

Infinite possible subsurface models from limited data

She notes a fundamental problem: for each dataset, there can be an infinite number of possible subsurface configurations[6:51]
The problem arises because they are building three-dimensional models to fit two-dimensional data[6:55]
She gives an example: a body that is smaller and closer to the surface versus a larger body farther away can produce the same measurement[6:59]

How the incumbent industry currently handles uncertainty

Makai states that the incumbent industry deals with this ambiguity by ignoring it[7:07]
They choose one possible answer and act as though other possibilities do not exist[7:09]
As a result, the industry designs sub-optimal mines and makes sub-optimal decisions[7:15]
She says this often leads to mining unnecessary material[7:17]

Kobold's alternative: modeling many possibilities with AI

Simulating many rock arrangements instead of choosing one

Makai says they have invented a different way[7:21]
They collect all the possibilities that are consistent with the measured data[7:25]
They do this by simulating the physical response of each possible arrangement of rocks[7:27]

Using AI to accelerate physics-based predictions

Makai explains they achieve this 10,000 times faster by training an AI[7:33]
The AI learns the relevant physics of the rock beneath[7:35]
She contrasts this with the conventional method, which can only test one configuration in the same amount of time[7:39]
This speedup means they can collect better data and make better predictions of where to look next[7:45]

Optimizing drilling decisions by targeting uncertainty

Example of a dense rock body and drilling strategy

Makai describes a hypothetical rock body that is denser than the surrounding material[7:49]
She notes that a typical approach might be to drill through the middle of such a body[7:53]

Targeting areas of greatest uncertainty

When they have hundreds of thousands of possible subsurface solutions, the best strategy is to collect data where uncertainty is highest[7:55]
By drilling in areas of maximum uncertainty, they can rigorously eliminate as many possibilities as possible[8:03]
This approach enables them to maximize the information obtained for every dollar spent[8:09]
They repeat this process to continually quantify their uncertainties[8:11]

From discovery to defining ore body size and shape

Uncertainty persists even after a discovery

Makai emphasizes that even once an ore body discovery has been made, uncertainty remains[8:17]
The challenge then is to define the size and shape of the ore body[8:21]

Thought experiment on spatial prediction from a single drill hole

She illustrates the difficulty by extending the earlier exercise[8:28]
Now, 1,000 meters below the audience member, a drill sample shows rock with 5% copper[8:30]
This gives one data point and observation[8:34]
She then asks the audience to predict the concentration of copper beneath the person sitting next to them[8:40]
She challenges them to consider how confident they would be in that prediction[8:46]
She escalates the challenge: predict the copper concentration below a person across the room[8:52]
She then asks about predicting beneath the next building, and then the next city[8:56]
She states that this illustrates the vast challenge: they have sampled only a tiny fraction of rock, separated by several football fields, yet must predict properties in between[9:04]

Applying the technology in Zambia to move quickly

Makai says their technology has helped them move fast in Zambia, where she comes from[9:12]
They have been able to design and develop a mine based on their predictions[9:16]
She reiterates that this design is derived from a situation where they have only sampled a tiny fraction of rock[9:20]

Using uncertainty quantification to guide drilling and project decisions

Many ore body scenarios and what uncertainties represent

Makai explains there are many possible ore body configurations consistent with the available data[9:52]
Some possibilities contain a lot more metal, while others contain less[9:28]
The difference between these scenarios serves as a measure of the uncertainties[9:30]

Deciding where to drill and when to start building

These quantified uncertainties enable them to know where they should collect information next[9:34]
They help determine where to drill the next hole[9:36]
The same framework also indicates when drilling can stop and when it is time to start building a mine[9:38]

CoboMine and optimizing mine design under uncertainty

Limitations of designing a mine from a single model

Makai notes that to build the mine of the future, they must continue to grapple with uncertainty[9:42]
She explains that the industry typically designs an entire mine based on a single model of the ore body[9:46]

Introducing CoboMine as a mine design optimization tool

Makai says they are developing CoboMine, a mine design optimization tool[9:50]
CoboMine evaluates many possible mine designs against the many possible ore body geometries previously discussed[9:54]
This allows for the best decisions on how much ore will be mined and how much waste will be produced[10:02]
The tool also informs how much water will be used and the expected cash flows[10:04]

Planning infrastructure and operations for efficiency and minimal waste

CoboMine supports decisions about where to place permanent infrastructure such as a shaft[10:12]
It helps plan where traffic and tunnels will be placed to enable efficient operations[10:16]
The aim is to maximize the ore and metal extracted while minimizing waste[10:20]
Makai adds that this technology will move into mine operations to guide day-to-day decisions for efficiencies[10:24]

Impacts: safety, environment, and community resilience

Beyond profitability: safety and environmental benefits

Makai stresses that better predictions are not only about profitability[10:30]
They enable a safer mine by revealing where rocks are weaker[10:36]
They also support an environmentally sustainable mine by helping lessen environmental impact[10:38]

Creating resilient mines that support communities

Makai says better predictions make for a resilient mine with cash flows that can support local communities and businesses[10:44]
This resilience is important across different commodity pricing cycles[10:50]

Mingamba project and call for responsible mining transformation

Mingamba as the mine of the future

Makai states that their Mingamba project in Zambia will be the mine of the future[10:52]
The project is being designed and developed by talent from around the world[10:58]
She highlights that this includes Zambians and Africans like herself[11:00]

Recognizing ongoing material demand and industry responsibility

Makai acknowledges the reality that the need for these materials will continue to grow[11:04]
She links this growth to advancing lifestyles that will demand more materials[11:08]
She argues that the mining industry must transform to become responsible miners[11:14]
The industry needs to build better mines with better technology[11:16]
Makai closes with "Asante and thank you"[11:20]

Outro and context of the talk

Identifying the talk and event

Elise Hu says, "That was Mfakeyi Makai"[11:23]
She notes that the talk was given at the TED Countdown Summit in Nairobi, Kenya in 2025[11:27]

Mention of TED's curation guidelines and production credits

Elise invites curious listeners to learn more about TED's curation at ted.com/curationguidelines[11:31]
She says that concludes the episode for the day[11:53]
She notes that TED Talks Daily is part of the TED Audio Collective[11:35]
She mentions the talk was fact-checked by the TED Research Team[11:39]
She lists production and editing team members: Martha Estefanos, Oliver Freedman, Brian Greene, Lucy Little, and Ida Gillespie[11:45]
She credits Christopher "Fazy" Bogan for mixing the episode[11:55]
She acknowledges additional support from Emma Taubner and Daniela Balarezo[11:59]
Elise signs off by saying she will be back tomorrow with a fresh idea and thanks listeners for listening[12:01]

Lessons Learned

Actionable insights and wisdom you can apply to your business, career, and personal life.

1

When facing complex, data-limited problems, explicitly modeling uncertainty and considering many plausible scenarios leads to better decisions than anchoring on a single "best guess" model.

Reflection Questions:

  • • Where in your work or life are you relying on a single forecast or plan instead of mapping out a range of possible outcomes?
  • • How could you start quantifying the uncertainty in a key decision you're making, rather than treating unknowns as fixed facts?
  • • What is one important choice this month where you could deliberately generate multiple scenarios and see how your strategy changes under each?
2

Investing in exploration and research upfront, even when it feels expensive, can unlock much larger long-term gains and prevent costly downstream mistakes.

Reflection Questions:

  • • In what area are you underinvesting in learning, testing, or exploration, and potentially paying for it later through rework or missed opportunities?
  • • How might a modest increase in experimentation or data gathering change the confidence you have in a major project or investment?
  • • What is one concrete experiment, pilot, or research effort you could launch this quarter to de-risk an important initiative?
3

Targeting efforts where uncertainty is highest often yields the most valuable information per unit of time or money, accelerating progress toward your goals.

Reflection Questions:

  • • What are the biggest unknowns currently blocking your progress, and how clearly have you defined them?
  • • How could you redesign your next week so that your most focused work directly addresses the questions you're least sure about?
  • • What is one high-uncertainty assumption you could test quickly with a small experiment or conversation?
4

Aligning technological innovation with safety, environmental responsibility, and community resilience creates solutions that are both sustainable and economically robust.

Reflection Questions:

  • • How well do your current projects balance financial goals with impacts on people and the environment around you?
  • • In what ways could integrating safety and sustainability into your planning actually strengthen, rather than weaken, your business or personal objectives?
  • • What is one project where you could explicitly map out benefits and risks for stakeholders beyond yourself and adjust your approach accordingly?
5

Global transitions-such as electrification and digitalization-depend on overlooked upstream systems, so transforming those foundational systems can be a powerful lever for large-scale change.

Reflection Questions:

  • • What foundational processes or "hidden" systems underlie the work you do or the industry you're in that most people take for granted?
  • • How might improving a less visible part of your workflow or organization create outsized benefits for everything built on top of it?
  • • What is one upstream bottleneck or fragile link in your current projects that you could strengthen to support future growth?

Episode Summary - Notes by Phoenix

The emerging science of finding critical metals | Mfikeyi Makayi
0:00 0:00