reSee.it - Related Video Feed

To transition the world to sustainable energy, we'd need 100 gigafactories. The Gigafactory will be huge, but Tesla alone can't build 100. Big companies worldwide must follow suit. Government support and a carbon tax are crucial for a quick transition.

The transcript argues that China’s export restrictions on indium compounds will “almost certainly” crash AI data center infrastructure build-out plans due to indium’s role in high-speed optical networking. The speaker, Mike Adams, describes indium as a rare, not-very-abundant element and connects it to periodic-table groupings: indium, boron, aluminum, gallium are grouped together, and indium’s outer-shell electron configuration is described as relevant to forming compounds. The key application claim is that indium is needed for materials that combine electrical conductivity with optical transparency. The speaker emphasizes indium tin oxide and other indium-based transparent conductors, and links this to optical microchips and telecommunications used for AI data centers. The transcript states that copper-based transmission is inadequate for the throughput required to connect large numbers of GPUs quickly enough for large model training. It asserts that very fast inter-GPU communication requires optoelectronics, including optical transceivers, switches, and optical modulators capable of terabits-per-second bandwidth. It claims there is “no substitute” in photonics for indium phosphide and that indium phosphide performs for lasers, photodetectors, modulators, and optical telecom functions. According to the transcript, China has already placed indium phosphide on an export control list in early 2025, which is said to have caused a price spike: indium phosphide wafers are reported to have risen about 250% in roughly a year and a half to around $5,000 per six-inch wafer. The transcript further claims that China has increased scrutiny on buyers of “straight indium,” requiring end-user information and destination country details for European and U.S. purchasers. It describes a reciprocal geopolitical pattern: the U.S. is said to have pressured ASML to block exports of high-end UV lithography equipment to China, while China responds by restricting exports of gallium, indium, and indium phosphide. The transcript claims gallium is used for night vision optics and radar systems and that China mines/refines about 70% of exported indium. The transcript identifies supply-chain bottlenecks: it states that 70% of the global indium market is controlled by China and that substrate manufacturing is largely handled by AXT Sumitomo, described as controlling about 80% of substrate production. It frames this as a “choke point” that would affect AI data center rollout, including “orbital data centers,” because high-speed optical transmission would still be required. The speaker cites an article from Mining.com as saying indium phosphide is a “powerful trade weapon” and quotes Semi Analysis’ Conrad Wong describing indium phosphide as one of several supply chain bottlenecks “collectively gating AI data center build outs.” It also mentions NVIDIA’s $2 billion investment into U.S. photonics product makers Coherent and Lumentum, and Lumentum/Marvell’s acquisition of Celestial AI for photonics work, as evidence that AI builders recognize dependencies on photonics and indium. The transcript expands to other element constraints, mentioning gallium and tungsten hexafluoride (WF6) as inputs for microchip manufacturing. It explains that indium is extracted as a byproduct from zinc mining and then refined from zinc ores, stating there are no dedicated indium mines and no large U.S. mining or sufficient reclamation to replace Chinese supply. It claims indium recycling exists but is not enough for the industry’s needs. Finally, the transcript asserts that China’s leverage can “flick a switch” to block exports and describes prior reversals when U.S. trade pressure is applied, with China cited as using these restrictions as negotiation leverage. It concludes by stating the U.S. AI industry is dependent on Chinese supply and warns that the AI data center “bubble” could face a brick wall due to these element bottlenecks.

It takes a massive amount of diesel to create concrete, steel, and transport materials using heavy machinery. The carbon footprint of these operations, along with solar panels and lithium batteries, may not be offset during their lifespan. The existing transmission lines are inadequate to power the world with electricity. We have a 120-year petroleum-based infrastructure that is essential to our lives and found in roads, car wheels, tennis rackets, lipstick, refrigerators, antihistamines, plastic products, cell phones, clothing, soap, and more. We will run out of petroleum before we find a replacement, which will kill us as a species. Oil extraction is dangerous, but we do it because we run out of options. The demand to keep pumping oil is to blame for the danger.

Speaker 0 notes that the energy solutions list for energy-hungry data centers was short and contained one thing: gas. They ask why not gas and renewables. Speaker 1 responds: "the what one has to appreciate is the intensity of energy." As an engineer, they state: "the mix of energy doesn't matter. How much is wind? How much solar? We like to advertise that. Kilohounces matter because energy intensity has to shift, not the mix." They argue that solar power cannot produce cement or steel and that "they are very energy intensive." Therefore, "you still need a gas based heating or" (implying gas is necessary). They add: "Physics. It's against physics. Fine. Absolutely. Physics don't allow do it." They emphasize evaluating energy mix changes in the context of "jewels of energy," noting the world still needs to progress and must build infrastructure—steel, cement, fuels. The challenge is how to change the energy mix while also building data centers and consuming more energy. They describe the current problem as "single threaded with the gas fired power plant, maybe a little bit of nuclear. Nuclear? Renewable remain in the mix, cannot bring the amount of jewels we need to produce this infrastructure which is required in the world."

The conversation centers on rare earths and critical minerals as key “leverage” in the ongoing war dynamic between the US and China, described as a modern equivalent to oil. Mario argues that Trump’s decision to get into the war—framed as a “big gamble”—could be tied to gaining leverage over China, which holds leverage through rare earths. He links this leverage to global energy and an “energy choke point,” and asks Lippy/Anupam to explain how big the issue is for the West and US security, when China gained the edge, whether it is reversible, and the national security risk. Anupam says oil once dominated geopolitics, but rare earth and other minerals that power modern economies are “the new oil.” He asserts that militarily, technologically, and for AI and supercomputing, nothing like the described way of life can be made without rare earths. He adds that anything powered by electricity and much consumer tech uses rare earths and critical materials. He claims that 90% or more than 90% of rare earth production is controlled by one country (China), and that attempted tariffs against China were not pursued because stopping Chinese rare earth shipments could shut down major production lines quickly. He cites an example where one large motor-company production line stopped within about six weeks after rare earths stopped shipping, and that defense primes would not be able to produce defense systems if disruption continued. The discussion distinguishes “lights” versus “heavies.” Anupam states EVs use light rare earth magnets, while defense equipment uses “heavy” rare earths that are temperature sensitive, and he claims drones and modern warfare rely on rare earths. Mario reinforces that everything becomes a switch for the US defense sector. They then discuss how rare earths became outsourced and why China gained dominance. Anupam says America offshored production to make goods faster and cheaper over decades, not necessarily maliciously, and that China developed an entire processing and supply chain over about 30 years. He says China got technology and know-how from earlier US processing instruction in the 1990s, and later grew into a competitor controlling critical materials needed for manufacturing. A key point is described as bipartisan and international: Anupam says the European Union policy proposes no country producing more than 60% of these critical materials, while today 90% or more is produced in China. He says the US is increasing government support through floor pricing, debt financing, and equity investments for critical materials companies. He describes deglobalization as accelerating beyond COVID-era trends, but says for rare earths it is an even bigger threat. A major operational deadline is raised: Lippy states the Pentagon cannot buy systems containing Chinese-origin rare earth materials after January 1st, 2027, “in less than six months.” Anupam says this is a law taking effect then, not an executive decree, and it creates a defense-specific requirement: anything sold to defense cannot have a “Chinese nexus.” He contrasts this with EV companies, which he says do not have that issue in the same way. They describe shifting restrictions and enforcement. Anupam says that three days before the conversation, the US banned certain rare earth material companies (including MP Materials and USA Rare Earths) from procuring Chinese equipment and chemicals, because most US processing depends on Chinese equipment, and without that equipment and know-how the US cannot process rare earth materials. He says regulations keep changing weekly, and that an economy cannot function if the ability to trade elements changes on a weekly basis. The group discusses company-level implications. Anupam says their focus is “heavy rare earths,” especially dysprosium and terbium used in defense. He claims their company is the only non-Chinese nexus outside China and argues that most Western rare earth companies still have Chinese connections through equipment, chemicals, control panels, and other parts. He describes an example involving a research organization and the inability to buy Chinese equipment after China stopped selling “to non-friends,” which forced rebuilding from scratch. He asserts that they were positioned by timing and location, and that their lack of Chinese nexus is a strategic advantage. They also cover permitting and geography. Anupam says all rare earths (specifically heavies) have uranium and thorium, and that processing creates radioactive byproducts. He claims another company attempted to build processing in the United States (Texas) but pulled the project because it could not get an EPA permit. He says Saskatchewan is suited because uranium-handling infrastructure and permitting exist, citing “Uranium City,” and asserts certain radioactive-related processing steps cannot be done elsewhere in the US but can be handled in Saskatchewan. They describe building facilities and scaling quickly. Anupam says they acquired a heavy rare earth mine in Canada within 12 months, acquired 80% offtake from an SRC facility shown behind him, set up heavy rare earth metallization, acquired PMT Critical Metals in Ohio, announced financing, and planned pilot magnet manufacturing. He says they formed an agreement with JOGMEC for magnet-related expertise and knowledge transfer. They state the US Army selected them to build facilities on an army base (Utah). Finally, they return to industrial base and replenishment. Mario argues the US has depleted munitions and has low domestic manufacturing share compared with WWII (15% to 20% now versus ~60+% then), and asks how the US will restock and rebuild capacity for hypersonic and drone-heavy conflicts, tying the industrial base challenge back into the rare earth supply chain. Anupam says scaling manufacturing is a 10–15 year journey because it took 40–45 years to give away capacity, and that the approach should focus on faster infrastructure for refining, metallization, and magnet manufacturing. He says their speed is part of the solution: they moved rapidly from being “on paper” to acquiring mines, off-take, metallization facilities, and magnet lines, and that similar acceleration is needed across sectors. Anupam adds technical points: he says their hydrofluoric-acid-free process reduces exposure and capex/opex, and they use automation and AI/robotics to reduce labor intensity, noting a plant scale comparison where “China” needed 60 people while their approach uses two. The conversation concludes that while the work is to support defense now, scaling to allied and broader supply is part of the longer roadmap.

Going all electric by 2035 is not practical because there is no such thing as a zero emission vehicle. Electric cars simply shift emissions elsewhere. Manufacturing a single 1,000 pound battery requires digging up 500,000 pounds of materials and 100 to 300 barrels of oil. This process can result in a carbon debt of 10 to 40 tons of CO2. Increasing battery usage will require more minerals like lithium, cobalt, and zinc, leading to a 400% to 4000% increase in demand. However, there isn't enough mining in the world to produce enough batteries for everyone's cars.

Electric vehicles are driving a surge in demand for minerals like lithium, nickel, rare Earth elements, and copper. By 2030, global lithium production needs to increase 8 times to meet Tesla's needs. These cars require 6 times more minerals than conventional vehicles. The mining industry generates $119 billion annually, with a projected 105% increase in nickel demand for transportation by 2026. By 2040, rare Earth element demand will rise by 1,000%. Additionally, copper production must increase significantly as wind turbines require 4.7 tons of copper each.

Copper and aluminum are the primary beneficiaries of the grid spending increase. $800,000,000,000 is going to buy copper, which is money. How big is the oil market compared to the metals market? Crude oil dominates. All metals—iron ore, gold, copper, aluminum, nickel—are thinly traded and critical. There is no chance to get off crude oil; you can’t build electric cars, windmills, solar, or a modern military without these metals. Underwater power cables are expensive, and offshore wind with transmission to Greening efforts illustrates copper’s central role. Copper is the focus: copper is the expected $270,000,000,000 per year market by tomorrow morning. Where will this metal come from? There is no copper inventory. Historically, since Mohenjo Daro, humanity mined 700,000,000 metric tons of copper; about 80% of all copper ever mined is still in human possession. Recycling can recover about 80% of that 700,000,000 tons, but to do so would require tearing down every building in the United States, Europe, Japan, and China. Copper is embedded in buildings and other infrastructure; it can be recycled, but extracting it at scale remains challenging. Currently, we consume 30,000,000 tons of copper a year, with only 4,000,000 tons recycled. To maintain global 3% GDP growth, without electrification and relying on burning oil and gas, we must mine the same amount of copper in the next eighteen years as we mined in the last ten thousand years. In the next eighteen years, we would have to mine the same cumulative amount as in ten thousand years prior, without electrification, without data centers, without solar and wind, and without the greening of the world economy. There is little appreciation for the challenge faced. Since 1900, the energy required to produce copper has increased 16-fold. As ore grades decline, more energy is needed to produce the same metal, while water consumption has doubled. The easy copper deposits are largely depleted; Chile accounts for 24% of global copper mine production, but costs are in the third or fourth quartile. Chile burns coal, and solar isn’t reliable for mining operations since the sun shines only ~five hours a day; solar is useless without grid-scale storage. We are heading for a train wreck in Chile. To meet copper demand, six giant Tier One mines must come online every year from now until 2050. To meet copper demand, 40% of production must come from new mines for electrification, data centers, and grid upgrades. All the talk about AI is fantasy without sufficient energy. Nuclear power could help, but its components require metals, and the U.S. lacks the capability to weld containment vessels in traditional nuclear plants; Korea can build a nuclear power plant.

Solar panel waste is highly toxic and requires special disposal. However, due to the high cost involved, discarded panels are being sent to landfills in poor countries instead. Research shows that by 2030, there will be around 8 million tons of green waste, which is expected to increase to 80 million tons by 2050.

Dan and Kelly discuss the outlook for copper and what’s baked into current prices. Kelly notes that short-term factors influencing copper include disruptions in mines in parts of the world, tariffs, and uncertainty about Federal Reserve policy. In the longer term, she says, prices reflect a growing need for copper as the world electrifies. They estimate that by 2040 the world will use 50% more electricity than today, which she equates to “building 650 nuclear power plants every year.” Copper is described as the “metal of electrification.” She explains that much of the demand growth will come from developing countries, and that with the rise of data centers and AI there is a voracious appetite for electricity that has surprised traditional utilities. She cites that data centers used about 4% of US electricity last year, and by 2030 it will be more like 14%, and none of that happens without copper. Dan recalls that copper was first discussed as a major story in 2022, noting that while prices have risen since, they haven’t surged like major tech equities. He acknowledges that commodities are highly cyclical and asks how investors can ensure continued upside given potential soft data points or supply coming online. Kelly responds by emphasizing copper’s link to GDP, describing it as a core economic demand vector. She notes that a key factor is government policy toward mining exploration: it takes an average of 17 years to bring a new copper mine online, so investing in copper is a bet on the future and depends on how governments regulate mining exploration. Overall, the conversation highlights the thesis that long-term copper demand will be driven by electrification and rising electricity use (especially from data centers and AI), while near-term price dynamics will be influenced by mine disruptions, tariffs, and macropolicy. The lag between discovering, permitting, and developing new copper mines (about 17 years) adds to the structural bullish case.

Speaker 0 argues that we are still completely underestimating how short we will be in terms of the global demand-supply dynamics of a handful of critical elements. In the view of the Trump doctrine, the world is no longer as multilateral, and there is a need for unilateral national security. From this lens, the asset set to go absolutely parabolic is copper. Copper is described as the most useful, cheap, amenable, conducted material that we have, and it manifests in everything from data centers to chips to weapon systems. Currently, Jason, we are on a path by 2040 where we will be short about 70% of the global supply at current course and speed. Copper.

The speaker, a long-time green energy supporter, was dismayed to learn about the environmental and human costs associated with green technologies. A single lithium mine allegedly creates millions of tons of waste annually, laced with sulfuric acid and radioactive uranium, polluting water for 300 years. Child labor is used to mine cobalt. Solar panels are allegedly made by laborers in razor wire enclosed camps exposed to quartz dust, causing silicosis. The Ethical Consumer Organization reports that forced labor in the solar panel supply chain is hard to avoid. Wind turbines consume vast resources, require diesel to start, gallons of oil to lubricate, and are hard to recycle. Solar panels are also extremely difficult to recycle, costing more than production. Lithium batteries pose steep challenges too. The speaker claims these "green" solutions are actually good marketing from the $1.5 trillion climate change industry. They urge people to prevent further escalation through unnecessary EVs and solar farms consuming farmland.

The discussion highlights lithium mining in Chile, part of the "lithium triangle," where nearly a third of the world's lithium is produced. One plant can power about 50,000 electric vehicles annually, with potential to reach 75,000. The Salar de Atacama boasts the best lithium brines globally, containing approximately 2,000 parts per million of lithium, making it the most cost-effective production location. Lithium is extracted through brine mining, where salty water is evaporated in ponds, concentrating the lithium. Despite high demand and limitless resources, Chile is losing market share to Australia and Argentina. Experts emphasize the need for Chile to quickly increase production before other countries surpass them or new battery technology emerges. Chile's president has announced a state-led plan for lithium industry development. Separately, it is mentioned that Piedmont Lithium has bought homes in North Carolina for a project.

Speaker 0 mentions a lack of coordination, but it is unclear what they are referring to. Speaker 1 questions the wisdom of becoming more dependent on and vulnerable to a perceived enemy. They express concerns about the enemy's actions in Latin America, America, and with currency, suggesting they are trying to take down America. Speaker 0 then brings up the supply chain of critical metals for electric vehicles and defense. Speaker 1 acknowledges the information about the need for a 2,000% increase in mining for 20 years to meet the demand for EVs and critical metals.

Oil, natural gas, and coal still dominate as the main sources of global energy, providing 84% of the world's energy. Despite claims of a rapid transition away from fossil fuels, the reality is that we have made little progress in shifting to green energy. The main challenge lies in the need for a significant increase in mining to obtain the necessary materials for solar panels, wind turbines, batteries, and other components. This mining process requires a substantial amount of energy, further contributing to the challenge. Additionally, the location of new mines is a concern, as China currently holds a monopoly on critical energy materials. Attempts to build mines in the United States and elsewhere face strong opposition. Future energy demands will only increase with population growth and technological advancements, making it clear that a diverse mix of energy sources, including fossil fuels, nuclear energy, and renewables, will be necessary.

President Biden initially stated that he wanted 50% of new cars to be electric by 2030, but it has now been updated to 60%. It is true that electric cars require six times the mineral inputs compared to conventional cars. However, if 50% of cars were electric today, the current electric grid would not have enough power to charge them all. Achieving EV targets globally by 2030 would only reduce global temperatures by 0.0002 degrees Fahrenheit by 2100. Despite this, unilaterally impacting the U.S. auto market, critical mineral supply chain, and grid stability is not seen as the solution for addressing temperature goals.

Speaker 0 introduces a myth that Trump waging war against Iran would close the Strait in a way that hurts China first, making Trump victorious, and asks for an answer to that perception. Speaker 1 argues that the perception isn't accurate, noting China has been building energy security for over twenty years. They travel to China frequently and see zero signs of energy scarcity; if there were any potential energy squeeze, it would be visible among the people and on social media, but it isn’t. He explains China’s energy composition is stable, and that even if Middle Eastern energy supplies were disrupted, China’s situation remains manageable. He states that China actually produces 30% of the crude oil it consumes domestically, so it does not import all its energy. Speaker 0 adds that people are often surprised by how much solar, wind, and hydropower China has, mentioning a special report noting that the aggregate annual terawatt-hours of output of China’s power grid is more than double the United States, and that this is growing rapidly. Speaker 1 confirms the rapid growth and attributes part of China’s diversification to the influence of Western financial practices, saying, “thanks to the Western banking cartel because they have been suppressing the price of silver to ridiculous low prices.” He claims China imports all the silver to manufacture solar panels, implying that by maintaining low silver prices, Western bankers have inadvertently helped China with energy diversification.

I've been exploring lithium mining, which is crucial for the energy transition in America, especially for AI technologies that require significant electricity. The U.S. power grid struggles to support this demand, leading to the installation of large lithium-ion battery facilities. Indigenous groups have fought against lithium mining at Thacker Pass due to its toxic nature, but the Biden administration allowed it to proceed. Interestingly, I discovered a plan to convert the Hoover Dam into a giant battery, similar to how ancient pyramids were believed to generate electricity. There's a connection between Tesla, Trump, and the push for a new power grid, raising concerns about how this will transform our land and energy systems into something resembling a computer chip. The implications of this transition keep me awake at night.

Mining uses toxic chemicals and creates hazardous waste, yet is needed for green technologies. Demand for minerals is expected to increase 400-600%. Years ago, a proposal for Pebble Mine in Alaska was vetoed by the EPA due to environmental concerns, despite scientific studies. A Republican administration removed the EPA veto, but President Biden vetoed it again. Environmental groups and regulators have allegedly killed new mines in America, with permitting taking decades. The Biden administration dealt a blow to Twin Metals mine plans. Environmental groups oppose American mines, but clean energy needs minerals. Windmills, solar panels, and batteries require a massive increase in minerals. The NRDC didn't provide examples of mines they support. The Green Movement has been happy outsourcing mining to disadvantaged countries with child labor. America has child labor laws, safer equipment, and environmental rules. America once led in mineral production, but now depends on other countries. Society can't exist without mines.

The IT industry relies on minerals like lithium and cobalt, and their extraction consumes massive amounts of water, causing pollution. As ore quality decreases and demand increases, extraction practices become more aggressive. The global demand for lithium is projected to rise 40 times by 2040. Disruptions like floods and droughts are forcing mining plants and factories to shut down. Big tech data centers, often located in drought-stricken regions due to incentives, are increasing pressure on water levels, leading to conflict with farmers and local communities. Big tech is competing for water with agriculture, which accounts for 70% of human water usage. The relentless push for AI adoption will multiply water consumption and energy demand, despite AI not being sustainable. AI-assisted searches consume up to five times more energy than conventional searches. Those pushing for AI adoption are often those who have invested heavily in it.

Jensen Huang (NVIDIA) discusses how the amount of compute—and the energy required for that compute—is likely to increase dramatically, moving from “a hundred times” to “a thousand times” compared with current levels. He frames future computing as two simultaneous shifts: it will be intelligent and contextually aware with generative outputs, and it will be continuous rather than based on prerecorded retrieval that is initiated only when prompted. The discussion contrasts concerns about today’s AI being “backward looking” and copying previous work, potentially leading to feedback loops where people rely on AI and become stagnant without new regenerative creativity. Jensen Huang’s described future addresses this by arguing that software will not remain static code stored on a hard drive; instead, people will ask AI to write software in real time as needed (for example, generating a Photoshop clone to edit an image or generating an original movie tailored to a preference). Creating such continuous generative experiences is said to require a tremendous amount of energy—“a thousand times more” than today’s levels. Speakers note that existing energy sources cannot easily support this scale. The conversation states that it cannot be done on hydrocarbons, not even on nuclear due to long build-out time, and not on solar because current energy sources are insufficient. It also emphasizes efficiency: having the ability to use vastly more energy does not mean it should be used, and continuous regeneration is not always the more efficient approach. Speaker 0 then argues for limiting market cap and having these groups invest themselves without government backing or government liability protection, suggesting a free-market approach rather than government-directed competition framed as an arms race. Speaker 2 responds that pursuit of “superintelligence” requires centralized power and therefore cannot be decentralized. The conversation claims this centralized effort is being directed toward a quest for superintelligence connected to world domination and competition, particularly framed as an attempt to “beat China,” and concludes that once superintelligence is achieved, humanity’s fate would be in question.

Because the plan is to cover the whole planet with this to produce enough power for these data centers. I don't think this is really a one for one swap on the positive side for humanity to cover our entire planet with this to to divert power when there's so many other ways to do it, you know? We can't get clean coal technologies. Only pure spring water slash artesian water slash deep well water punching into aquifers will work. So the call is once they get the electrification route from Eritrea, Ethiopia down through Tanzania, you're gonna watch a bunch of AI data centers pop up along there and they're gonna tap all those sandstone aquifers beneath to get that water. No data center left behind.

To make a wind turbine, you need a large amount of iron ore, concrete, and steel. The concrete production emits carbon dioxide, and the steel requires rare earth elements, which are often sourced from China and come with environmental concerns. Additionally, the cobalt used in wind turbines is often mined by child slaves in dangerous conditions in the Congo. The turbine blades are made from balsa wood obtained by clearing parts of the Amazon forest, and they contain a toxic chemical called Bisphenol A. These blades cannot be recycled and end up as landfill, polluting the soil and water. Supporting wind and solar power means supporting pollution, slavery, and environmental damage.

The speaker argues that China’s export restrictions on indium compounds will likely disrupt the AI data center infrastructure build-out. They describe themselves as an AI developer and “elemental scientist,” running a mass spec laboratory for elemental analysis, and emphasize indium’s rarity and lack of natural abundance. They connect indium to periodic table groupings: indium is in the same periodic table column as boron, aluminum, and gallium (group 13), and like other group 13 elements it has three outer electrons. They state that combining a group 13 element with a group 15 element such as phosphorus produces compounds with characteristics “like silicon” but with better light transmission. They assert that for systems requiring optical transparency and conductivity—solar panels, optoelectronics, optical telecommunications, touch screens, solar cells, and electrodes embedded in displays—indium enables transparent conductors such as indium tin oxide. The speaker links indium to high-speed data center networking, claiming copper cannot provide the required throughput for massive GPU clusters (they mention setups like 100,000 GPUs). They say extremely fast GPU-to-GPU interconnections require optoelectronics and optical transmission rather than copper wiring, noting that they personally use copper at 10G but that it is “getting really slow,” while large AI builders (SpaceX, OpenAI, Meta, Google) rely heavily on optical infrastructure. They claim data centers thus “depend severely on indium.” They then describe escalation in export controls: China is restricting indium exports (and “scrutinizing” exports of straight indium). They say that even in 2025 China added indium phosphide to an export control list. They explain that indium phosphide is indium and phosphorus configured together. They state that indium phosphide wafer prices increased by about 250% in a little over a year and a half, reaching about $5,000 per 6-inch wafer, and they portray this as the “template” for optoelectronics fabrication. The speaker further claims that China is asking extra questions of buyers of just indium, including European and U.S. purchasers providing end user information and destination country details. They connect this to prior U.S. pressure on ASML to block high-end UV lithography exports to China and say China is countering by blocking gallium exports and indium/indium phosphide exports. They argue this will “dramatically hamper” U.S. AI data center build-out, stating that silicon does not work at required wavelengths while indium phosphide works for lasers, photodetectors, modulators, and optical telecom equipment for terabits-per-second bandwidth. They claim “there is no substitute in photonics” for indium phosphide and state that without indium there is no high-speed optical networking in data centers. They present supply chain choke points: they say China controls about 70% of the global indium market and also point to AXT Sumitomo as handling about 80% of substrate manufacturing, while non-Chinese buyers depend on China-controlled input. They reference a Mining.com story stating China’s control over indium phosphide exports threatens AI data center rollout and quotes Semi Analysis analyst Conrad Wong on indium phosphide as a supply chain bottleneck gating AI data center build-outs. They mention NVIDIA’s $2 billion investment into U.S. photonic product makers Coherent and Lumentum and Marvell acquiring Celestial AI, claiming these moves reflect an industry need for photonics dependent on indium. The speaker expands to related shortages and production constraints, mentioning gallium and tungsten hexafluoride (WF6) as bottlenecks for microchips and optoelectronics. They explain indium comes as a byproduct from zinc mining rather than from dedicated indium mines, stating there are “no dedicated indium mines” and that indium is extracted from zinc ores using solvent extraction and electro-refining. They claim China mines/refines around 70% of indium supplied globally, followed by South Korea, Japan, Canada, and others, and state none is the United States. They assert that while indium recycling exists (especially reclaiming indium tin oxide from displays in Japan), there are “almost no spare reserves,” and they say there is no U.S. mining or large-scale U.S. reclamation sufficient for AI data centers. They conclude that if China “flick[s] a switch” to block exports, the U.S. AI industry could be stopped quickly due to dependence on Chinese supply, and they argue that without indium there is no quick substitute. They add element trivia, stating indium is named from “indigo” due to its bright indigo blue spectral line and the Latin indicum, and they mention other elements as named after places or scientists. They end by urging caution toward AI company hype, warning that AI data center expansion could hit “a brick wall called no indium,” tied to ongoing export restrictions and supply bottlenecks.