Newsletter · 2026-09-21 · Ashutosh Agarwal

Rare Earths and Tariffs Set the Stakes for the Trump Xi Summit - Trade War, Tariffs & Reshoring - Week of September 21, 2026

Trade War, Tariffs & Reshoring for the week of September 21, 2026. Podcast synthesis on the Bessent and He Lifeng talks ahead of the Trump-Xi summit and the November 10 truce deadline, the 12 billion dollar Project Vault minerals reserve and a Department of War tungsten deal, the copper-tariff stockpile distortion, honest reshoring reality checks, the power and transformer crunch, humanoid robots meeting their first real financials, and the measurable U.S. EV retreat.

Trade War, Tariffs & Reshoring

Week of September 21, 2026: Rare Earths and Tariffs Set the Stakes for the Trump Xi Summit

The week in one read

The whole story this week points at one room. On Saturday, September 20, Treasury Secretary Scott Bessent sat down with China's Vice Premier He Lifeng in New York to clear a path for a Trump–Xi summit in Washington later this week, with tariffs, AI, and rare-earth minerals all on the table, and the current trade truce set to expire November 10. Everything else in this issue is downstream of that meeting. China still controls the processing of the minerals America needs to build almost anything advanced, so Washington is racing to buy, mine, and stockpile its way out of that dependence: a $12 billion minerals "vault," a $150 million+ tungsten deal signed by the Department of War, and a copper market so distorted by the mere threat of a tariff that the third-largest pile of copper on earth is now just sitting in U.S. warehouses. Meanwhile the reshoring dream keeps running into the same wall: power. The grid isn't ready, and the people building it say we're "behind in every way." And the two technologies that are supposed to make American factories competitive again (humanoid robots and automation) had their most honest week yet, as real revenue numbers finally landed next to the hype. Below, what the operators and the analysts actually said, and why it matters.

1. The summit that frames the week: rare earths are China's card, tariffs are America's

The single most important event this week isn't a factory or a mine; it's a negotiation. On the September 20 broadcast of NTD Evening News, the setup was laid out plainly: Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng met in New York to "lay the groundwork" for a Trump–Xi summit in Washington later this week. The two sides are negotiating over tariffs, artificial intelligence, and China's supply of rare-earth minerals. Washington also wants Beijing to buy more American farm products and aircraft. The clock matters: the current trade truce expires November 10.

Why the leverage runs both ways is the part worth understanding. Grant Newsham of the Center for Security Policy, interviewed on the same broadcast, put China's chokehold in one number: China controls "up to 98% of the processing capacity" of the critical minerals needed to build AI hardware, smartphones, data centers, and weapons. And he reminded listeners this is not theoretical. Early in the Trump administration, when the White House first got tough, "the Chinese played the rare earth card and they said, not so fast." Within a couple of weeks, Newsham said, "US businesses started coming into the White House saying, if this continues, we're going to have to stop production." That episode, American factories brought to the edge of a halt by an export threat, is the reason critical minerals now dominate this newsletter.

America's counter-leverage is the tariff itself, plus a newer weapon: Trump has just signed a law giving him authority to impose tariffs of up to 100% on major buyers of Russian oil, which includes China. Newsham's read: China's economy has "only one bright spot," its export machine, and "tariffs are something that really frightened them." So the summit is, in his words, each side "sizing each other up to see who makes a mistake." Taiwan hovers over it: Trump is withholding a $14 billion arms sale and has floated it as a bargaining chip, which Newsham warned against reading too much into.

Why it matters: Everything in the reshoring and critical-minerals trade is a bet on how this standoff resolves. If Beijing loosens rare-earth exports in exchange for tariff relief, the urgency (and the government money) behind domestic mining projects cools. If the talks break down before November 10, the squeeze tightens and every Western minerals project with an offtake deal gets more valuable. Newsham is a security analyst, not an operator, so treat the 98% figure as the widely-cited ceiling it is, but the direction is not in dispute.

A second, more grounded voice reinforced the same picture. On American Potential (September 15), Senator James Risch (chairman of the Senate Foreign Relations Committee and the senior Republican on the Intelligence Committee) described the same rare-earth episode from inside the room: China "drew that pistol and laid it on the table" and threatened to ration critical-mineral exports during trade talks, before backing down under U.S. pressure. His framing: critical minerals and semiconductors are "one of, if not the most critical national security issue there is," and the fix has been "an all-of-government approach."

2. Critical minerals stop being a mining story and become industrial policy

If there was one theme the podcasts hammered this week, it's that the U.S. government is now an active buyer, lender, and equity partner in the minerals supply chain, not a bystander. Three concrete deals showed what that looks like in practice.

The $12 billion "vault." The clearest window into the new playbook came from Emerging Tech Horizons (September 16), where Brett Lambert, a former Pentagon official (Deputy Assistant Secretary of Defense for manufacturing under Ash Carter), walked through Vault Co, the private company created out of the White House's "Project Vault." The mechanics are clever and worth understanding because it's a template others will copy:

• It's backstopped by $12 billion: $10 billion in loans from the Export-Import Bank plus $2 billion in private capital.
• It is a demand-driven "stock buffer," not a traditional government stockpile. Manufacturers (from aircraft makers to dishwasher makers) tell Vault Co exactly what minerals they need; Vault buys them on the open market, "U.S. providers first, allies second."
• The minerals sit in storage with the title held by Vault Co. Each year a company can draw down up to 33% of its holding, but must replenish an equal dollar value within a year, and it can swap: pull out $33 of yttrium, put back $33 of tungsten if that's the better value. It's a 10-year program.
• The scope is huge: Lambert referenced 60 critical minerals, and the framing came from the 2026 U.S. Geological Survey report, which found China is the leading supplier of 18 critical mineral commodities.

Lambert's point about why this exists is the tell: he traced it back to when NASA canceled the Space Shuttle and "our solid rocket base was basically eviscerated" because a single supplier of one obscure ingredient held 80–90% of demand. "Primes do not fully grasp the demands... that the subs have at the fifth and sixth tier level." Vault is an attempt to fix that blind spot with taxpayer-backed balance sheets.

The $150 million+ tungsten deal. On Mining Stock Daily (September 14), Trevor Hall reported a US$150–175 million commitment into the Springer Tungsten Complex in Nevada, structured as a $50 million tungsten prepayment facility and backed by the U.S. Department of War. The players: Blue Moon Metals (NASDAQ: BMM), the Elmet/LMET Group (NASDAQ: ELMT), and EQ Resources (ASX: EQR). Production is expected in Q4 2027, with the APT processing plant restarting in the second half of 2028. Tungsten (used in armor, cutting tools, and munitions) is one of the materials China restricts, so a domestic source with a government prepayment is exactly the kind of deal Vault Co is designed to enable.

Germanium and antimony, the quiet chokepoints. Two smaller but telling stories. On Company Interviews (September 17), Group Eleven Resources CEO Sean Eindricks discussed germanium (used in fiber optics, night vision, and chips) trading around $200 per ounce and facing Western supply constraints since China's 2023 export restrictions. And on Company Interviews (September 18), Rua Gold's CEO noted its antimony project is fielding interest from Mercurium, a metals trader with "a huge mandate from the US government... to help with the critical mineral supply," and is "in discussions today with about four different groups" for offtake. The pattern: whoever can produce a non-Chinese pound of these materials now has government-linked buyers lining up.

Why it matters: The investable version of the rare-earth standoff isn't the rare earths themselves; that market is "still somewhat small in terms of total dollars," as one investor noted below. It's the government money flooding into tungsten, germanium, antimony, copper, and lithium, and the offtake deals that de-risk projects that would never have cleared a return hurdle on market prices alone. These are mostly operator interviews (CEOs pitching their own projects), so weigh the promotion, but the government backstops are real and named.

Copper: the tariff that isn't here yet is already reshaping the market

The single best explanation of how a tariff distorts a market, before it even exists, came from Sprott Radio (September 14), where Albert McKenzie, copper analyst at Benchmark Mineral Intelligence, laid out what he called "a tale of two copper markets."

The price move first: the London Metal Exchange (LME) averaged $10,000 a tonne for the first time ever last year, with an all-time high of $11,000. It now sits around $14,000 a tonne: "the copper price has shot up an incredible amount." And McKenzie's honest admission is the striking part: last year the market was in a massive surplus, not a deficit. So why does it feel tight?

Because of an American tariff that hasn't arrived. U.S. buyers price copper on the Chicago exchange (CME), which includes duties, while the LME is a duty-free global price. Because traders expect a roughly 15% U.S. tariff on copper, the CME price has floated $500–600 a tonne above the LME, an arbitrage that, for a while, looked like "free money": buy on one exchange, sell on the other, book the spread. The result: U.S. copper imports more than doubled, and McKenzie's team calculated that there is now more copper sitting idle in American warehouses "than any other country in the world consumes annually after China and the US." In effect, U.S. warehouses have become the third-largest copper "consumer" on earth, buying nothing, just hoarding ahead of a tariff.

The kicker is the setup for volatility. If the administration says clearly no tariff, that demand evaporates and the U.S. draws down its stockpile, which is bearish. If it announces a tariff with a lead time, there's a mad rush to beat the deadline: a spike, then a cliff. Either way, McKenzie warned, the volatility is "good for people who trade it" but "not necessarily good for the day-to-day functioning of what can traditionally be a very orderly market," and the consumer absorbs it in the end ("maybe a microwave is an extra $3").

Against that backdrop, capital is chasing the ounce in the ground. A sampling of the operator interviews this week: Power Metallic Mines (TSXV: PNPN) CEO Terry Lynch, on The KE Report (September 15), touted a maiden resource at his Canadian NISK/Lion project with a mine-construction cost "under $200 million" and a "one-year payback"; on Mining Stock Daily (September 16), the morning briefing flagged Surge Copper's Berg project in British Columbia at a C$4.6 billion after-tax NPV and 24% IRR, and Gunnison Copper's Johnson Camp mine in Arizona adding domestic cathode capacity with a payment from a Rio Tinto venture; and on ITM Trading (September 20), Copper Giant CEO Ian Harris described a 1.1-billion-tonne Colombian resource with a 10-year Trafigura offtake covering 20% of future output. These are pitches, but the throughline is that a decade-long supply deficit narrative plus the tariff distortion has made copper developers financeable again.

Lithium: China's control is quieter but just as deliberate

On The Global Lithium Podcast (September 14), host Joe Lowry (one of the most-followed voices in lithium) argued China's control of the lithium chain is "methodical and well-executed," citing a Ganfeng Lithium executive's point that warehouse inventory data is misleading because much material is unusable for specific battery makers (Ganfeng reported negative usable inventory in September despite high reported stocks). His warning: don't trust the headline inventory numbers that suggest a glut. On the domestic-supply side, Rock Stock Channel (September 16) featured Standard Lithium CEO David Park on pulling lithium from Arkansas oilfield brine (the Smackover): a Franklin project targeting 70,000 tonnes a year and a claimed $5 billion NPV, the kind of "America's next shale" pitch that only pencils if Western buyers pay up to avoid Chinese material.

3. Reshoring reality checks: the honest version

Every week brings triumphant reshoring headlines. The podcasts this week were refreshingly honest about where it's real and where it isn't.

Semiconductors, moving, slowly, and with the government "greasing the skids." Senator Risch, on American Potential, used his home-state Micron as the case study: the U.S. invented the chip, then "the industry chased the siren of cheap wages over to Asia," and now "it was going to take the government to grease the skids to make that happen." His verdict, "the semiconductor industry is moving back," is genuine, but note the framing: a policymaker crediting policy. On the retail-investor side, The Minority Mindset (September 15) framed memory-chip reshoring as a "once in a 100-year opportunity," citing tariffs, Project Vault, and export restrictions, and Tim Cook's warning that "price increases are unavoidable." Useful as a read on sentiment; less so as operator detail.

"China Plus One" is dead; "world plus three" is the new model. The most sophisticated take on the shape of reshoring came from Bain's global chairman Manny Maceda on The Morning Brief (September 14). His argument: the old model of "make everything in China and ship it to the world" created "populists and tariffs," so companies moved to "China Plus One" (add Vietnam, add Southeast Asia). Now even that is being replaced by "the more extreme version... world plus three, or multi-local": building resilient, regional supply chains. His candid warning to boardrooms was about capital: reshoring bets like Tata's move into semiconductors require "patient capital" for a "deep J-curve" 10-year return, and "a public shareholder can sell your stock tomorrow." The companies that win, he said, are the ones with "the courage and the conviction, sometimes contrarian" to make the bet, but that patience is "just challenging to find" in public markets.

What real reshoring actually looks like: 30 jobs, not 3,000. The most instructive story was the small one. On Today in Manufacturing (September 14), the hosts (drawing on an interview with 30-year Northern Tool leader Jonathan Bakke) detailed Northern Tool moving production from Monterrey, Mexico, back to a repurposed 40,000-square-foot space in Faribault, Minnesota, adding about 30 jobs. The drivers weren't primarily tariffs; they were quality problems, long lead times, and unpredictable container and fuel costs. (The company had run a factory in China for nearly 20 years, then near-shored to Mexico, and is now insourcing.) The hosts' honest aside: the small job count "makes me believe automation is also playing a factor here." Their takeaway is the real lesson of the week: reshoring that works often "doesn't have to be transformational, it can just make sense... blocking and tackling."

Solar shows the limit. On Core Memory (September 16), a founder from Swift Solar (which acquired the Arizona and Colorado assets of failed German maker Meyer Burger) described the reshoring frontier: perovskite-silicon tandem cells with a theoretical 45% efficiency (34.8% in the lab today). But the reality check, echoed by an investor in the power section below: solar module assembly has come back, but making the ingots and wafers has not; it's "very energy intensive, rather dirty," and brutally hard to compete on margin. The hardest, dirtiest, most upstream step is always the last to reshore.

4. The power crunch: reshoring's real bottleneck

Here's the uncomfortable truth running under every reshoring and AI story: you can't run new factories or data centers without electricity, and the people building the grid say America is dangerously behind.

"Behind on infrastructure in every way." On Squawk Pod (September 15), Revathi Advaithi (CEO of Flex, soon to lead its spinoff "Axiom," an AI-infrastructure company building power, compute, and cooling from data centers to utilities) did not mince words: "we're behind on infrastructure in every way... particularly around the power sector." Her forecast: by 2035, data centers will use 20% of what's available from the grid. Demand is so far ahead of supply that Axiom is seeing 70% growth this year, 80% next year, is "booked out for the next few years," and is "capacity constrained in every way." The only things slowing deployment, she said, aren't demand or money; it's "labor availability and permitting" and material shortages. (Axiom just announced a $4.4 billion acquisition of a power company, funded with debt and equity.) On the growing political backlash ("people are seeing their utility bills go up"), her answer was an "education process," and a warning that if America slows the buildout, "you don't want to wake up in five years and realize this is what held America behind."

The hidden bottleneck: transformers and the metals inside them. On The Contrarian Capitalist (September 18), investor Dr. Carly Anderson went deep on the unglamorous chokepoints. Transformer lead times are now "two, three, five years." The fix, "solid-state transformers" that use semiconductors instead of "a big chunk of iron with copper wrapped around it," depends on wide-bandgap materials like silicon carbide and gallium nitride, and "absolutely," China has a foothold in both. Her other under-discussed pinch point was aluminum: the U.S. exports about two million tons a year of scrap it can't refine to the grade that electrical cable and aerospace need, because "there really are not aluminum smelters in the West at any scale," and new U.S. tariff policy, given heavy reliance on Canada for primary aluminum, "has not helped." Her investing conclusion is a neat summary of the whole reshoring dilemma: new mines and smelters don't fit "a 10-year fund life," so "my stance is secondary production, I just want the stuff that's already out of the ground."

Nuclear as the only answer for base load. Senator Risch was blunt on American Potential: "you are not going to deliver the load that the tech sector needs with windmills and solar... the only answer to this is nuclear." He also captured the local politics precisely: communities are pushing back because "you're going to build this in my backyard and then I'm going to have to foot the bill for the new electricity," so now, "if you're going to do a data center, you better come with a plan for producing or buying your own electricity and not competing with the neighborhood."

How the industry plans to build that nuclear fast came from The Atomic Show (September 14), where Assil Halimi, CEO of Apollo Atomics, made the manufacturing argument for small reactors. His pitch: the U.S. is "very good at manufacturing," not construction, so the answer is a reactor you build in a factory. Apollo offers 10, 50, and 300 megawatt designs, all truck-transportable, and he claims "there is no factory-manufactured reactor above 80 megawatts that can be transported on a truck." His 300 MW unit is "about half the size of the NuScale power module and delivers four times the power," with a target of being "competitive with natural gas." A striking detail for cost: "soft costs" (engineering, documentation, permitting) are 30–40% of a new plant's cost, and Apollo is building AI tools to attack them.

Why it matters: Power is the constraint that turns reshoring from a slogan into a schedule. Every AI data center, every new fab, every automated factory competes for the same scarce megawatts, transformers, and skilled electricians. The investable edges the podcasts pointed to: transformer and grid-connection equipment, wide-bandgap semiconductors, secondary/recycled metals, and factory-built nuclear. The constraint isn't going away this decade.

5. Humanoids meet their financials: the hype gets a spreadsheet

For two years, humanoid robots have been sold as the answer to America's factory-labor gap. This week, for the first time, real numbers landed next to the vision, and they were sobering.

Agility Robotics: $1.8 million in revenue, $100 million+ burned. The most valuable episode was The Robot Report Podcast (September 18), which paired hard financials with a real product launch. Agility unveiled Digit 5, now with human-oriented knees (reversed from the old backward "ostrich" bend), 284 pounds, 5-foot-11, reaching over 7 feet, a 90-minute run time on a 9-minute charge (20+ hours a day), and CE certification so it can actually be deployed in European warehouses (customers had refused to deploy uncertified Digit 4 units). The company now has over 400 employees. But the money: Agility made just $1.8 million in revenue in 2025, burned $100 million-plus that year, and won't start early Digit 5 rollouts until early 2027, with full production by the end of 2027. The bullish counterweight: over $300 million in back orders specifically for Digit 5. One of Jonathan Hurst's most contrarian claims cut against the whole field: "general purpose hands don't exist... there are hands that look like a human hand, but they're not very good." Agility is sticking with two- and three-finger grippers because, per a prosthetics competition he cited, a two-finger gripper beat the fancy hands.

1X wants to ship 50,000 robots next year. On Relentless (September 18), 1X founder Bernt Børnich laid out an audacious plan: 50,000 humanoid units next year across two factories: Hayward (10,000-unit capacity) and a new San Carlos plant (100,000-unit capacity). His engineering argument for why humanoids can scale faster than skeptics think: a well-designed humanoid is "about 1,000 parts" and "less than 70 pounds," versus a car's "50,000 parts" and "4,000 pounds." The company (11 years old) makes its own motors and tendon drives in-house because "there is no supply chain for this," and can now go "from major changes in CAD until a new robot walks off the line" in about four weeks. His honest caveat: the gating item isn't shipping 50,000 units, "it's ensuring you don't get them back"; safety and reliability, not ambition, are the real constraints. And on the perennial debate, he sided with generalists over specialists but conceded that today "if you want to fold laundry... a specialized model will work better," with the shift to general intelligence coming in "2027, not 2026."

The number that keeps humanoids in perspective. Both episodes made the same quiet point: nobody is selling many of these yet. As the Robot Report hosts put it, "no company building humanoids for industrial environments has sold many of these." The revenue is a rounding error; the back orders and the factory capacity are the bet.

The China angle, and a fresh trade wall. Buried in the same Robot Report episode was a genuinely new tariff/reshoring story. The founder of RoboInk (which distributes and customizes Chinese Unitree robots) said the FCC has moved to restrict new advanced foreign quadruped and humanoid robots, aimed squarely at Chinese imports, even as Unitree filed for an IPO that could value it around $900 million. His response captures the whole reshoring impulse in one line: the rule is "something to really just push companies like ours to get your shit together in the United States and start globally sourcing everything," so American buyers aren't "totally reliant on this Chinese technology." His fear for the consumer: a China-made robot "in your house now getting all this data on you." RoboInk is pivoting from reselling Chinese hardware to localizing purpose-built robots for healthcare, security, education, and small business (a concierge robot-as-a-service at $2,500 a month). This is the humanoid version of the same decoupling playing out in minerals and chips.

Automation is quietly working where humanoids aren't. The nearer-term money is in boring, task-specific automation. On Inside the ICE House (September 16), GXO CEO Patrick Kelleher said the logistics giant will have humanoids working in warehouses within two years with ROI in that timeframe, already piloting across 45 locations, while tackling a 25% warehouse labor turnover rate across 26 countries by using automation to fill "unfillable jobs" rather than displace workers. And on The New Warehouse (September 14), Carhartt's VP of operations described ripping out an obsolete crane-based storage system and deploying Exotech robots across five distribution centers, justified by total-cost-of-ownership, not novelty. Meanwhile, on Business Wars (September 17), WSJ's Christopher Mims highlighted Apptronik (Austin-based, partnered with Google DeepMind) demonstrating a robot folding laundry in messy, real-world settings, with a CEO focused on "the cheapest, best humanoid" and leaving the AI to partners. (For corroboration, Figure's widely-covered Helix 2.5 breakthrough (its Index program paid $15 million to 4,000 daily contributors for 16 million videos from 108 countries, against a $2 billion annual compute budget) appeared across roughly a dozen AI podcasts this week, e.g. Personal AI, September 18.)

Why it matters: The humanoid trade is a 2027+ story, and this week's numbers ($1.8M of revenue, $100M+ of burn) are a reminder to underwrite it as venture risk, not near-term earnings. The nearer, cleaner exposure is task-specific automation with real ROI today (warehouse robotics, ASRS replacement), and the FCC's move against Chinese robots is a new, investable decoupling vector to watch alongside chips and minerals.

6. Autos: the EV retreat is now measurable

The clearest cost-benefit accounting of a policy reversal came from Reuters Econ World (September 16), "Trump's EV U-Turn." The numbers, from Reuters' auto team drawing on Atlas Public Policy data:

After the $7,500 EV tax credit ended in September 2025, EV sales "fell off a cliff," down 40%+ in the months that followed (EVs had climbed from under 2% of U.S. sales to about 10% by end-2025). The administration also froze fuel-economy fines and stripped California's ability to set its own standards. Notably, 87% of the canceled EV investment was in red states: the "battery belt" of Georgia, the Carolinas, Ohio, Kentucky, Tennessee, Indiana, and Michigan.

The paradox: this has been good for Detroit's near-term profits. Automakers "lose money on EVs" (only Tesla and BYD make them pay), so cutting EV capex and leaning back into big trucks and SUVs sent GM's profits to records and freed cash for buybacks. Total auto employment (~1 million at automakers and suppliers) is roughly flat, down about 1% since January 2025, despite the administration's jobs claims, though tariffs are pushing some production from Mexico back to the U.S., which could help later. One genuine silver lining: some idled EV battery plants are being repurposed for grid-scale energy storage, riding the same AI-data-center demand from Section 4.

The long-term worry the reporter raised is the one PMs should hold onto: the U.S. has become "this hermetically sealed market... protected from Chinese competition." China will export 7–8 million cars this year (from almost none six years ago), and its entry into Canada is "a dress rehearsal" for an eventual U.S. push. With no competitive pressure and no regulatory forcing function, the risk is that protected U.S. automakers under-invest and get caught flat-footed: "Jim Farley at Ford is on record saying we've seen this movie before" (the Japanese in the 1970s).

Across the Atlantic, the restructuring is brutal. On Today in Manufacturing (September 14), the hosts detailed VW CEO Oliver Blume's plan to invest "three-figure billions" (i.e., $100 billion+) while cutting 50,000 jobs, slashing the model lineup by 50% (to ~75 models across 12 brands) and cutting complexity 75% by 2035. The damning stat: "VW has about 60% more employees than Toyota, despite producing the same number of vehicles." The hosts' contrast with Northern Tool was sharp: VW is "reactive to market conditions," while the small Minnesota reshorer was "proactive." Both are the same story from opposite ends: the cost of a manufacturing footprint built for a world that no longer exists.

Operators vs. pundits: whose word to weight this week

• Operators / insiders with real skin in the game (weight these most): Bernt Børnich (1X, on his own factory ramp), Jonathan Hurst / Agility (via Robot Report, with audited-style revenue and burn figures), Revathi Advaithi (Flex/Axiom, on booked-out power demand), Assil Halimi (Apollo Atomics), Patrick Kelleher (GXO, warehouse robotics ROI), Brett Lambert (Vault Co, on the actual mechanism of the minerals reserve), and the Carhartt and Northern Tool operators on why they actually moved production. When these people give numbers, they're describing their own P&L.
• Grounded analysts / reporters (high signal, one step removed): Albert McKenzie (Benchmark) on the copper-tariff arbitrage, the single most useful market explanation of the week; Joe Lowry on lithium inventory games; the Reuters auto team on the EV retreat; Manny Maceda (Bain) on the shape of multi-local supply chains.
• Policy voices (directionally useful, but they're selling the policy): Senator Risch on minerals, semis, and nuclear; the NTD security panel on the summit.
• Handle with skepticism: the mining-explorer CEOs pitching their own projects (Power Metallic, Copper Giant, Surge, Element 29, Standard Lithium): the government-backstop and offtake themes are real, but every project is "the next big mine"; and retail-investor framings like The Minority Mindset's "once in 100 years" memory-chip pitch and ITM Trading's "$200 billion Pentagon sovereign wealth fund" claim, which we could not corroborate and would not lean on.

The week ahead: what to watch

• The Trump–Xi summit in Washington and the November 10 truce deadline. The whole minerals/tariff complex hinges on whether rare-earth access is traded for tariff relief. A thaw is bearish for domestic-minerals urgency; a breakdown is bullish for every Western project with an offtake.
• Any clarity on the U.S. copper tariff. Per Benchmark, a clear "no tariff" or an immediate tariff both trigger a U.S. stockpile drawdown (bearish); a tariff with lead time triggers another import rush (a spike, then a cliff). This is a live volatility catalyst.
• The FCC's restriction on Chinese robots. A new decoupling front. Watch how it's scoped, whether it names Unitree, and whether it pulls forward "made-in-USA" robotics demand, and Unitree's IPO progress (~$900M mooted valuation).
• Vault Co's first named participants. Lambert said they're courting Fortune 100 and top defense/tech firms. The first announced offtakes will show which minerals (and which producers) get the government backstop first.
• Whether reshoring shows up in jobs. Auto employment is flat despite the rhetoric; watch the next payroll prints and any Mexico-to-U.S. production shifts actually breaking ground.