Trump Rally or Bessent Put? Elon Back at Tesla, Google's Gemini Problem, China's Thorium Discovery
Summary
- The panel rejected a literal “Bessent put,” arguing that the rally reflected fading catastrophe risk rather than a government backstop. Chamath noted that markets were only 500-600 basis points below prior levels after “upending 50 years of economic policy,” while the 145% China tariff looked increasingly like an extreme negotiating anchor. The tradeable shift was that investors no longer believed Trump and Scott Bessent were “crazy enough to tank the global economy.”
- China’s rare-earth dominance turned the tariff fight into a national-security stress test, not merely a dispute over trade balances. Sacks said China processes over 90% of rare earths and makes over 90% of the magnets used in electric motors. Chamath extended the hypothetical leverage to pharmaceutical APIs and battery cathodes in a Taiwan crisis. Sacks’s conclusion: “Security has to be worked out first, then you have trade.”
- The debate over America’s damaged brand came down to trust versus coercive leverage. Sorkin relayed Ken Griffin’s warning that the dollar and Treasuries are unmatched brands now being put at risk, while Chamath relayed that Starbucks, Nike and McDonald’s increasingly market themselves as local rather than American in China. Chamath countered that scaled capital still has “no investable alternative” to the US—and that dependence on Chinese inputs already limits America’s freedom to act.
- Apple’s reported India shift illustrated both supply-chain diversification and how little manufacturing is likely to return directly to America. The report envisioned US-bound iPhone production moving from China to India by the end of 2026; Chamath answered, “See you in 2035,” while others argued Foxconn’s existing Indian lines make rapid scaling plausible. India offers roughly one-fifth of China’s labor cost, itself estimated at one-fifth of America’s, but the panel warned that investing there remains difficult outside infrastructure.
- Alphabet’s quarter showed a resilient, cheaply valued business whose strategic problem is Gemini distribution, not model quality. Revenue reached $90.2 billion, up 12%, net income rose almost 50%, cloud grew 30%, and $70 billion of buybacks plus roughly $10 billion of annual dividends implied a 4-5% shareholder yield at about 18 times free cash flow. Yet ChatGPT is teaching users a new habit while Google leaves Gemini buried despite 270 million paid subscriptions: “Do something so that you can start to blunt the growth of OpenAI.”
- Alphabet’s AI upside is paired with a cloud-compliance risk that could hit GCP, Azure and AWS. Chamath said about 47% of Nvidia’s reported revenue was associated with billing addresses in China, Singapore and Taiwan even though much of the silicon is shipped elsewhere; separately, cloud providers may not be performing enough KYC to identify restricted end users. The panel still expected Google’s planned $75 billion infrastructure spend to hold because hyperscalers view the buildout as “totally inelastic.”
- Tesla’s rebound priced in Elon Musk returning more attention to the company, but neither DOGE savings nor robotaxi readiness was settled. Tesla rose 23% in five days; Sacks said Musk can move DOGE into one- or two-day-a-week maintenance mode, but Congress must turn its claimed $160 billion annualized savings into actual appropriations cuts. On autonomy, one owner described recent FSD as roughly “99.5” quality, while Friedberg preserved the crucial counterexample: direct sun can blind the camera system and trigger an emergency disengagement.
- China’s thorium and fusion programs were framed as compounding industrial advantages built partly on research America pioneered and shelved. Friedberg described a one-million-ton Inner Mongolia thorium reserve said to be sufficient for 60,000 years of Chinese power, a running 2 MW molten-salt reactor, and a planned 10 MW unit by 2030. A newly observed fusion facility was said to be 50% larger than the US National Ignition Facility, making regulatory reform “the number one mission-critical strategic imperative” for America.
Deep dive
1. The rally removed catastrophe risk, not policy risk
Chamath’s answer to the “put” question was no: markets had merely returned to roughly May or August 2024 levels. “If you had said that we would have upended 50 years of economic policy and the markets would only be off five or six hundred basis points, I would have been shocked.”
His sharper question was who wanted a put. Active traders had suffered from the volatility and were not monetizing volatility as they once had; that did not mean the market had fallen far enough for the Federal Reserve or Treasury to rescue anyone.
Sacks argued that commentators cannot blame Trump whenever markets fall, then credit only Bessent when they rebound. He praised Bessent’s reassuring command of bond markets but called the “Bessent put” framing another way to avoid giving the administration collectively any credit.
The competing explanation was a repricing of intent: investors had assigned some probability that 145% tariffs would remain and global trade would stop. Once the administration signaled it wanted deals, the “crazy enough to let it all fall apart” premium came out of the market.
2. Trump’s extreme tariff anchor created leverage that must now be converted
The negotiating logic was an opening bid set “as far away as you possibly can.” A 145% tariff creates room for a deal that would have looked impossible from a 0% starting point, though Sorkin pushed back that visible market sensitivity may now tell counterparties Washington needs agreements too.
Chamath separated equity prices from actual trade flows. Starting from a hypothetical US deficit of $2 billion a day and Chinese surplus of $3-4 billion, he speculated that tariffs might have flipped America toward roughly $1 billion of positive daily cash flow while taking China toward zero or negative territory.
His prescription was political as much as economic: communicate the actual cash-flow change and have every cabinet member repeat it. “We just flipped cash flow. We went from negative cash flow to positive cash flow”—a message Americans might support long enough for negotiations to play out.
3. Rare-earth dependence made the trade conflict a security stress test
Sacks traced the vulnerability to China retaining WTO developing-nation status, which permits tariffs, subsidies and different compliance timelines. He argued Beijing used that framework to identify global choke points and dominate them, rather than participating under genuinely reciprocal free-trade rules.
Rare-earth ore exists globally, but Sacks said China controls over 90% of the difficult processing stage and over 90% of cast rare-earth magnets. China’s cutoff therefore reaches electric motors, automotive manufacturing and many other products, exposing “critical dependencies we’ve created on a nation that is not our ally.”
Chamath extended the mechanism beyond supply restriction: China reportedly told South Korea what it could do with rare earths supplied to it. In a Taiwan crisis, he asked, could America choose freely if China threatened pharmaceutical APIs, battery cathodes and rare earths? “If they can tell you what to think, that is not winning.”
4. Regulatory asymmetry matters as much as tariff arithmetic
Friedberg’s example was Monsanto’s Indian cotton seed, launched around 2005-06. Farmers paid roughly 1,400-1,500 rupees per acre instead of 450 but earned about 5,000 rupees in incremental profit; after adoption reached 90%, he said the government imposed price controls, litigation followed, the IP was taken and Monsanto left.
His broader case was that foreign companies can establish an LLC, bank account, lease and storefront in America with few barriers, while US pharmaceutical, software and hardware companies face price controls, IP risk or local-ownership requirements abroad. In China, he said, companies must set up a 51% joint venture owned by a local company, exposing equity and IP. Regulatory parity could therefore unlock revenue that tariffs alone cannot.
Sorkin preserved the trust objection through Ken Griffin: “No brand compares to the brand of US Treasuries, the strength of the US dollar, the strength and creditworthiness of US Treasuries.” Griffin’s concern was that policy volatility puts that brand at risk even if the underlying objective is defensible.
Chamath’s relative-value rebuttal was that Griffin remains massively long “Pax Americana,” while a Brazilian multibillionaire told him there was still nowhere else to deploy scaled capital. China and India have distinct problems; Europe was described as roughly uninvestable. “The United States is still the shining beacon on a hill.”
5. America’s brand loss may persist even if the deals succeed
Chamath said multinational CEOs see Chinese consumers pulling back from American identity, not merely individual products. Starbucks, Nike and McDonald’s increasingly present themselves as local businesses; a Starbucks job that once impressed a Chinese family no longer carries the same association with the American dream.
Chamath also relayed a Chinese CEO’s memorable distinction: “We’re cool with the US being the leader… It gets a lot more complicated when you have to be the winner.” Chamath replied that this sentiment should be tested against China’s leadership, not only its CEOs.
Sacks recalled a 2003 Hu Jintao speech identifying ten industries that would determine supremacy 25-30 years later and promising national champions in each. Xi, despite ideological differences, followed through; for Sacks, that continuity outweighs softer discussion of branding.
Sacks accepted that the administration must “stick the landing,” but rejected nostalgia for a gentler path from critics who never challenged the prior consensus. His Trumpian mechanism, borrowing Jared Kushner’s phrase, was “controversy elevates message”: disruption first makes unfair trade conspicuous, then negotiations create a new equilibrium.
6. India is the diversification winner, but not an easy investment market
Sorkin cited a report that Apple would move production of all US-bound iPhones from China to India by the end of 2026. Chamath’s “See you in 2035” drew an immediate bet; the rebuttal was that Foxconn and others already operate Indian model lines, making the task replication and scale rather than a cold start.
Sacks framed the destination through security: India sees China as its principal threat, aligning it with the US for decades. Chamath added that India has an educated, young workforce and labor costs estimated at one-fifth of China’s, while China’s are one-fifth of America’s—so this production was never likely to return wholesale to the US.
The investment experience was less clean. Chamath said his decade of Indian technology investing returned perhaps 50 cents on the dollar because American VCs funded familiar replicas while overlooking locally native models, such as food moving from farmers through warehouses and rickshaws. He later pivoted toward infrastructure after the Ambanis and Adanis told him: “Stop beating your head against this wall.”
7. A reverse Kissinger strategy divided realism from moralism
Sacks called China America’s only peer competitor and argued that the US should improve relations with Russia instead of pushing it toward China, much as Nixon and Kissinger engaged Mao when the Soviet Union was the larger Cold War threat.
Sorkin’s pushback was moral and practical: choosing between Russia and China cannot be reduced neatly to rankings when both raise profound objections. Sacks replied that “extreme moralism” is dangerous because foreign policy’s first purpose is American security, not rectifying every injustice.
Their agreement was narrower but meaningful: Iraq, Afghanistan, Libya and Syria did not spread American-style democracy as intended. Sacks argued that democracy works better as soft power—“the shining city on the hill”—than as regime change. He also invoked the open internet’s promise that freedom of information would produce democratic proliferation, a promise he said failed to deliver the expected outcome.
8. Ukraine’s refusal on Crimea became Sacks’s test of realism
Sorkin quoted Trump after Russian strikes on Kyiv: “Vladimir, stop. 5,000 soldiers a week are dying. Let’s get the peace deal done.” He then asked why negotiations remained stuck after Ukraine rejected recognition of Russia’s annexation of Crimea.
Sacks said Crimea should be Kyiv’s easiest concession: Russia annexed it in 2014, Western polling cited by him showed over 80% of residents wanting Russia, and Ukraine’s 2023 counteroffensive failed to sever the land bridge or break the first Russian defensive line. He treated military recovery as impossible.
His proposed settlement was “Istanbul plus”: Ukraine abandons NATO ambitions and accepts Crimea, while Russia keeps additional territory won since the aborted early-war talks. Sorkin raised the Poland domino risk; Sacks saw no evidence Putin wanted a NATO war and said that, if Zelensky rejected Washington’s offer, Europe should become his patron.
9. Alphabet’s earnings broke the simple “search is dead” thesis
Alphabet reported $90.2 billion of revenue, up 12%, with net income almost 50% higher. The panel highlighted roughly $77 billion in services, $12 billion in cloud, $9 billion from YouTube ads, $7.5 billion in non-Google advertising, $10 billion in subscriptions and platforms, and cloud growth of 30%.
Friedberg estimated that $70 billion of buybacks and about $10 billion in annual dividends against a roughly $2 trillion market value create a 4-5% shareholder yield. At about 18 times free cash flow, even losing half of search to ChatGPT could be absorbed if cloud, subscriptions and YouTube continue compounding.
The upside case was that AI reinvents search without destroying the audience. The hidden assets strengthened that asymmetry: Waymo was said to be completing 250,000 rides a week in four cities and worth perhaps $100 billion, while Chamath also remarked on Alphabet’s ownership of part of SpaceX without specifying a percentage. “The market’s giving them zero for that right now.”
10. Gemini’s problem is distribution, taste and the innovator’s dilemma
Sacks said Gemini now performs well on benchmarks, but ChatGPT is growing much faster in actual usage. Users are acquiring a new default habit while Google faces a brutal choice: putting Gemini at the center could give up more than half of its search revenue.
Chamath’s answer was to avoid google.com initially. Give the best Gemini model to the 270 million customers already paying for YouTube or Google One, then make Gmail, YouTube and other under-monetized entry points Gemini’s front door. That could habituate 300-500 million weekly users before Google redesigns search.
His criticism was less about capital or talent than product taste. Gemini’s numerous Gmail and Workspace pop-ups looked like decisions made “at way too junior a level”; Sundar Pichai, Larry Page and Sergey Brin should spend their political capital imposing one coherent experience instead of allowing group decisions to produce clutter.
Chamath’s Facebook analogy was a fight over whether contact importing should be opt-in or opt-out. His instruction then was, “Honestly, just shut up and enjoy the success that will come from keeping this as an opt-out”—evidence, in his telling, that consequential product choices sometimes require one person to impose a decision.
11. AI infrastructure is inelastic, but cloud KYC could disrupt revenue
Clarifying his prior Nvidia comments, Chamath said that about 47% of reported revenue was associated with billing or invoicing addresses in China, Singapore and Taiwan, yet billing addresses do not show where chips ultimately ship. Major US clouds and OEMs invoice suppliers through Singapore while products move to the US, Mexico, Taiwan and elsewhere.
A separate concern was inadequate cloud KYC: restricted users might access Western compute remotely to distill leading models. If governments require GCP, Azure and AWS to identify end users more aggressively, revenue could face “some blowback”; nevertheless, the panel expected Google’s planned $75 billion capex program to continue because AI infrastructure is strategic to hyperscaler survival.
12. Tesla’s recovery still depends on execution beyond Elon’s return
Sorkin linked Tesla’s 8% post-earnings rise and 23% five-day gain to Elon Musk signaling that he would devote more attention to the company. Sacks rejected the claim that Musk was leaving DOGE: he described Musk’s pattern as an intense full-time burst, followed by delegation once he understands the organization “down to the database level.”
As a special government employee, Musk was said to be limited to 130 days a year. Sacks expected one or two DOGE days weekly, with the U.S. Digital Service operating under DOGE, because a complete exit would invite spending to be restored.
DOGE was said to have identified $160 billion in annualized savings, equivalent to $1.6 trillion over ten years if sustained, but Polymarket priced an 84% chance that actual 2025 savings would be below $50 billion. The panel also cited reporting that roughly $92 billion was not itemized, reinforcing Steve Bannon’s demand for “the receipts.”
Sacks and Friedberg agreed that contract cancellations are insufficient without Congress eliminating appropriations. Logrolling rewards legislators for securing more local spending, producing a tragedy of the commons and a $2 trillion deficit. Chamath warned that if the deficit is not addressed, Treasury rates could spike, capital could disappear and the US could enter a “debt death spiral.”
13. FSD impressed owners but did not settle the sensor debate
Friedberg said a recent FSD revision probably moved the experience from roughly “90.8” to “99.5 or something,” with few disengagements; the most frustrating part was still having to watch the road while the driver-monitoring camera enforced attention and imposed temporary suspensions.
Friedberg supplied the hard counterexample: driving directly into sunlight at certain angles can blind Tesla’s cameras and produce an emergency FSD disconnect. That makes a never-disengage robotaxi service materially different from impressive supervised autonomy, especially because Sacks said Elon is never doing lidar.
Waymo’s architecture was presented as the contrast: lidar-equipped cars carry their models onboard, operate without distributed inference points, then return to synchronize. Systems requiring remote human intervention need extremely low-latency connectivity, adding a different infrastructure and reliability constraint.
14. Thorium shows how China converts old research into new leverage
Friedberg described a million-ton thorium reserve at Inner Mongolia’s Bayan Obo complex, said to contain enough fuel to power China for 60,000 years. The US was estimated to have 64,000 tons and Canada 172,000; Friedberg said the US amount alone is enough, in his framing, to power the country for centuries.
China’s 2 MW experimental molten-salt reactor reportedly replaced fuel while operating last June, demonstrating operation without a shutdown and conventional refueling cycle, with a 10 MW unit planned for 2030. Because molten salt is low-pressure and can shut down passively, Friedberg presented it as safer and more modular than a centralized 1,000 MW plant.
Thorium’s fuel case was abundance and utilization: it absorbs a neutron and transmutes into uranium-233, while Friedberg said essentially all mined thorium can become fuel versus less than 1% of mined uranium after enrichment. China’s current forecast already moves electricity production from 3 TW to 8 TW over 15 years without counting this upside.
The bitter historical note was that Oak Ridge pioneered molten-salt work before America shelved it. Friedberg quoted the Chinese program’s leader: “Americans let the research wait for the right successor. We were that successor.” A newly observed Mianyang fusion facility, said to be 50% larger than America’s National Ignition Facility, made his conclusion categorical: US regulatory barriers are now a strategic liability.