Trump Rally or Bessent Put? Elon Back at Tesla, Google's Gemini Problem, China's Thorium Discovery

Evidence since April 2025 shows that the rare‑earths cutoff was directly addressed in U.S.–China trade talks and partially rolled back through formal agreements.

1. China did halt rare‑earth exports after the April 2025 tariffs. On April 4, 2025 China imposed export controls on seven categories of medium and heavy rare earths in response to U.S. tariffs, and reporting in early April noted that shipments of those rare earths effectively stopped, with exporters waiting indefinitely for licenses.(energynews.oedigital.com) This is the “cutoff” Sacks was referring to.

2. A formal trade agreement explicitly tied to rare earths was reached in June 2025. On June 11, 2025, Reuters reported that President Trump announced a new U.S.–China trade agreement under which China would supply the U.S. with magnets and rare‑earth minerals, while the U.S. kept 55% tariffs; the deal, framed as a trade truce, was described as lifting China’s export restrictions on rare earths, pending final approval by Trump and Xi.(reuters.com) A contemporaneous Gulf News analysis summarized the same London deal as committing China to provide “full magnets and any necessary rare earths” to the U.S., resuming exports after roughly two months of severe restrictions, via export permits for U.S. firms (valid for six months and renewable).(gulfnews.com) That is a formal, negotiated trade arrangement that explicitly addresses the rare‑earth cutoff.

3. That deal restored at least civilian U.S. access to Chinese rare earths. Industry and policy analyses describe this London agreement as a rare‑earths “deal” or truce that restarted flows of Chinese rare‑earth magnets to U.S. manufacturers, even while emphasizing it is temporary, license‑based, and leaves China with major leverage.(rareearthexchanges.com) These accounts stress that military‑grade magnets and some specialized rare‑earth products remain outside the fast‑track system, so access is partial and focused on non‑military uses rather than comprehensive.

4. Later negotiations reinforced (but did not fully liberalize) this framework. A subsequent deal after Trump–Xi talks in Busan on October 30, 2025 suspended a planned expansion of Chinese rare‑earth export controls for a year as part of a broader trade agreement, effectively preserving the existing flow regime rather than tightening it further.(en.wikipedia.org) U.S. and Chinese officials have since referred to a continuing, repeatedly extended truce under which lower tariffs and ongoing Chinese rare‑earth exports to the U.S. are maintained, while both sides argue over draft regulations that could re‑tighten controls.(uk.marketscreener.com)

5. Assessment relative to the prediction. Sacks’ normalized prediction was that the rare‑earth cutoff would be “explicitly resolved in a formal agreement, restoring US access to rare earth materials or processing from China.” That did occur in substance: rare‑earth exports were shut off in April, then explicitly addressed and partially rolled back in a negotiated London trade deal that committed China to resume supplying magnets and rare earths to the U.S., followed by later agreements that kept that flow going (though under restrictive licensing and with military‑grade items still constrained). The resolution is imperfect and fragile, but a formal agreement did explicitly restore significant U.S. access to Chinese rare‑earth materials. On that basis, the prediction is best judged as right (directionally correct, albeit with important caveats about scope and durability).

The prediction concerns whether Apple will fail to move assembly of all iPhones destined for the US from China to India by the end of 2026 (an 18‑month timeline from the April 2025 reporting).

Multiple reports in late April 2025 (e.g., Financial Times summaries via MacRumors, Business Standard, The Guardian) state that Apple’s goal is to assemble all US‑bound iPhones in India by the end of 2026, but they describe this as a future target, not an already-achieved shift. (macrumors.com) As of mid‑2025, Apple is rapidly ramping up Indian production and exports to the US, but is still only part‑way there, with China remaining a major manufacturing base. (reuters.com)

However, the evaluation date here is November 30, 2025, while the prediction’s deadline is end of 2026. There are still about 13 months remaining before that deadline, so we cannot yet know whether Apple will ultimately meet or miss the target. Therefore, the prediction’s truth value cannot be determined yet, making it too early to judge.

The prediction is explicitly about "the next couple of decades"—i.e., roughly 20 years of stable, closely aligned U.S.–India security and trade relations. As of today (30 November 2025), only about seven months have elapsed since the podcast release (26 April 2025), which is far too short to evaluate a 20‑year forecast.

Available evidence so far actually shows deepening U.S.–India alignment in security and trade: a new 10‑year defense framework agreement has been signed and described by U.S. officials as indicating that military ties have "never been stronger"; joint leaders’ statements from early 2025 reaffirm a comprehensive global strategic partnership and launch new defense, trade, and technology initiatives; and recent commentary describes the relationship as entering a phase of “structured stabilisation,” with growing defense sales and efforts to resolve trade frictions. (whitehouse.gov) None of this, however, can confirm or falsify the 20‑year stability claim yet.

Because the prediction concerns a much longer time horizon than has passed, the correct rating is inconclusive (too early to tell).

The prediction concerns the entire 21st century, asserting that no other country will surpass China as the primary security threat for both India and the United States before 2100. As of today (November 30, 2025), fewer than 75 years remain in that horizon, and we are only about a quarter of the way through the century. There is no way to definitively determine now whether some other power (for example, a resurgent Russia, a future unified bloc, or a currently minor power) might become the top security concern for either country later in the century. Because the claim is explicitly about a long-term future state extending to 2100, it is too early to judge its accuracy. Therefore the correct classification is: inconclusive (too early to tell).

The prediction explicitly concerns a 20–30 year period starting around 2025, claiming India will have a prolonged high‑growth phase comparable in impact to China’s 2003–2012 boom and become a major global manufacturing engine. As of November 2025, less than one year of that window has elapsed, so the long‑run claim cannot yet be verified or falsified.

Available data show that India is currently the fastest‑growing major economy, with recent and projected real GDP growth around 6.5–8% per year, according to the IMF’s Article IV reports and World Economic Outlook updates.(imf.org) India has also been expanding manufacturing and exports, supported by government schemes like production‑linked incentives and rising FDI, and is now roughly the fifth‑largest manufacturer globally.(ibef.org) However, its share of global manufacturing is still only about 2.8–3.2%, versus roughly 29% for China, far from the dominance China achieved.(ndtvprofit.com)

India does enjoy a substantial labor‑cost advantage over China—manufacturing labor is on the order of $1.4/hour versus about $7/hour in China, and roughly a small fraction of U.S. manufacturing wages—supporting one premise of Chamath’s argument about cost competitiveness.(refteck.com) But whether this advantage will translate into a China‑style, multi‑decade manufacturing‑led boom comparable in impact to China’s 2003–2012 surge can only be assessed many years from now. Given the very early stage relative to the stated 20–30 year horizon, the prediction is too early to call.

Evidence from the quarters immediately after the April 26, 2025 podcast shows that hyperscaler data‑center and AI infrastructure capex was maintained or increased, not meaningfully reduced, despite tariff uncertainty.

1. Alphabet/Google (the specific ~$75B reference)

   • On Feb. 4, 2025 Alphabet said it expected to invest about $75B in 2025 capex, mostly for servers and data centers.(cnbc.com)
   • In its Q1 2025 earnings commentary (late April, i.e., right around the podcast), Alphabet reiterated that it still expected to invest approximately $75B in 2025 capex, noting this could fluctuate by quarter due to delivery and construction timing but not cutting the total.(alphabet2025ir.q4web.com)
   • On April 9, 2025, amid market volatility from President Trump’s tariffs, Alphabet publicly reaffirmed its ~$75B 2025 spending plan “despite turmoil over U.S. tariffs,” explicitly tying this to data center and AI build‑out.(investing.com)
   • By Q3 2025 (within two quarters of the podcast), Alphabet had raised its 2025 capex guidance first to $85B and then to $91–93B, with the “vast majority” going to technical infrastructure (servers, data centers, networking).(datacenterdynamics.com)
     This is the opposite of a cut: Google’s planned ~$75B was maintained and then increased.

2. Meta

   • In its late‑2024 guidance, Meta expected 2025 capex of $60–65B.(marketscreener.com)
   • In Q1 2025 (May 2, 2025—the very next quarter after the podcast), Meta raised that 2025 capex outlook to $64–72B, citing additional data‑center investments to support AI and higher infrastructure‑hardware costs.(datacenterdynamics.com)
   • Meta explicitly attributed some of the increase to uncertainty and higher costs from Trump’s tariffs and trade discussions, not to cutting investment.(datacenterdynamics.com)
     So tariffs raised Meta’s costs, but its DC/AI capex plans grew rather than being reduced.

3. Microsoft and Amazon

   • Microsoft signaled plans to spend about $80B in fiscal 2025 on AI‑enabled data‑center capex, and Q3 FY25 results (reported April 30, 2025) showed capital expenditures rising sharply, with management saying capex would continue to grow in the next fiscal year even after Trump’s tariffs were announced.(reuters.com)
   • While Microsoft later said it was “slowing or pausing” some specific data‑center projects (e.g., an Ohio site), it still planned to invest over $80B in AI infrastructure globally, and analysts continued to model roughly that spend, describing project moves as portfolio adjustments rather than a broad capex pullback.(apnews.com)
   • For Amazon, contemporaneous coverage of Q1 2025 results noted that it left its 2025 capex outlook roughly unchanged around ~$105B, most of it AI/cloud‑related, even as trade tensions rose.(mitrade.com)
   • Subsequent announcements of large new AWS government and regional data‑center investments later in 2025 are consistent with an ongoing, elevated build‑out rather than a near‑term cut.(reuters.com)

4. Industry‑wide picture under tariffs

   • Analysis of the top hyperscalers ahead of and during 2025 shows them collectively targeting around $350B+ in 2025 cloud/AI datacenter capex, up dramatically from prior years, with commentary that tariffs and trade war risks might eventually force moderation but had not yet done so.(forbes.com)
   • Another industry summary puts Q1 2025 total datacenter capex up ~50% year‑over‑year, with the top four cloud providers’ capex up ~73%, noting that hyperscalers were “investing proactively in anticipation of tariffs” rather than pulling back.(linkedin.com)

Across Google, Meta, Microsoft, and Amazon—the core “hyperscalers” Sacks was talking about—capital‑expenditure plans and data‑center buildouts in the one to two quarters after April 26, 2025 were maintained or increased, even as tariffs raised costs and created uncertainty. There is no evidence of a meaningful reduction in these strategic investments over that time window; if anything, the trend was further upward.

Given that, Sacks’s prediction that hyperscaler data‑center/AI capex (including Google’s ~$75B plan) would “hold up” over the next quarter or two despite tariff turmoil is best evaluated as right.

Sacks predicted that (1) the DOGE program in the US federal government would continue operating beyond 2025 and (2) Elon Musk would dial back his involvement to about 1–2 days per week instead of being there full time.

The second part largely happened: following Musk’s own comments on a Tesla earnings call, contemporary coverage reported that he planned to scale back his DOGE time to a day or two per week while refocusing on Tesla. (businessinsider.nl) Subsequent reporting and the DOGE overview note that he then left his formal White House/DOGE role entirely by late May 2025. (en.wikipedia.org)

However, multiple later reports show that DOGE itself did not continue past 2025. In late November 2025, major outlets (Guardian, Time, SF Chronicle, New York Post, People) all reported that the Department of Government Efficiency (DOGE) was dissolved or dismantled months ahead of its scheduled July 4, 2026 end date, with its functions largely absorbed by the Office of Personnel Management and other entities. (theguardian.com) These articles consistently describe DOGE as having been disbanded as a centralized program in November 2025, not continuing into 2026.

Because a core element of the prediction was that the DOGE program would keep operating beyond 2025, and it in fact ended in November 2025, the overall prediction must be judged as wrong, even though Musk’s short‑term shift to a 1–2‑days‑per‑week mode did occur before he exited the role entirely.

As of November 30, 2025, the United States has not experienced a singular, acute debt or fiscal crisis of the kind typically described as a break-point event (e.g., default, failed Treasury auctions, sudden loss of market access). There are ongoing concerns about high federal debt and deficits, but nothing that is widely characterized as a distinct crisis that clearly forced Congress into a new structural regime on spending/appropriations.

Likewise, there is no evidence that Congress has, in the wake of such a crisis, broadly recognized and praised Elon Musk’s deficit-cutting efforts via Doge (Dogecoin) or any Doge-linked fiscal mechanism; no such mechanism has been adopted in U.S. fiscal policy, and there has been no wave of congressional credit to Musk along those lines.

However, Sacks’s prediction is explicitly framed as eventual (no time bound) and depends on future political and economic developments that may still occur after November 30, 2025. Because:

1. The hypothesized crisis has not occurred in a clearly recognizable form yet, and
2. There is no specified deadline by which it must occur,

we cannot say the prediction is wrong; it simply has not had enough time to be tested. Thus the correct status, as of now, is “inconclusive (too early)”.

As of November 30, 2025, Tesla has launched a limited “Robotaxi” ride‑hailing service in Austin using a small fleet of Model Y vehicles, but all rides still involve a Tesla employee acting as a safety monitor in the front seat, and similar pilot services or tests elsewhere also continue to require human safety personnel. These services are geofenced, closely supervised, and not yet fully driverless commercial robotaxis available broadly to the public. (en.wikipedia.org)

The normalized prediction says Tesla will be operating real robotaxi services—vehicles driving themselves without active human drivers, offered commercially to the public—within roughly two years of the April 26, 2025 conversation, i.e., by around April 2027. Since it is still late 2025 and that deadline has not yet arrived, there is not enough information to determine whether the prediction will ultimately be correct or incorrect. Therefore, its status is inconclusive (too early to tell).