If Donald Trump actually fires Jerome Powell as Federal Reserve Chair, US equity markets will fall by more than the roughly 1% intraday drop seen on the rumor; i.e., a significantly larger decline will occur immediately following a confirmed firing.
“I do think the market would go down quite a bit more if Trump did fire Powell.”
View on YouTube
Explanation
Trump never fired Jerome Powell, whose term as Fed Chair ran its course, so the conditional premise of this prediction was never tested.
If US federal deficits continue at roughly current levels and existing federal debt is refinanced at around the current long‑term Treasury yield (~5%), then within a few years (on the order of the mid‑to‑late 2020s), annual net interest outlays of the US federal government will exceed annual spending on at least one of Medicare/Medicaid, Social Security, or the military, and eventually become significantly larger than those individual line items.
“you can kind of run a couple of scenarios, but like pick your pick your metric. If the deficit kind of continues at current levels and we were to refinance the debt at the prices that, you know, David was talking about, you know, it's only a few years before spending on interest is significantly larger than spending on Medicare and Medicaid or Social Security or the military.”
View on YouTube
Explanation
US net interest expense has already grown to exceed defense spending and is closing in on other major entitlement line items, consistent with the predicted trajectory.
On the current fiscal trajectory (unchanged policies on spending and taxes), US federal net interest expense will become the single largest budget line item within roughly the next decade, surpassing each of Medicare/Medicaid, Social Security, and defense individually.
“our current course and speed, it's it's in the not too distant future where interest expense is the biggest line item for the government”
View on YouTube
Explanation
Net interest expense had grown substantially but had not yet surpassed Social Security as the single largest federal budget line item as of 2026.
The next major frontier language models from xAI (successor to Grok‑4 or Grok‑5), OpenAI (successor to GPT‑4.1), and Google (next Gemini generation) will be trained on Nvidia’s Blackwell‑generation GPUs rather than Hopper, and their capabilities will show a large step‑function improvement over current frontier models.
“This model was trained on Hopper. So this this model is probably about as far as you can take the last generation of Nvidia GPUs. The next models we see grok 505 from OpenAI. Whatever they're going to call Jim the next Gemini, they will be trained on Blackwell. And I think that will be a really big step function.”
View on YouTube
Explanation
Subsequent frontier models from xAI, OpenAI, and Google were trained on Blackwell-generation Nvidia hardware and delivered substantial capability improvements through 2025-2026.
Assuming Adams or Cuomo do not withdraw from the race, Zahran Mamdani will win his targeted New York office (the race they are discussing), and if he fully implements his stated socialist policies, those policies will fail badly within a few years and will rapidly discredit and curtail the broader socialist political movement in the U.S. (a "quick end to socialism").
“unless Adams or Cuomo pulls out, he's almost certainly going to win. And then if he wins, you know, does he go through with these policies? Because if he goes through with them, they're going to be epic failures... So they won't work. And I think that'll be a quick that'll be like a quick end to socialism.”
View on YouTube
Explanation
Zohran Mamdani won the November 2025 NYC mayoral election with over 50% of the vote against Cuomo and Sliwa, confirming the electoral premise; whether his policies will prove to be 'epic failures' remains to be seen since his term only began January 2026.
U.S. tariff revenues in calendar year 2025 will likely fall short of the higher targets discussed (e.g., $250B+), ending closer to the low end of the $200–500B range, but by 2026 it will be plausible and reasonable for annual U.S. tariff revenues to fall in the $200–500B range on a full-year basis.
“it's going to be hard to get to that number just because the tariffs only really kicked in middle of the year... 30 billion. That's $360 billion annualized, right? Bang in the middle of the estimate. But you didn't have a lot for the first six months. Yeah, maybe you hit the low end of that with 30 or 40 billion for the last six months. But um, but we'll see. But for 2026, that does seem reasonable.”
View on YouTube
Explanation
2025 US tariff revenue came in substantial but below the very highest projections given the mid-year tariff start, while 2026 revenue trended toward the predicted $200-500 billion range as tariffs applied for a full year, roughly matching the prediction's shape.
Despite tightened U.S. export controls, smuggling of high-end Nvidia GPUs into China will continue at a meaningful level that is difficult to fully halt, but the total volume (on the order of about $1B over a few months as of 2025) will remain small relative to U.S. hyperscaler deployments and will not be decisive in determining the overall outcome of the U.S.–China AI capability race.
“I think it's going to be hard to stop it... China is doing everything they can to bring these chips in... I don't think it's all that dispositive.”
View on YouTube
Explanation
Smuggling and circumvention of US chip export controls into China continued through 2025-2026 despite enforcement efforts, remaining a persistent but not war-deciding factor in the overall US-China AI compute competition.
Within roughly 3–4 years of August 2025 (by 2028–2029), Apple will have brought to market the leading consumer augmented reality glasses product, superior to competitors’ offerings.
“it would shock me if they do not have the best augmented reality glasses on the market in 3 or 4 years. It's so plays to their strengths.”
View on YouTube
Explanation
Apple continued developing AR/smart glasses products through 2025-2026 without yet having released a market-leading consumer AR glasses product as of mid-2026, so this 3-4 year prediction (to 2028-2029) remains too early to call.
Over roughly the next 50 years from the time of this recording (by ~2074), solar power will become the dominant global energy source, with a large share of the world’s energy usage running on solar; however, solar-generated electricity will still not be cheaper than nuclear power, but its cost will approach that of coal.
“I think it is highly likely that in my lifetime the world just runs on solar. Like if you just, you know, we all know compound interest is the greatest force on earth. But if you just look at the rate at which photovoltaic cell efficiency is compounding, battery efficiency is compounding. And people will make these balance of system arguments, but it will it will never be as cheap as nuclear, but it will likely approach coal. And I think a lot of the world will run on solar. But that's going to take 50 years.”
View on YouTube
Explanation
Solar remains one of the fastest-growing electricity sources globally but a roughly 50-year horizon for global dominance is too far out to evaluate.
At some point in the future (no specific date given), Bitcoin will grow to become a serious competitive threat to the US dollar’s role, likely as a store of value and/or medium of exchange, prompting notable policy reactions from future U.S. administrations.
“But I do think Bitcoin will at some point be a serious threat to the US dollar. And that just is what it is. And we will see how different administrations react.”
View on YouTube
Explanation
Bitcoin has not threatened the US dollar's role as the global reserve currency; the dollar remains dominant in global trade and reserves as of 2026, with Bitcoin's role shifting more toward an institutional/strategic-reserve asset than a currency competitor.
After the next generation of leading AI hardware is released—specifically Nvidia’s Blackwell GPUs, new AI-focused chips from AMD, and new ASICs from Broadcom, expected starting in the year following this recording (2025)—China will no longer be able to keep up with the United States at the leading edge of AI compute capability.
“but you know, Nvidia's Blackwell chip comes out next year. You're going to have new chips from AMD, new Asics from Broadcom. And I, I think at that point it is not going to be possible for them to keep up anymore.”
View on YouTube
Explanation
The US maintained a lead in cutting-edge AI compute (Blackwell and successor chips) through 2025-2026, but China made significant strides via domestic chipmakers, narrowing rather than eliminating the gap.
By the end of 2025, Vladimir Putin will be in a significantly weaker geopolitical position: Xi Jinping will have visibly begun to distance China from Russia, and a Trump administration will have imposed or negotiated terms that are widely perceived as very unfavorable to Russia in the context of the Ukraine war.
“I think Putin I think Putin is going to lose bigly... I just think she is going to begin decoupling from Putin... I think Trump is going to be a lot tougher on Putin than people think. And I think he's going to get a very, very, uh, deal that's very, very bad for Russia and Ukraine.”
View on YouTube
In 2025, demand for AI inference compute (GPUs/accelerators) will again exceed available supply, producing a notable industry‑wide shortage similar to the GPU shortages experienced in 2023.
“we are the other like inference compute is going to be the the kind of derivative winner of that. We're going to run out of GPUs, accelerators compute in 2025 the same way we did in 23.”
View on YouTube
In 2025, companies deriving more than 35% of their revenue from U.S. government services/contracting will underperform comparable peers financially or in valuation, making them relative business losers.
“Government service providers. You do not you do not want to have the United States government at any level is over 35% of your revenue.”
View on YouTube
In 2025: (1) overall corporate M&A activity will surge after several slow years, with a marked increase in deal volume and size; (2) Intel will undergo at least one major strategic transaction (e.g., large acquisition, divestiture, spin‑off, or transformative partnership); and (3) multiple independent frontier AI labs will be acquired by larger technology or cloud companies.
“I think the biggest business deal is that they're going to be deals. I think you're going to see a tidal wave of M&A after four years of not being able to get anything done... something will happen with Intel and that will be big... I do think you will see a lot of these frontier AI labs that are independent be acquired.”
View on YouTube
AI will be the defining political issue of the 2026 midterms and the next US presidential election.
“I obviously AI is I think going to be the defining political issue of the midterms and for sure the the next presidential election.”
View on YouTube
Explanation
AI has become a significant political topic including regulation, jobs, and safety debates, but whether it is 'the defining' issue of the 2026 midterms (versus economy, immigration, or healthcare) remains a subjective and not yet fully settled characterization.
DRAM will account for 30-40% of all hyperscaler capital expenditure next year (2027).
“Something I've been thinking about memory is DRAM is probably going to be 30 to 40% of all hyperscaler capex next year.”
View on YouTube
Explanation
This is a forecast about 2027 hyperscaler capital expenditure allocation to DRAM that cannot yet be verified against actual full-year spending data.
China's CXMT will flood the global market with cheap consumer-grade DRAM after going public, easing the price pressure driving Apple's cost increases.
“CXMT is going public in China. They are going to they may be the cure for Apple's ills. They will flood the market with to some degree cheap consumer grade DRAM”
View on YouTube
Explanation
CXMT did complete a large IPO in July 2026, raising about $8.6 billion and becoming China's most valuable listed company, positioning it as a major new source of DRAM supply, consistent with easing memory price pressure, though 'curing Apple's ills' is a debatable overstatement.
Once Starship is fully reusable, launching a gigawatt of compute into orbit will cost about $5 billion, and within three to four years the cost of terrestrial gigawatt data centers will rise to roughly $70 billion versus roughly $40 billion for orbital compute, making orbital compute economically favorable.
“when Starship is reusable, it's going to cost $5 billion to put a gigawatt of compute into space... it may be that in three or four years it's 70 billion verse 40 billion”
View on YouTube
Explanation
Starship has not yet achieved full operational reusability at the scale needed to test these specific orbital-versus-terrestrial compute cost comparisons.
Anthropic is currently worth approximately $3 trillion on a private-market basis.
“I think Anthropic is worth $3 trillion today and it's very important.”
View on YouTube
Explanation
Anthropic's valuation as of mid-to-late 2026 was reported around $965 billion following its Series H round, with IPO targets discussed around $2 trillion, but it had not reached a $3 trillion valuation at the time this claim was made.
Anthropic will end the year (2026) with an annualized revenue run rate well over $100 billion.
“They're going to end this year well over 100 billion.”
View on YouTube
Explanation
Anthropic's revenue run rate reached roughly $65 billion by late July 2026, and full-year 2026 actual revenue was estimated around $20-26 billion, both well short of the $100 billion annualized run-rate level predicted for year-end.
Within roughly a year, SpaceX/Elon Musk will begin selling orbital or Colossus compute capacity to Google (and potentially OpenAI), extending the Elon Web Services model beyond its current Anthropic deal.
“Well, I think it's much easier to see him selling compute to Google and I think already there have already been posts about that.”
View on YouTube
Explanation
There is not enough specific public confirmation of a formal SpaceX-Google compute sales arrangement to verify this claim.
SpaceX's Starship will achieve rapid reusability (flying and landing the same booster/ship multiple times per day) within about one to two years, possibly sooner.
“I don't know when. I don't want to make a prediction. I would guess a year or two, maybe sooner.”
View on YouTube
Explanation
As of mid-2026, SpaceX had not yet achieved rapid same-day multi-flight Starship reusability; the upper-stage tower catch milestone was still pending, so this window has not been definitively resolved.
Space-based (orbital) data centers will become commercially real sometime between the second half of 2028 and the first half of 2030.
“But I think second half of 28 to first half of 2030 would be my point prediction.”
View on YouTube
Explanation
This prediction concerns commercial orbital data centers arriving between late 2028 and early 2030, a timeframe not yet reached.
There will not be a near-term U.S. dollar crisis, given America's relative advantages in energy self-sufficiency, AI leadership, and market strength versus other major economies.
“I don't think a dollar crisis is around the world.”
View on YouTube
Explanation
No US dollar crisis has occurred; the dollar has remained the dominant global reserve currency.
Within approximately 5–10 years from April 2023 (i.e., by April 2028–April 2033), Starship-based point-to-point transport will eliminate essentially all conventional transatlantic and transpacific air cargo flights, causing those long-haul air cargo routes to disappear.
“There will be no more transpacific or transatlantic cargo flights. I think in five, six, seven, eight, ten years. You're going to need a big Starship fleet to accomplish that. But I think the transatlantic and Trans-Pacific aerospace cargo routes go away.”
View on YouTube
Explanation
As of mid-2026, Starship has not eliminated any transatlantic or transpacific air cargo routes; conventional air cargo remains the dominant mode and Starship point-to-point cargo transport has not been commercially deployed.
Anthropic's IPO will be priced around a $2 trillion valuation and will trade up to roughly $3 trillion afterward.
“I think that's probably where it could where where it might trade... Maybe it is two and it trades to three.”
View on YouTube
Explanation
As of August 2026, Anthropic has not yet IPO'd; reporting says investors are targeting an October 2026 listing at $2 trillion or more. The IPO pricing and post-IPO trading price are not yet known.
Anthropic will exit 2027 with an annualized revenue run rate of roughly $400-500 billion (assuming it ends 2026 around $100-120 billion).
“I think four to 500 billion is within the zone... of like realistic and achievable... I think the overunder really is 500 billion exit ARR for next year. And I think they think they'll do the over.”
View on YouTube
Explanation
A 2027 exit-ARR forecast; 2027 has not happened yet, so this cannot be assessed.
xAI's Grok 4.7, a larger and significantly more capable model than Grok 4.6, will launch within a few weeks of this episode (late July/August 2026).
“this is still a relatively small model. It's 1.5 trillion parameter models. A larger model, Grock 4.7 that should be significantly more capable is coming out, you know, in a few short weeks.”
View on YouTube
Explanation
As of August 23, 2026 (well past 'a few short weeks' from the July 2026 episode), Grok 4.7 had not been released; xAI reportedly delayed it to early September after adding supplemental SpaceX training data.
Within the lifetimes of the people in the room, average human lifespan in developed countries will extend well past 100 (potentially toward 125-150), driven in part by GLP-1/obesity drugs acting as de facto longevity drugs.
“I would take the over on that uh in no small part because we already have one of the best longevity drugs out there and folks don't even realize it in the glip ones and the obesity drugs.”
View on YouTube
Explanation
This is a very long-run lifespan prediction that cannot be evaluated on a near-term basis.
LEO satellite constellations will replace all ground-based RTK networks and traditional GPS within the next decade.
“for me the geodet piece uh I I just think the space thing is is likely the path. It's going to replace all RTK and all GPS in the next decade.”
View on YouTube
Explanation
This prediction concerns satellite positioning fully replacing RTK/GPS within a decade, a timeframe not yet elapsed.