September 9, 2025
Summary
Market pricing for 3 rate cuts in 2025 and a fed funds rate around 3% by mid-2027 is “pretty fair,” but a sustained plunge to 1%-2% is highly unlikely. The speaker sees a September cut as the base case, with cuts also possible in October and December; a 50bp move is not the baseline. Regardless of how hard Trump pressures the Fed or whom he replaces as chair, the central bank will ultimately take a rational view of markets and rate cuts.
The 10-year Treasury yield is near the end of its downside move around 4.05%, making further bond gains an “asymmetric risk.” Fiscal expansion is creating both inflation concerns and ongoing Treasury issuance; the speaker’s reasonable range is 4.2%-4.5%. Investors should not chase TLT or long-duration bonds. If yields return to around 4.2%, a swaption on higher yields may offer a better trade than directly shorting the note. He stressed that 4.2%-4.5% is only his personal market read and may not be highly accurate.
The two major themes for the next several years are fiscal expansion and exponential growth in AI compute demand, both of which will make it difficult for rates to return to the lows of the old cycle. This is not 2019-style fiscal and monetary easing, but a more fiscally driven expansion: “there is always someone issuing debt, which is equivalent to someone always dumping it.” Earnings power will therefore matter more than liquidity alone.
AI chips may be comparable to “steel in 1870”: annual growth can stall, while long-term demand could still expand hundreds or thousands of times. Cloud providers initially planned to increase investment by 15% in early 2024, but ultimately delivered about 55%; their 2025 plan was 20%, yet also reached roughly 55%. The current 2026 plan is about 18% and could still be revised higher. Nvidia growing more than the market’s roughly 20% expectation next year remains “quite likely,” but another 50% growth year is unlikely.
The US employment data are weak enough to justify rate cuts, and the economic trough may come between 4Q25 and 1Q26. August nonfarm payrolls rose by only 22,000, averaging roughly 30,000 over the past 3 months, with July revised up by 6,000 and June revised down by 27,000. Tariffs, restrictions on small-package imports, and deteriorating Chinese exports to the US will continue to weigh on the data, but this is a “short-term illness” that does not alter the two long-term cycles.
Nasdaq’s push to tokenize stocks could strengthen RWA, but may also drain liquidity from native crypto and altcoins. One problem with existing securities RWA is that stocks and tokens cannot be fully exchanged, leaving no complete arbitrage loop. If approved, the entry of state-backed players would provide a more powerful route for securities tokenization. During the last ETH rally, capital could more easily spread on-chain into altcoins; this time, BTC and ETH gains can be converted directly back into dollars, while US equities offer leverage and are easier to buy, weakening the mechanism for a broad altcoin rally.
The concentration of investment bankers and consultants asking about tokens and tokenized US equities looks to the speaker like a market-top signal akin to “the shoeshine boy giving you stock tips.” He remains explicitly bullish on 3 areas—RWA, stablecoins and stablecoin companies, and tokenized US equities—but thinks the altcoin market is nearing its end. That does not apply to Bitcoin: BTC could still benefit from fiscal expansion and follow an independent path.
Deep dive
1. The Fed Will Rescue Growth, but Not Return to a 1%-2% Rate World
After Jackson Hole, the market is pricing a policy rate around 3% by mid-2027, which the speaker considers “pretty fair.” The odds are reasonably high of 1 cut each in September, October, and December 2025, but a 50bp September cut is “not particularly likely.”
His explanation of Trump’s pressure is highly personal: he believes Trump sees Powell’s rate cuts during the previous election as having helped Biden—“he takes everything personal.” But that political resentment does not mean the Fed will unconditionally force rates down to 1%-2%.
Even if Trump appoints a loyalist, the new chair could, like Bessent, have more influence with Trump and persuade him. Whoever occupies the Fed’s top seat, the speaker still expects the institution to take a fairly balanced, rational view of markets and rate cuts: “I think the probability of sustained, aggressive cuts is very, very small.”
2. The 10-Year at 4.05% Is Near Its Downside Limit; Fair Value Is Closer to 4.2%-4.5%
The speaker’s core trading view is that “4% on the 10-year is a move that has run its course.” Even if the Fed stops or slows QT, with the policy equilibrium rate around 3%, there is limited room for a further sharp decline in long-end yields.
Fiscal expansion is pushing up both funding demand and inflation risk. The US government keeps issuing debt, like a stock that “always has someone bidding it up”; for yields, that means a constant supply of bonds being dumped into the market. Renewed financing needs over the next year will cap the upside in long-duration bonds.
China’s post-2008 experience is his key comparison. Even with heavy monetary expansion, large-scale fiscal spending under local-government influence once made local-government bonds appear risk-free, while yields on many wealth-management products also remained high. “When the demand for money rises, interest rates don’t stay very low.”
In terms of positioning, he does not recommend chasing TLT or the long end around 4%; insurers locking in 4% for asset-liability matching are a separate case. With yields around 4.2%, a swaption positioned for higher yields could be worth a shot; that does not necessarily mean directly shorting the note. The curve could steepen, mainly because of front-end rate cuts. He also noted that 4.2%-4.5% is only his personal market read and may not be highly accurate.
3. Fiscal Expansion Replaces Deleveraging, and Earnings Will Overtake the Liquidity Narrative
From 2008 to 2019, the economy was in an “ice age” cycle: post-crisis deleveraging, weakening demographics, and a mobile internet that did not lift efficiency as dramatically as the earlier internet revolution. Demand remained structurally weak, leaving low rates to sustain roughly 2%-3% growth.
After 2019, policymakers began using fiscal policy to “directly inject blood” into the economy; the pandemic proved that this approach can directly generate inflation. The fiscal cycle in Europe and the US is only beginning, and markets may swing between deflation fears and inflation fears over the next few months rather than trade a one-way recession narrative.
Rising support for the global right is reinforcing this direction. The typical policy mix is tax cuts and deregulation, while deficit reduction is usually harder to deliver. In the US, financial deleveraging could inject “more than $1T” of liquidity into risk-based and non-risk-based capital; if tax cuts pass in November, the effects should emerge over the following months.
This cycle looks more fiscal than the fiscal-and-monetary dual cycle of 2019. The ECB, for example, may not cut rates while Germany can still launch fiscal expansion. Investors therefore should not mechanically replay the old “buy gold and Bitcoin” script; “corporate earnings will become more important.”
4. AI Is Steel in 1870: Annual Volatility Does Not Change the Exponential Demand Curve
The speaker compares chips to “steel in 1870” and AI compute to railways and skyscrapers. Steel demand may have been temporarily insufficient in 1871, but over a longer cycle output could eventually grow 1,000-fold. Demand for AI compute and tokens could likewise grow hundreds or thousands of times over several years, 5 years, or 10 years.
The social consequences will be equally severe: “Machines displaced people who were good at physical work; AI will displace people who are good at mental work.” Conventional high achievers, computer science, and basic programming demand will all face substitution. ASICs and algorithms may be next after GPUs, but every track is part of the same major cycle.
Capex from 4 cloud providers plus Oracle is rising exponentially, with investment driving both EPS and margin expansion. That differs from the declining marginal efficiency of investment after 2008: “The marginal efficiency of investment is now rising.”
He remains cautious about Nvidia’s growth at the annual level. The market expects roughly 20% growth next year, and he sees a meaningful chance of more than 20%, but “another 50%” is not realistic. The key counterexample is that cloud providers revise plans dynamically: 15% became 55% in 2024, 20% became 55% in 2025, and 2026’s 18% may not be the final number.
He also warned that AI and chip companies could still be dragged lower if the economy deteriorates, and that capital will rotate among AI subsectors. Over the longer cycle, however, GPUs, ASICs, and algorithms could all benefit.
5. Employment Is the Immediate Ice; the Trough May Come in 4Q25-1Q26
August nonfarm payrolls rose by only 22,000; July was revised up by 6,000, June revised down by 27,000, and the 3-month average is roughly 30,000. Even with the S&P at an all-time high, this sequence “tells you something” and is enough to justify rate cuts.
The speaker sees the Fed’s priority as simple: “Growth is greater than everything.” This is not entirely about Trump’s pressure; when employment deteriorates this severely, the probability of rate cuts is already very high.
Tariffs, restrictions on small-package imports, and Chinese exports to the US being “a complete mess” will keep weakening US data, while China will absorb the initial shock. He had already expected in February and March that “after summer, the data would deteriorate.”
The economic trough could fall between 4Q25 and 1Q26. Even if tax cuts are enacted in November, transmission will take several months; it is impossible to know whether the low comes in January or April. But this is only a “short-term illness.” The long-term themes of fiscal expansion and AI remain intact.
6. Stock Tokenization Helps RWA but Shrinks Altcoins’ Room to Survive
Nasdaq has applied to the SEC to tokenize stocks. One problem with existing securities RWA is that many stocks and tokens cannot be fully exchanged, eliminating arbitrage. If approved, the speaker sees it as “the state team entering RWA,” giving securities RWA a more powerful push—but not signaling a crypto bull market.
Many existing platforms, including Milestone, are essentially stock-gambling platforms where the platform and market maker bet against users. Securities tokens could bring liquidity to blockchain markets while also siphoning liquidity from native crypto: buying US stocks directly does not require holding ETH, and the speaker believes US equities offer leverage and may even outperform ETH. If regulation loosens, platforms that survived in lightly regulated jurisdictions could lose room to operate.
In the last bull market, ETH was the base currency of crypto. An ETH rally effectively expanded on-chain “M2,” making it easier for capital to flow into altcoins. This time, gains can be quickly converted into dollars, making a broad altseason difficult to reproduce.
Bitmain is building an ETH treasury on one hand while using treasury funds to trade small coins such as WLD on the other, reminding the speaker of the previous cycle’s model of using ETH to speculate on altcoins. Some investment-banking and consulting contacts who did not buy when he advised them to buy Bitcoin at just over $20,000 2 or 3 years ago are now asking about tokens and crypto opportunities. He compared this to “the shoeshine boy giving you stock tips.”
He is explicitly bullish on 3 areas: the RWA market, stablecoins and stablecoin companies, and tokenized US equities. But he believes the altcoin market is nearing its end. That does not apply to Bitcoin; BTC could follow a completely different path during the major fiscal-expansion cycle.