Summary

The speaker’s base case is unchanged: the US economy maintains strong growth, inflation itself will not materially worsen over the next two quarters, and the Fed can therefore watch the data and cut rates 1–2 times this year. Wednesday’s core CPI met expectations, while growth data came in far stronger than expected; even with a short-term payrolls spike to 240,000, he still sees the broader trend centered around 150,000–200,000, or roughly 180,000: “The data are more volatile, but the underlying trend has not changed.”
The fair trading range for the 10Y Treasury is 4.3%–4.5%, and a break above that range now does not mean 5% is an easy next stop. Yields have already fallen from around 4.8%; even with markets pricing two Fed cuts, supply will keep the long end from coming down easily, but another move to 5% would require a degree of growth and inflation pressure that is not currently present. US assets remain in a bull market, but volatility will be high and the eventual move may be smaller than last year’s and the year’s before.
A single BOJ hike will not fundamentally reverse the strong dollar. The speaker sees a high probability of a January hike to 0.5%, but that outcome is already priced; even if the 2025 JGB 10Y rises to Goldman’s 1.6% forecast and the US 10Y falls to 4.3%–4.35%, the US-Japan rate gap will still be wide enough for the carry trade to make sense. The yen may do better against the euro and renminbi, but could remain weak against the dollar.
CNH may not have finished pricing Trump’s tariffs. CNH, a government-regulated rate, is included in the dollar index basket; the speaker says the People’s Bank of China has asked Hong Kong to manage the process centrally and convert its foreign-exchange reserves into CNH, yet CNH remained under pressure while holding around 7.3 over the past two weeks. Trump will likely announce China policy, perhaps easing somewhat in technology, while raising tariffs and collecting money in core areas. His stated fully priced range for the renminbi/CNH was “4.7 to 4.5”; the direction and units need verification and should not be rewritten as 7.4–7.5.
Credit spreads are already too tight; the more compelling alpha may be in Agency Bonds such as Fannie Mae and Freddie Mac. The speaker cites US credit spreads at roughly 260bp and says Agency Bonds may offer about 200bp over IG; in his view, these bonds are theoretically close to US Treasury risk, with the spread mainly reflecting a smaller buyer base, making them suited to long-and-hold rather than a bet on a sharp fall in rates.
The TRUMP coin put political rent-seeking and on-chain liquidity migration on the same stage, and changed the speaker’s assessment of Trump’s risk. Its value briefly reached roughly $60B before falling back to around $30B, and it was described as “incense money” anyone could buy; most of the initial wallets reportedly withdrew funds from OKX and Gate, after which the coin surged on-chain and was quickly listed across exchanges. Solana and other chains may benefit from the near-term activity, but the bigger signal is that “he is here to make money.” His wife Melinda then said she would issue a coin too, adding that his son would do the same.
If Trump governs mainly as an influencer, policy may be loud in rhetoric but slow in execution, reducing but not eliminating the odds of a black swan. China tariffs will likely be announced within 1–2 months, but may be “thunder with little rain,” with negotiations continuing to extract concessions; energy deregulation, withdrawal from the Paris Agreement, and offshore oil and gas drilling are likely to come first, while DOGE, universal tariffs, and mass deportation may be constrained by their economic and budgetary costs. On foreign policy, the speaker thinks Ukraine and Taiwan are more likely to come under pressure.
NVIDIA’s earnings visibility remains high this year, but the position should be sold or reviewed in March–June because 2026 depends on whether enterprise AI applications truly take off. Old COBAS S capacity used for H chips has reportedly been cut 80%, COBAS L has not fully caught up, and B100 packaging and rack-level coordination also have issues, but these may not derail this year’s results. The question NVIDIA cannot answer itself is enterprise AI demand; “you can bet this year’s earnings will be fine,” but you should not blindly bet on 2026.

Deep dive

1. Strong growth and tame inflation remain the US economy’s base case

The speaker first returned to his view before adding Trump’s policy variables: inflation itself will not materially worsen over the next two quarters; if conditions change, the more likely catalyst would be Trump’s tariff or immigration policies. The US growth backdrop remains strong and inflation relatively tame, so the Fed can wait for the data before acting. Two rate cuts this year remain his base case.
The latest data reinforced that framework. Wednesday’s inflation print was not especially good, but core CPI met expectations; at the same time, growth data came in far stronger than expected, easing fears that the US might hike rates or make no cuts at all this year. Rates also fell back from around 4.8%.
He does not want to overread a single-month payrolls jump to 240,000. Long-term charts for job openings and wage costs show that post-pandemic data have become more volatile, but the combined trend has not materially changed; employment overall remains consistent with his 150,000–200,000 range, centered around roughly 180,000.

2. Long-end rates will struggle to fall much; US assets can still rise, with more volatility

Last year he was a “big short” on the 10Y Treasury because issuance supply made the 10-year yield behave like Sisyphus pushing a boulder: “always pushing up, always falling back down.” Supply pressure remains this year, but is not as intense as last year, putting the fair year-end trading range at 4.3%–4.5%.
Even with two Fed cuts, the 10Y may not fall materially below that range, primarily because of supply. Conversely, another move to 5% would require stronger growth and a renewed inflation impulse, and he does not see those conditions in the market today.
US assets should still rise on a relative basis, but expectations for good growth and low inflation are already higher than they were last year and the year before. Realizing those expectations can still lift markets, but the path will be more volatile and the eventual move may be smaller than in the prior two years: “It is still a bull market, but there will be a lot of volatility.”

3. A BOJ hike will not overturn the strong dollar; CNH is the unfinished pricing

The dollar’s current strength was described as rare in the post-Plaza Accord era, rooted in economic divergence between the US and Europe and Japan. Trump’s decision not to impose China tariffs immediately on his first day did trigger a rebound in the renminbi and a sharp turn lower in the dollar, but the speaker sees this more as a correction after the dollar’s short-term gains had become excessive.
The probability of a January BOJ hike to 0.5% is high because waiting until inflation worsens further would carry greater risk. The decision not to hike in December was more about waiting for clarity on US tariffs, Trump’s policies, and his team’s stance. A January hike is already widely expected, so it may not provide the yen with a fresh, powerful catalyst.
Goldman’s 2025 forecast for the JGB 10Y is 1.6%, driven by both BOJ hikes and a rise in term premium as inflation expectations increase. But even if the US 10Y falls as low as 4.3%–4.35%, the US-Japan rate gap will remain wide and the carry trade will still make sense. The yen may perform better against the euro and possibly the renminbi, but could remain weak against the dollar.
The speaker says CNH is included in the dollar index basket and is a government-regulated exchange rate. He also says the People’s Bank of China has asked Hong Kong to manage the process centrally and convert Hong Kong’s foreign-exchange reserves into CNH; in his view, this reflects funding pressure, with CNH still under pressure despite holding around 7.3 over the past two weeks.
It remains unclear what policy Trump will ultimately adopt toward China, but he will likely announce something at the statement level. He may ease somewhat on technology or technology exports, while raising tariffs in China’s most core and painful areas to collect money. The original remarks put the range after CNH and the renminbi finish pricing at “4.7 to 4.5,” but the direction and units are unclear.
Fiscal spending could rise materially ahead of Germany’s election, and Europe may also use fiscal policy this year to offset recessionary forces. Taken together, these factors make it difficult for the yen to stand apart; as long as US long rates remain relatively high, BOJ hikes may not alter the dollar-led dynamics of foreign-exchange pricing.

4. After credit spreads get too tight, Agency Bonds offer hold-to-own alpha

Both credit and equities are priced richly: the speaker cites US credit spreads at roughly 260bp and believes rates will not fall much from current levels. This year’s alpha may therefore lie not in another bet on tighter credit spreads, but in structural bond spreads.
The portfolio he favors is Fannie Mae and Freddie Mac Agency Bonds versus IG, with a quoted spread of roughly 200bp. In his view, the underlying risk of Agency Bonds is close to zero, theoretically equivalent to US Treasuries; the spread mainly reflects a smaller pool of potential buyers. Held on a long-and-hold basis, their premium may be better than that of Treasuries and ordinary investment-grade bonds.

5. The TRUMP coin became on-chain “incense money” and exposed the risks of political rent-seeking

Trump issued the TRUMP memecoin before taking office. Its market value briefly reached roughly $60B before falling back to around $30B. What mattered most to the speaker was not the price move but that “it is not really clear who is buying”: anyone who wanted to send Trump money only needed to buy a little TRUMP, turning it into “incense money.”
At launch, most wallets reportedly withdrew funds from OKX and Gate. After the TRUMP coin was issued on-chain, its price surged and it was listed across all exchanges almost immediately. The speaker contrasted this with traditional token listings, where an exchange might take 70% of the liquid float and 30% of the whole pool, effectively making the exchange a major shareholder of the listed stock; TRUMP could not be handled that way and instead used on-chain trading to bypass centralized exchanges and direct funds on-chain.
Under the framework that “95% of the money used to sit on centralized exchanges,” this liquidity migration could make Solana and other chains unusually active over the next several weeks, with on-chain capital rushing everywhere. Melinda then said she would issue a coin, and that Trump’s son would issue one as well, reinforcing the speculative signal.
But compared with money moving from equities into crypto—“a baby sucking money from a giant”—the more important change is Trump’s own motivation. The speaker thinks Trump may be highly interested in money, profits, and making money, and that this has changed his previous view of Trump.

6. “Influencer-style governance” means creating noise first, then weighing the cost of execution

The speaker says Trump is fundamentally an influencer, not a successful businessman: he claims Trump squandered the $700M his father gave him in the 1970s, never succeeded in business, and later made the money back through a TV show. On that reading, Trump may care more about attention, personal gain, and visible wins than about fully implementing every policy.
He summarizes Trump’s agenda under three main themes: deregulation, especially in energy and oil; immigration and its supposed security concerns; and tariffs. China tariffs are likely because they can both extract money from China and be used to pressure China into granting further concessions, but they may not be strictly enforced after announcement—“thunder with little rain.”
The speaker expects China policy to arrive quickly, within 1–2 months, creating confusion on release but slowing at the execution stage while negotiations and cooperation with China continue: “the dance goes on, the song goes on.” If this view is correct, the practical result for China could be favorable, but that remains a conditional judgment. Technology may see some easing, while the most core areas could remain subject to higher tariffs.
Deregulation and energy policy are the easiest to implement first because, in his view, they would not affect the economy or relations with allies. Withdrawal from the Paris Agreement, offshore oil drilling, natural gas, and many environmental policies could move quickly. On Europe, he is still unsure whether tariffs will be imposed across the board; Trump may start with specific areas such as autos, while universal tariffs and their scale remain unclear.
DOGE, meaning deficit reduction, and mass deportation both face practical constraints. Deportation would require fiscal resources that are not yet in place and would reduce the labor force; Silicon Valley elites around Trump may not support the mass expulsions demanded by the MAGA camp. The speaker also believes that people who come to make money do not want the stock market to fall too far, so Trump may be reluctant to genuinely shock the economy.
If all the policies are implemented, Trump could still become a black swan, but the speaker thinks that probability is falling. His foreign-policy logic is closer to “an enemy’s friend is an enemy, and an enemy’s enemy is a friend,” making Ukraine and Taiwan more vulnerable. He also reads the attendance at the inauguration as a signal: Han Zheng was present, but TSM, Nvidia, and the Tainan faction were not.

7. NVIDIA’s earnings are relatively secure this year, but 2026 must be validated by AI applications

Market concerns center on the supply transition: old COBAS S capacity used for H chips has reportedly been cut 80%, COBAS L has not fully caught up, and B100 servers face packaging and rack-level coordination problems. These could create quarterly noise, but the speaker does not think they will derail NVIDIA’s earnings this year.
Based on the information he has received, including TSMC’s indication that NVIDIA’s earnings are not a problem this time, he believes this year’s results are secure. When he bought the dip around $90 last year, his calculation was that annual EPS would be at least $5: a 20x PE implied $100 and a 30x PE implied $150, making it possible to “buy without thinking.”
The real variable is 2026: whether enterprises begin experimenting with and calling AI, and at what scale, must be demonstrated jointly by cloud providers and front-end software. Cloud providers are doing well partly because enterprises are experimenting with AI and first moving their data to Amazon’s cloud, AWS, or Microsoft. That phase has already happened, but future AI call volumes and application demand remain difficult to assess.
You can bet that NVIDIA’s earnings this year will be fine, but do not blindly bet on 2026. If enterprise AI applications fail to take off, the impact will extend beyond NVIDIA to large cloud providers and front-end applications. The speaker still leans toward AI applications developing because IBM and consulting firms of all kinds are pushing a large number of zero-to-one experiments.
Even with problems around “72,” and possible problems with “36,” Microsoft is still willing to place orders directly before large-scale testing is complete in order to build compute capacity as quickly as possible. The speaker also believes that quarterly earnings at Microsoft and similar companies reflect conditions from roughly 6 months earlier, making misses by MSFT and Oracle in their fourth-quarter and first-quarter reports unlikely.
For NVIDIA, the focus should not be on tracking the narrative around any single quarter, but on watching the longer-term trend in AI applications. The speaker thinks it is fine to hold now, but around March–June, when first-quarter earnings are released, investors should sell or review the position and first see whether AI applications truly take off.

Verification Notes

The original remarks put the range after CNH finishes pricing at “4.7 to 4.5”; the reference text does not specify the direction or units. This summary preserves the original wording and does not infer 7.4–7.5.

Dylan Field - Designing The Future - [Invest Like the Best, EP.407]Hussein Kanji, Founder @Hoxton Ventures: Why AI Means London Can Compete with the US | E1248