Summary

The biggest significance of this round of U.S.-China tariff talks is not the short-term rate cut, but that China turned “equal pullbacks under maximum pressure” into a new model. Markets had assumed the first talks would produce no substantive result, while Trump’s stated reasonable range was 50%-80%. Instead, the two sides “reached a deal within a day and a half,” largely following “you cut by how much, I cut by how much.” Li Feng’s summary: what began as “the U.S. wanting to use China as a model” ended with “China using the U.S. as a model” (“中国用美国打了个样”); long-term preparation collided with short-term decisions made barely 100 days into the administration, and the outcome did not follow the U.S.’s preset Scenario A.
The figures often cited as the U.S. retaining 30% and China 10% do not represent a complete effective-tariff comparison; Li Feng’s rough trade-weighted calculation is closer to 30% for the U.S. and roughly 20% for China. The U.S.’s 30% includes two separate 10% fentanyl-related hikes. China’s more than $100B of imports from the U.S. are concentrated in chips, medical products, energy and agricultural goods; chips and medical products bear roughly the baseline 10%, while energy and agricultural goods face an additional 10%-15% retaliatory tariff. The gap remains, but almost nobody expected China to negotiate reductions on an equal basis.
China’s negotiating outcome could expand the strategic options available to India, the EU, Japan and South Korea when dealing with the U.S., even if they do not copy it outright. Israel can eliminate its trade deficit through arms purchases, while the U.K. accepted the baseline 10% in exchange for targeted exemptions. China demonstrated another high-pressure negotiating model. India then filed tariff countermeasures at the WTO, while Europe said a deal might not be reached within 90 days and prepared new measures. “At least this option exists” may have more medium-term significance than any single tariff change.
What global long-duration capital really needs to reassess is the probability of winning in the U.S.-China competition over the next 10-15 years, and this round of talks will drive a slow but potentially multitrillion-dollar asset-allocation rebalance. Li Feng believes the worst year may have been 2023. Since the end of last year, changes in U.S. political priorities, the period around Chinese New Year and the outcome of the trade war have pushed sovereign wealth funds and mega-funds to pull back from allocations that may previously have been tilted 80/20 toward one side. Short-term volatility will continue, but this probability reset is the main capital-markets story.
Tariff détente does not mean technology détente; the AI supply chain may instead be pushed toward “two complete systems.” Wang Yiquan noted that on May 13 the U.S. Bureau of Industry and Security proposed restricting global use of Huawei Ascend chips and restricting the use of U.S. chips to train Chinese models. The blunt translation is: “I want the entire world to use Nvidia chips and American models.” That could benefit Nvidia in the short term, but over time it will force China to complete the loop from chips and software ecosystems to model applications. The main highway remains constrained by lithography tools and the CUDA ecosystem, but the “side roads” opened by DeepSeek-driven inference, ASICs and heterogeneous chips could produce the next major route.
The probability of a U.S. recession and rate cuts fell sharply after the agreement, but Li Feng believes much of that move was emotional recovery after “a portion of enormous pressure suddenly came off.” China’s April exports rose by more than 9% year on year, well above the 1%-2% expected, although direct exports to the U.S. fell by more than 20%. U.S. inflation did not jump materially, helped by oil falling from $63-$64 to roughly $55. Wang Yiquan added that the real risk lies in services: if tighter immigration policy disrupts the supply of low-wage labor, wages and services inflation could rise together while labor demand falls, trapping the Fed in a stagflationary economy.
The recent appreciation of Asian currencies looks more like a repatriation of dollar assets amid uncertainty than a new “Plaza Accord” engineered by the U.S. The Hong Kong Monetary Authority triggered the strong-side convertibility undertaking several times in May, signaling repatriation pressure. The renminbi initially weakened on tariff shock and expectations of depreciation as a hedge, then offshore yuan rapidly moved from roughly 7.25 back to 7.20 after the agreement. Li Feng’s instinct: “It still seems more sensible to bring the money home first.”
Domestic policy is simultaneously supporting services consumption, capital-markets liquidity and technology financing; the banking system’s second transformation is more important to track than any single consumption number. Zhang Xiaoyu noted that during this year’s May Day holiday, traveler numbers rose by more than 6% from 2024 and total spending by more than 8%, implying higher per-capita spending. More importantly, lower-tier city residents are increasingly visiting other lower-tier destinations, accelerating the spread of experience-led consumption. On April 25, policy language shifted from “stabilizing the capital market” to “stabilizing and invigorating it”; on May 7, authorities announced RMB500B in relending for services consumption and eldercare, while using technology bonds and AICs to channel bank capital into direct financing. Zhang’s view: the turn that did not fully happen in 2018 “should have happened this time.”

Deep dive

1. A day-and-a-half deal shattered the consensus that a prolonged deadlock was inevitable

Li Xiang relayed a description circulating among friends of major news as “streaming-like, like a waterfall”: by the time an analytical framework has explained one action, the action itself may have been overturned 3 days later. The U.S.-China talks were the clearest example.
The reasonable pre-talks expectation was that the first meeting would produce no substantive result, followed by repeated haggling during the 90-day cooling-off period, with the eventual tariff rate perhaps landing in Trump’s previously mentioned 50%-80% range. Instead, the two sides “reached a deal within a day and a half,” at rates below that range.
Claims that a delegation had walked out midway were initially just self-media rumors, saying the group had gone out for lunch. State media quickly clarified that talks continued after lunch and into the following day. Li Feng therefore judged that this did not look like a process in which everything had been agreed in advance and the meeting merely formalized it; substantial content appeared to have been negotiated on site.

2. “U.S. 30, China 10” ignores fentanyl tariffs and import weights

Of the 30% retained by the U.S., 20% came from two separate 10% fentanyl-related hikes imposed since February. Chips and medical products imported by China from the U.S. largely face the baseline 10%, so comparing the two headline numbers alone is misleading.
China’s more than $100B of imports from the U.S. are concentrated in chips, medical products, energy and agricultural goods. Chips and medical products largely received no additional retaliatory tariff, while energy and agricultural goods face another 10%-15%; those two categories account for more than 30% to 40% of imports, close to half.
Li Feng’s rough weighted calculation puts China’s effective blended rate at roughly 20%, versus roughly 30% for the U.S. Even on a fully reciprocal basis, a gap remains; the more unusual result is that the two sides actually used a “you cut by how much, I cut by how much” pullback mechanism.

3. China rewrote the U.S.’s intended surrender model into a model of reciprocity

Israel was the first country to reach an arrangement with the U.S. Li Feng speculated that Israel occupies an exceptionally special position and may also have been able to erase its deficit through weapons purchases. The U.K. subsequently accepted the U.S. baseline 10% tariff in exchange for exemptions on specific goods, which was likewise not comprehensive reciprocity.
China had already lived through the tariff conflict 6 years ago, and had spent at least more than a year preparing since Trump formally entered the presidential race and reiterated his tariff agenda. The U.S., by contrast, was roughly 100 days into the administration; the escalation was abrupt and may not have been part of the original playbook.
Li Feng described it as “a contest between a long-term decision and a short-term decision”: the U.S. likely assumed China would quickly give in, then planned to use that outcome to force concessions from other countries. But “in the end, everything was different from A.”
That reversed the original plan of “the U.S. wanting to use China as a model”: “In the end, China used the U.S. as a model. You can see that I did not surrender, and I still obtained these at-least-temporary reciprocal conditions.”

4. The new model is expanding other countries’ strategic room to negotiate with the U.S.

The day after China announced the result, India said it had initiated tariff countermeasures against selected U.S. imports at the WTO. Li Feng explicitly reserved judgment on causality: he could not confirm whether China had influenced the move, but “it may have had an impact” given the timing.
Europe said it might not reach an agreement within 90 days and was preparing new tariff countermeasures. At the same time, signals emerged between China and the EU around dropping comprehensive restrictions, inviting EU leaders to visit China and marking the 50th anniversary of diplomatic relations. Together, these point to a slow medium-term shift.
During Trump’s Middle East trip, he reached an arms-sale agreement with Saudi Arabia worth roughly $140B. Li Feng believes the trip and the amount could not have been decided at the last minute within 2 or 3 days; they were more likely reactions formulated by various countries after the U.S. launched its tariff threats in early April.
Japan, South Korea, India and the EU may ultimately choose not to replicate China’s approach; their state capacity and bargaining chips differ. But “China gave a new model” means future negotiators will know that “at least this option exists.”

5. Trade outcomes will enter global long-duration capital’s 10-to-15-year probability model

Li Feng believes the largest sovereign wealth funds and mega-funds outside the U.S. are still betting on “who wins, and by what margin” in the U.S.-China competition over the next 10-15 years. This is long-term asset allocation, not market timing.
Li Feng believes the worst year may have been 2023. Since the end of last year, changes in the U.S. government and political agenda, developments around Chinese New Year, the trade war and this agreement have all forced long-duration capital to recalibrate its prior probabilities.
If the previous allocation was an extreme 80/20 tilt toward one side, even a small move back could involve several trillion dollars. The process will be slow, but like China’s creation of a new negotiating model, it is a medium-term impact more important than short-term market action.

6. The day after tariff pullbacks, AI restrictions immediately reminded markets that détente is not linear

On May 13, the U.S. Bureau of Industry and Security issued broad but vague rules: first, restricting global use of Huawei Ascend, with users potentially placed on the Entity List; second, restricting the use of U.S.-made chips to train Chinese models.
The latter will be difficult to trace in practice, but directly constrains AI companies operating in the U.S. whose technical teams may be in China: using Huawei could trigger the first rule, while using Nvidia to train Chinese models such as DeepSeek could trigger the second.
Wang Yiquan’s simplified translation was: “I want the entire world to use Nvidia chips and American models.” That could increase demand for Nvidia in the short term, but over time, as in the 2019 technology war, it could force China to build an independent, complete AI supply chain.
Huawei was also scheduled to hold a product launch in Germany on May 15, returning to Europe after 6 years. The products were mainly smart devices such as watches rather than Ascend chips, but Wang Yiquan believes the event was macroeconomically connected to China-EU cooperation and U.S. restrictions.

7. Russia shows that extreme sanctions do not necessarily force capitulation, but its GDP resilience has special sources

Li Xiang added another “model” from the Russia-Ukraine war: the world is watching both Russia’s actual capabilities and how much sustained support Western allies can provide to a smaller country.
Li Feng believes Russia’s refusal to compromise under collective extreme sanctions and the pressure of active war provides another example. Its roughly 4% GDP growth over the past 2 years partly reflects rigid demand for energy and commodity exports.
Another portion comes from “civilian-to-military conversion”: a prolonged war continuously consumes resources, while defense demand pulls up GDP. Civilian and household sectors may be squeezed as a result. An economy not being dragged down does not mean its growth structure is cost-free.

8. Why the U.S. became more eager to negotiate still has no single answer, and April data offer no final evidence

Possible drivers include rare earths, inventory building for the summer and back-to-school seasons, feedback from major retailers, and voters beginning to worry about their quality of life and losing confidence in economic policy and Trump. Li Feng’s honest conclusion: “I can’t tell” which factor mattered most.
China’s April exports rose by more than 9% year on year, well above the market consensus of 1%-2%. Direct exports to the U.S. fell by more than 20%, but the decline has not yet derailed overall exports.
The upside surprise may reflect transshipment through Southeast Asia, or revenue from first-quarter front-loading showing up later in the statistics. It proves that April aggregate exports remained strong, but does not show that tariff effects have disappeared after May.
U.S. inflation in April also did not surge as pessimists expected. Both sides had reasons to compromise, but they could not continue confronting each other over an extended period using the previous tariff levels.

9. Policy communication is speeding up, while CPI and PPI need to be separated

During the talks, state media quickly clarified that negotiations were continuing in response to social-media rumors. After He Lifeng returned to China, the Ministry of Commerce held an unscheduled press conference, then released the complete agreement at 3:30 p.m. the next day, seeking to compress the space for rumor-driven narratives.
Li Feng linked this to policy communication since last year, including the “9/24” financial press conference: officials have begun responding more promptly to suggestions, questions and data interpretations, “putting out positive first-hand information before other interpretations and information emerge.”
China’s April CPI “fell a little year on year and rose a little month on month.” Officials emphasized that the month-on-month reading was better than normal post-holiday seasonality. Food remains the largest weight, energy prices exerted a negative drag, while services prices showed a clear recovery.
The renewed declines in PPI on both a year-on-year and month-on-month basis more likely reflect the synchronized weakening of energy and commodity futures and spot prices after the tariff war depressed global growth expectations. Not every decline should be attributed to domestic end demand.

10. The same oil-price drop pushed down China’s PPI and U.S. CPI

Weaker global growth expectations and an unexpected increase in OPEC output pushed crude temporarily from $63-$64 per barrel to roughly $55. For China’s industrial system, the effect appears more directly in lower PPI; for the U.S., it is more visible in CPI, where energy carries a higher weight.
U.S. shale oil and shale gas have relatively high extraction costs. Li Feng estimates that producers need “the high fifties, above $60” for reasonably attractive profits; falling below $60 would materially increase pressure on producers.
Li Feng initially thought Trump’s Middle East trip might include discussions on coordinating oil production and prices in addition to arms sales, Boeing aircraft and investment in the U.S. He stressed that the news did not report such an agenda, so it remains speculation.

11. Post-deal rate-cut and recession odds fell, perhaps with too much emotional release

After the first-stage U.S.-China outcome, markets sharply lowered the probability of a U.S. recession and simultaneously cut the probability of Fed rate cuts. Li Feng questioned the scale of the move because both repricings were driven by sudden pressure and its partial removal.
His analogy was that when people “suddenly face enormous pressure, and then part of that pressure suddenly comes off,” their perception can reverse sharply. That does not necessarily mean economic fundamentals improved by a comparable amount in 2 days.
Li Xiang believes Powell’s stance remains firm. Wang Yiquan sees him more as watching a seesaw with constantly changing weights: tariffs, commodities, employment, immigration and capital flows keep shifting the balance, and even the experience of the previous round may not apply to the next.

12. Asian-currency appreciation looks more like capital repatriation than a forced replay of the Plaza Accord

The dollar weakened against most currencies in April. Some Asian currencies appreciated significantly, while the renminbi initially weakened against the dollar and also declined on a relative basis against the euro, pound and other Asian currencies.
The Hong Kong Monetary Authority triggered the strong-side convertibility undertaking several times in May, requiring it to sell Hong Kong dollars and buy U.S. dollars. Li Feng inferred that trade uncertainty was prompting Asian capital to sell some dollar assets and return home: “It still seems more sensible to bring the money home first.”
Markets at one point suspected that Japan, China or other Asian holders were selling Treasuries. He was unsure of the specific actors and leaned toward viewing it as common behavior driven by insecurity, rather than a unilateral U.S. order for Asian currencies to appreciate.
After the agreement, offshore yuan moved quickly from roughly 7.25 back to 7.20. If the earlier depreciation had mainly been an active hedge against tariffs, it should have remained weak while a 20-to-30-basis-point gap persisted. The rapid rebound instead supports the interpretation that growth expectations and depreciation fears recovered together.

13. Trump’s rate-cut horizon is 1 to 1.5 years, not a full economic cycle

Li Feng summarized Trump’s decision-making as linear reasoning: “If the economy is bad, we should cut rates”; “If we have a trade deficit, we should impose tariffs.” Other second-order effects are temporarily set aside.
During his previous term, Trump fought with the previous Fed and failed to push through rate cuts, but did prevent the Fed from continuing to hike. He then promoted large corporate income-tax cuts in 2017, using fiscal policy to create short-term stimulus.
The reason is that the midterm elections in the second half of the following year would determine whether Republican candidates could win, and how many people would still be willing to follow his policies. “The timeline for each individual year matters most,” rather than the total outcome over 4 to 5 years.

14. Tighter immigration could turn a services boom into services inflation and stagflation

Wang Yiquan believes the resilience of the economy during the second half of the Biden administration was tied to service jobs in education, health care, construction, restaurants, lodging and entertainment. A substantial share of low-wage jobs was filled by undocumented immigrants who had obtained work authorization.
The cycle relies on cheap labor supporting services supply, with employment and consumption resilience then feeding back into finance and spending on food, drink and entertainment. If Trump tightens immigration and disrupts labor supply, companies would have to raise wages to hire, while demand for positions could also fall.
Higher underlying costs in services would materially increase services inflation, while the “double-high” cycle of employment and job demand could break. Goods inflation caused by tariffs might not even be the biggest problem.
In the optimistic scenario, the economy and AI-driven growth remain strong, so the Fed does not need to cut. In the pessimistic scenario, the economy weakens while inflation persists, leaving rate cuts trapped by “stagflation.” Rate cuts become feasible only if reality lands between the two, toward the middle.

15. War technology has moved from precision automation and unmanned systems toward full-system integration

The Iraq War showed the world the power of information warfare and advanced automated weapons: targeted kills and precision strikes could replace some broad-based military activity. Wang Yiquan speculated that U.S. arms sales and weapons may now be entering a higher phase.
After the Russia-Ukraine war entered 2023, drones and unmanned boats were used extensively for offense and defense in a major war for the first time. The restrictions, easing and renewed restrictions around DJI also reflect the rising importance of unmanned equipment.
Much of the information from the India-Pakistan conflict remains impossible to fully verify, but Wang Yiquan believes it may reveal the next phase: once early-warning aircraft, radar, satellite communications, data, aircraft and air-to-air missiles are connected, a complete system can overwhelm equipment that is excellent in isolation but cannot coordinate effectively.

16. The next arms-sales cycle will sell “delivered outcomes,” not individual weapons

Li Xiang summarized it as: “What gets delivered is the outcome, the solution, the integrated system.” From satellites to aircraft, missiles, phased-array radar, electronic systems and cyber systems, hardware, software and control across time and space must work together as one system.
Wang Yiquan compared Huawei and Apple’s smartphones: in an environment with extremely poor signal, if the carrier equipment, communications system, operating system and terminal chip are all optimized within the same architecture, the individual chip may not be the strongest, but connection efficiency can be higher.
Arms sales also carry real political content. Wang Yiquan relayed a child’s formulation: “The main thing is paying protection money, with strengthening your own defenses as a secondary goal.” Future buyers may purchase advanced individual systems from allies while also buying 2 complete alternative systems “just in case.”
Li Xiang’s reservation is that a complete system requires the buyer to open its entire gateway, something states will approach very cautiously. Wang Yiquan noted that the compromise is ultimately about whether the buyer is purchasing a protection relationship, individual superiority or a complete operational capability that truly works when it matters.

17. In catching up on the AI-chip highway, China’s hardest problems are lithography tools and the CUDA ecosystem

Wang Yiquan defines the first route as “using the entire Nvidia system to catch Nvidia,” meaning building general-purpose AI chips similar to Nvidia’s. China’s most prominent participant is Huawei Ascend, which has consequently become an explicit target of BIS hardware restrictions.
Huawei’s distinctive difficulty is solving the entire manufacturing process from software to hardware. Individual problems in materials and process technology may be addressed gradually, but the biggest challenge at present remains lithography tools.
When SMIC released its first-quarter results, it said it had been affected by certain “special events” from last year into this year without specifying what they were. Market speculation includes being unable to buy new lithography equipment or having to recalibrate processes on old equipment. Wang Yiquan did not want to confirm anything on the company’s behalf.
For most companies, the harder problem may be a CUDA-style software ecosystem: developers, tools, standards and habits all require years to build. Hongmeng would also be difficult to scale without a special moment and special reason, showing that ecosystem barriers often take longer to overcome than hardware gaps.

18. Inference and ASICs are splitting the main highway into many side roads that could grow large

Wang Yiquan’s investment view differs from “there is no Nvidia substitute for now, so there is only one path.” His team has invested in various heterogeneous and ASIC-based chips. One inference-chip company backed 2 years ago has not yet completed tape-out, but gained momentum as DeepSeek took off.
A year and a half ago, models were mainly compared on parameter count, input length and tokens; the core question was “who is bigger.” DeepSeek showed that “being bigger is not the only way to be better.” Optimization on the inference side can achieve better results with a different resource structure, also driving a phase of DeepSeek appliance-style deployment.
Intel-style core processors need to do the operating system, computing, storage, display and command processing reasonably well. Early Nvidia, by contrast, focused on making one thing—graphics—exceptionally well. Future application chips may likewise score 95 in a specific capability while needing only 60 elsewhere.
Whether in the cloud or on the edge in phones and computers, real-world applications will force compute demand to split according to cost and resource scheduling. Not every side road will work, but one could produce “the Nvidia of 20 years ago.” The main highway admits only the companies with the most resources; side roads offer more opportunities for new entrants.

19. Full-stack industrial completeness is becoming a harder-to-copy advantage than any single technology

From weapons and robotics to AI, competition is no longer about isolated soft capabilities but increasingly about tightly coupled complex systems. Wang Yiquan used embodied AI as an example: the robot’s “brain needs a body,” and algorithms must be coupled with hardware to achieve their full effect.
Zhang Xiaoyu noted that China has brought together supply chains, talent and scale in a relatively short time through industrial policy, industrial capital, investment capital and policy guidance. Not every investment becomes a productive asset, but a small number of winners will remain in China, while the length and breadth of the supply chain and its talent pool continue to accumulate. This is “using capital to buy time.”
Wang Yiquan sees the U.S. as a highly practical, utilitarian and money-driven country. If some advantages are shown to no longer exist, containment may cease to be the most effective approach, and policy may shift toward other forms of competition. When that happens cannot be determined.
Li Xiang warned that historical psychology can obstruct purely rational adjustment: long-term victories create a “we will always win” gene, and when an advantage is challenged, social sentiment may rebound like “it was always our Axe Gang bullying everyone else.” Wang Yiquan also believes China’s modern historical memory will intensify the impulse to push back under maximum pressure.

20. DeepSeek proved another technical path, but not who will win commercially in 3 years

DeepSeek’s first significance is that, at the frontier, it changed how the world sees U.S.-China competition in AI technology. Wang Yiquan compared it to a technology event of historical importance within a particular phase.
Its second significance is that it proved a strong model does not have to be built “the way OpenAI does it.” Through inference, cost convergence and open source, it opened a side road next to the main highway that could eventually lead to a major road.
But what kind of company DeepSeek will be in 3 years, and whether it can remain ahead after the side road becomes a major road, “is impossible to see clearly today.” From 2012 to 2014, many big-data startups looked impressive; more than a decade later, very few had actually reached the front.
Wang Yiquan repeatedly emphasized: “Technology does not make money when it is valuable; technology is no longer valuable when it makes money.” When technology is sophisticated, 99% of people do not use it. Once 50% of people use it, it has become a general capability, and value shifts to supply chains, products and application companies. Autonomous driving may go through the same process.

21. May Day data are not simply “more people, less money”; the denominator and customer base have changed

Compared with the 5-day May Day holiday in 2024, this year’s traveler numbers rose by more than 6% and total spending by more than 8%, so per-capita spending actually increased. Consumption recovery remains weak, but the data do not support saying that per-capita spending fell year on year.
The comparison with 2019 is more complicated: May Day lasted 4 days in 2019 and 5 days this year, so the statistics use a “comparable basis” to adjust the totals. Fixed high-cost items such as transportation and lodging are spread across more days, making per-person, per-day figures naturally harder to increase.
The holiday system has repeatedly been adjusted to support consumption. Golden Week appeared in 1999; in 2008 it was shortened while Qingming and Duanwu holidays were added; in 2019 it was adjusted to 4 days; during the pandemic it was expanded to 5 days. This year added 1 day while requiring only 1 day of make-up work. Zhang Xiaoyu speculated that the pattern may continue, but made no firm conclusion.
The more important observation is the change in travel flows. During the pandemic and immediately after reopening, the pattern was “higher-tier cities going to lower-tier cities, and lower-tier cities going to higher-tier cities.” This year it became “people from higher-tier cities going to lower-tier cities, while people from lower-tier cities are going both to higher-tier cities and to other lower-tier cities.” Residents of lower-tier cities are beginning to visit other counties and towns for distinctive experiences, materially broadening participation in tourism consumption.

22. Services consumption and banking transformation are becoming new anchors of domestic capital circulation

The spread into lower-tier cities during May Day and the Pangdonglai phenomenon point to the same shift: lower-tier cities are no longer seeking only the lowest price. They increasingly want products and experiences that are good while still reasonably priced—a spread of experience-led consumption rather than one-way consumption downgrading.
Alcohol data show a similar barbell: Moutai, Wuliangye and Fenjiu, the top 3 brands, are still growing, while mid-tier and mass-market baijiu remain broadly weak and beer is performing well. This could reflect a shift from RMB100-200 baijiu to beer, or an expansion of drinking experiences and craft-beer consumers; both mechanisms may be operating at once.
On April 25, policy language described the capital market as needing to be “stabilized and invigorated” for the first time. On May 7, authorities launched RMB500B in relending for services consumption and eldercare. Zhang Xiaoyu believes policy is seeking not only to stimulate consumption through appliance purchases, but also to use eldercare, finance and other services to drive consumption and increase capital-markets liquidity.
Technology bonds allow private-equity funds to receive support through a debt structure familiar to banks. AICs allow banks to establish new structures for direct investment or fund-of-funds investment. After the 5 pilot banks—Bank of China, Agricultural Bank of China, ICBC, China Construction Bank and Bank of Communications—Industrial Bank became the first national joint-stock bank approved, while China Merchants Bank and CITIC Bank also began applying.
This extends the unfinished transformation initiated by the 2018 asset-management rules. At the time, policy sought both to look through to underlying assets and dismantle funding pools, and to move banks from real-estate and local-government financing toward equity financing for the new economy. Now, as old assets lose their appeal, policy and market pressure are jointly pushing banks through a “second turn.”
Li Xiang asked whether duration, capabilities and risk were mismatched. Zhang Xiaoyu described it as “walking on 2 legs”: technology bonds have underlying central-bank participation, so some risks may be borne more by policy banks; AICs may start with the low-risk segments familiar to banks and gradually learn higher-risk investing. Whether duration, pricing and stage matching can be handled properly remains to be seen.
This contrasts with Chinese bank capital entering direct financing. In the U.S., disputes over government grants and taxes involving Harvard and other universities are prompting some university endowments to redeem mandates, mainly affecting secondary-market funds. China is bringing in its largest financial players for the long term, while the U.S. is seeing a temporary exit of long-duration capital.

Agent Neo launches: Derek and 拐子 on the ultimate AI creation tool The Quest for Autonomous Web Agents with Div Garg, Cofounder and CEO of MultiOn