硅谷Fintech观察④｜为什么中国出不了 StripeSilicon Valley Fintech Watch ④ | Why There's No Chinese Stripe

"Why There's No Chinese Stripe" series · Part Four (Finale)

In brief — China has the best payment experience on earth, yet it can't grow a Stripe of its own. This isn't a technology gap, still less a talent gap—Futu and Airwallex proved long ago that Chinese founders can build world-class financial products; they just have to build them offshore. So what's missing? The soil. This piece delivers the final answer to the whole series—and at the end, two pieces of unsolicited personal advice for ordinary readers.

Setting the Stage

In June 2026, Meta did something worth pausing over: it put $900 million into the Indian finance app Cred, at a valuation of around $4.5 billion; almost simultaneously, it installed Cred's founder, Kunal Shah, as WhatsApp's global head.

Cred is no bit player—it serves 17 million Indians and, by transaction count, processes more than 40% of India's credit card bills. But the truly interesting part is Meta's calculation: it wants to turn WhatsApp into India's payment super-app. Yet as of May 2026, WhatsApp Pay's share of India's UPI was still just 0.65%. One commentator nailed the crux in a single line—"WhatsApp is everywhere in Indian commerce, and almost nowhere in Indian finance."

Anyone who knows China will smile knowingly at that sentence. Because it's the mirror image of the Chinese story—except China finished this move long ago: once a super-app fills up a country's commerce gateway, independent financial infrastructure struggles to grow out from under its feet.

Put plainly: India's Google is still Google, but India's "JPMorgan" is becoming WhatsApp. Swap "India / WhatsApp" for "China / WeChat, Alipay, Meituan," and this piece's central question is sitting right in front of you.

Across the first three parts, I walked a complete chain: Part One—the unbundling of U.S. brokerages and asset managers finished long ago, fees fell to zero, no meat left on the bone; Part Two—the smartest money pivoted to attack the banks, but most who went after the consumer side fared badly, and the ones who truly rode out the cycles were the pick-and-shovel sellers; Part Three—enterprise payment infrastructure is the crown jewel, and Stripe, Adyen, and Airwallex each proved that "selling shovels" can be a business with software-grade margins.

But the end of Part Three, with Airwallex—that Chinese-founded specimen—already spoiled the shape of the answer: it could grow to $11 billion, yet had to put its headquarters, its data, and its growth outside China, and even actively keep its distance from the word "China" itself.

This is no accident. In this piece, I answer the series' ultimate question head-on—why can this layer of pick-and-shovel infrastructure grow into a crown jewel in America, yet barely grow at all in China?

My answer: China has never lacked fintech talent or product chops—Futu and Airwallex already proved that. What China lacks is the soil that grows a company like Stripe. That soil is made of four variables: enterprise willingness to pay for software; a unified-but-platform-captured market; independent distribution; and an M&A/exit ecosystem. Let me take them one at a time.

I. Soil Variable One: Whether Enterprises Will Pay for Software

Stripe, Adyen, Ramp—these "shovel sellers" were, at bottom, all raised by the same thing: American enterprises are willing to pay for software and infrastructure.

How lopsided is this? Shixiang once cited a set of numbers from Battery Ventures: total software spend in North America is roughly 13 times that of Greater China (over the same period, U.S. per-capita consumer spending was also about 13 times China's). If you find that yardstick dated, take a newer one: in 2019, China's SaaS market was just $3 billion, 2% of the global total (IDC); by 2025, the composite read from several research houses is still that "the U.S. market is roughly 10 times China's."

More damning is the value per customer. In China, about 60% of small and medium-sized enterprises are still on the fence about digitization, and only 3.2% have reached a genuinely high level of it; and the average lifespan of a Chinese SME is far shorter than an American one—when customers don't live long, software's lifetime value (LTV) can't hold up. The result is that China's enterprise SaaS has spent years marinating in a swamp of "red-ocean low prices and chronic unprofitability," with one study estimating that annual per-customer spending power in the U.S. and China differs by nearly tenfold.

(Plain-language aside: SaaS = enterprise software paid for annually; ARPU = average revenue per customer; LTV = the total value a customer contributes over their entire lifetime. Whether an enterprise keeps paying for software, and how long it survives, directly decides whether the shovel sellers get to eat.)

Shixiang put it well: "SaaS is a robot for white-collar workers." In America, software is an essential purchase for enterprises—the billboards lining Silicon Valley's highways are almost all SaaS companies advertising; in China, software is more of a "cut-it-if-you-can" cost line.

This gap in willingness to pay, landing on a single company, is staggering: in America, a decent enterprise software company can have a single large customer contribute tens of millions of dollars a year in recurring revenue—the data-warehouse company Snowflake's big client Capital One contributed tens of millions annually all by itself (per its early prospectus). It is precisely this rich soil—enterprises willing to keep paying for software—that grew Stripe, grew Adyen, and grew that entire row of "shovel sellers" from the previous part's table. Conversely, when a market's enterprises haggle even over software subscriptions and, on average, don't survive a few years, companies in this row can only subsidize, burn cash, and chase ever more downmarket traffic—they simply can't accumulate the kind of repeatable, self-expanding revenue that "selling shovels" requires.

No enterprise willingness to pay, no revenue soil for a Stripe. A company can build beautiful payments, but if no one will keep paying for "beautiful infrastructure," it won't grow into a crown jewel—it can only subsidize, burn cash, and fight for consumer traffic. This is the fated starting point of Chinese fintech.

II. Soil Variable Two: A Unified Market, but One Eaten Clean by Two Platforms

Someone will object: China's market is huge and unified—isn't that the perfect hotbed for a Stripe?

Quite the opposite. China's "unified market" is genuinely unified and genuinely large—but that unity was monopolized into being by two super-apps, not paved by a layer of neutral infrastructure.

Look at payments: Alipay and WeChat Pay together hold over 90% of China's mobile payments. Look at wealth management: on the back of Alipay's distribution, Yu'e Bao was briefly the world's largest money-market fund in 2017–2018, peaking at around ¥1.69 trillion (roughly $268 billion)—it proved China can conjure a giant financial player out of thin air, but the winner was the platform holding the users (Alipay), not some independent asset-management startup.

Nothing illustrates the point better than restaurant SaaS. In America, a company like Toast can grow up independently: starting from restaurant POS, embedding payments into the software stack, collecting SaaS subscription fees on one side and payment take-rates on the other, and growing into a public company.

Here I have to correct a popular claim. People often say "the reason there's no Toast in China is that Meituan embedded it and killed it off." The precise fact is this: Meituan in China built its own restaurant POS from the start (smart terminals, scan-to-order), bundling this vertical SaaS layer together with fulfillment and ad traffic into its own operation; it even went abroad in reverse, assembling a team to build a direct Toast rival called Peppr—which, because Meituan was unwilling to pour money into an overseas war, has already been shut down in the U.S. The in-and-out tells the whole story: in China, the restaurant-SaaS layer simply leaves no crack for independent players; and the super-app's playbook, transplanted to an America without a traffic monopoly, doesn't take to the soil.

This is the paradox of China's "unified market": the market is big enough, but every vertical opportunity—payments, credit, wealth management, restaurant SaaS—gets casually eaten at the gateway layer by Alipay, WeChat, and Meituan using ready-made traffic. An independent Stripe, an independent Plaid, an independent Toast can't find the crack to grow up on its own in a landscape where two super-apps fill every gateway.

Look one layer deeper: the reason super-apps can pick off vertical opportunities one by one isn't how good their products are—it's the thing in their hands that no one else has: ready-made traffic in the billions. An independent startup has to spend years and burn big money to acquire customers, whereas WeChat, Alipay, and Meituan need only add an entry point inside their own apps to casually do credit, wealth management, and restaurant SaaS, at a marginal customer-acquisition cost of nearly zero. This is exactly India's script: WhatsApp dares to dream of doing payments precisely because of its dominant penetration in Indian commerce, not because WhatsApp Pay itself is any good. When distribution can be handed out free via traffic, the road of independent infrastructure—crawling up slowly on product and licenses—looks both slow and unnecessary. This is the most counterintuitive thing about a super-app market: the bigger the market and the stronger the platforms, the smaller the crack left for independent, neutral infrastructure.

III. Soil Variable Three: Independent Distribution, and the Impossibility of "Building a New Network"

The third patch of soil is independent distribution channels and network moats.

As I said last time, the shovel sellers' real moat isn't technology—it's distribution and licenses. Here's a judgment that could serve as the whole series' thesis line: since Discover was born in 1985, no one in America has built a new payment network at retail scale—and that's unlikely to change. And Discover succeeded back then precisely because it rode the distribution of Sears, at the time America's largest retailer.

The barrier in payments also shows up in the layered nesting of licenses—someone likened it to a "Russian doll": open up one payment company and there's another nested inside, all propped up by licenses across jurisdictions; even the largest banks hold direct licenses in only around 60 markets.

A one-line analogy (echoing the whole series): a bank's core system is the bank's pickaxe; enterprise payment infrastructure is the pickaxe of the entire digital economy. Whoever sells the pickaxes stands upstream of every prospector.

So who holds China's "pickaxe"? The answer: no independent third party. UnionPay is the state team's clearing network; WeChat and Alipay are the closed loops of Tencent's and Alibaba's respective traffic platforms. In a market where the distribution gateways are filled up by the state team and the super-apps, a Stripe-like "neutral, developer-facing third-party infrastructure layer grown on enterprise payment" has never lacked technology—what it lacks is the independent-distribution soil it needs to grow.

Say it one layer more bluntly: America's fee-pool chain is commercialized—the card issuer is a commercial bank, the clearer is a listed company like Visa/Mastercard, every segment carries profit and can each be pried loose by a slicker third party. Stripe pried its way in from the "acquiring" segment. And China's chain? The issuing banks are state-owned, the clearer UnionPay is built by the state team, and the gateways are Tencent's and Alibaba's respective closed loops—from end to end, everything is either state or platform; no segment is left as a commercial fee pool for a "neutral third party." This is a more root-level layer than "the market is eaten clean": in America, the fee pool is a commercial cake that can be pried at; in China, it was never a cake open to outsiders at all.

Airwallex's solution is itself proof of this conclusion: it didn't go toe-to-toe inside China, but chose its battlefield in cross-border and overseas markets, laying a network outside China over nearly a decade with 85+ licenses. It even wants to move its staff and data out of China, and has to answer accusations of a "China backdoor"—Keith Rabois, a Khosla partner who also sits on the board of rival Ramp, publicly alleged that its China team and Chinese investors may leave it "obligated to hand over customer data," and U.S. senators have already called for an investigation. A global infrastructure company with Chinese DNA, to grow up, must first prove it is far enough from China.

IV. Soil Variable Four: The M&A and Exit Ecosystem—America Even Made "Getting Acquired" Dignified

The fourth patch of soil, and the most underrated: whether there's a path for a startup to exit with dignity.

America's exit ecosystem is like a densely woven net—every strand a road. In just the last year or two, we've seen: Brex acquired by Capital One for $5.15 billion (closed); Bullish buying transfer agent Equiniti for $4.2 billion; Kraken's parent spending $600 million to buy Hong Kong stablecoin-card platform Reap. Even a failed acquisition isn't fatal—Adobe's $20 billion purchase of Figma collapsed on antitrust grounds and cost a $1 billion breakup fee, and Figma turned around and IPO'd independently. In America, getting bought by a big bank, by a card network, being consolidated by a peer, or simply going public yourself is an exit ecosystem so mature that "even failure lets you walk away whole."

The most fitting footnote is Stash's story. In February 2026, Southeast Asian super-app Grab acquired the U.S. investing app Stash at a $425 million enterprise value (first taking 50.1% of equity, with the rest collected at fair value within three years). What Grab bought wasn't Stash's code—it was its ready-made "48-state regulatory framework + $5 billion AUM + 1 million paying subscribers"—saving itself the time of building a U.S. financial-infrastructure stack from scratch.

But this Stash case has a colder truth underneath, practically tailor-made for this piece: Stash raised roughly $747 million in total, so the $425 million exit was actually an "underwater exit"—late-stage investors took a haircut and common stock went essentially to zero. Around the same time, even MrBeast's company acquired the teen bank Step. This batch of "second-wave neobanks" can't reach the IPO escape velocity of a Revolut, and can only be repriced and absorbed according to "how much tool-value they still hold for a super-app."

Turn the lens back to China, and the contrast is glaring: the A-share market isn't friendly to these companies, the window for overseas listings opens and shuts, and industrial M&A is far less active than in America; the most emblematic scene of all was Ant Group's largest-ever IPO in 2020, halted at the final moment. Once the exit ecosystem is thin, there's less patient capital willing to bet on ten-year-long infrastructure—and companies like Stripe and Airwallex need precisely a decade of patience.

By the way, America's exit ecosystem has a counterintuitive feature: the strongest financial infrastructure actually stays private. Two of the world's three largest asset managers are private—Fidelity is family-held, Vanguard is mutually owned by its clients, Capital Group is held in partnership by its employees. Unshackled from quarterly earnings, they can bet on the longest cycles. This "private giant" form barely exists in China's financial system.

The depth of American capital markets also shows in how much room it leaves for different exit forms. On one side, IPOs come ever later and ever more exclusively serve insiders—when SpaceX went public, someone lamented that "for an ordinary American, your first chance to buy the stock is at a $2 trillion valuation"; on the other side, even an alternative asset like private credit has opened a door to retail investors (though in Q1 2026 retail redemptions clustered, several semi-liquid funds hit redemption gates at once, and Goldman Sachs expects private credit to keep seeing net outflows over the next two years—there's fragility hidden inside the boom, too). But whatever the form, America has prepared a full set of pipes that "let you in and let you out"; the more numerous and free-flowing the pipes, the more patient capital dares to bet early on a ten-year run. This is exactly the foundational condition that companies like Stripe and Airwallex—the ones that "have to burn for a decade"—need most.

Not to mention that the depth of the capital markets isn't remotely in the same league. How deep is the U.S. equity pool? It can swallow a $2 trillion behemoth IPO like SpaceX in one gulp without even a ripple of liquidity—even if a second $2 trillion came at the same time, it would take that too. Hong Kong? Forget $2 trillion—it can't even take a single ByteDance: actually floating a book the size of ByteDance on the Hong Kong exchange would instantly drain the entire market's liquidity. An exit needs not only a buyer (M&A) but also a harbor deep enough to hold a big ship (capital markets)—America has both, China is thin on both. That's why, for the same "ten-year burn," money dares to invest in America but not over here.

V. Pressing Down on the Soil: Regulation, and That Dollar Rail

Beyond the four patches of soil, there's a harder layer pressing down—regulation's attitude toward "competition."

An often-overlooked detail: the White House's 2026 fintech executive order was the first to write "promote competition, clear away rules that obstruct innovation" plainly into the regulatory objectives. As one commentator put it bluntly: "Before this moment, I don't think there existed a clear 'promote competition' mandate in regulation." That even America has to issue an order specifically to encourage competition shows how hard the thing is.

In China, the direction is reversed. In May 2026, the CSRC and other agencies issued advance notice of proposed penalties against Futu, Tiger, and Longbridge, finding them guilty of illegally operating securities businesses: Futu faces a fine and disgorgement of over ¥1.85 billion, with CEO Leaf Li personally fined ¥1.25 million; Tiger, about ¥411 million in total. This is part of an eight-agency "two-year concentrated crackdown on illegal cross-border securities business"—apps have been pulled from mainland stores since 2023, existing mainland clients are permitted only to sell and transfer out one-way, and everything winds down entirely at maturity. In other words, Futu's most successful line of business (serving mainland cross-border clients) is being structurally sealed shut. This echoes exactly the conclusion of Part One: Futu is not "a Schwab grown in mainland China"—it's a global brokerage that a group of Chinese people had to go offshore to grow.

Finally, there's a hidden thread laid down in Part Three: on the newest financial rail—stablecoins—about 99% of the entire market is dollar-denominated. The more successful stablecoins are, the more they distribute dollars globally. Should—and can—a country that guards its monetary sovereignty through capital controls get on this rail? That question is itself one of the sharpest slices of the U.S.–China soil difference: America's financial infrastructure naturally stands in the tailwind of "global dollar distribution," while the RMB's rail is blocked by the wall of capital controls.

Even more counterintuitive is the scene below: knowing full well that stablecoins are already a dollar world—about 99% of the whole market pegged to the dollar, all local currencies together under 1%—Asian countries have still scrambled these past couple of years to issue their own. Japan will launch a regulated yen stablecoin in 2026, issuable only by banks and trusts; Hong Kong's Stablecoins Ordinance took effect in August 2025, purpose-built to license Hong Kong dollar stablecoins; South Korea is still bickering endlessly over legislation for a won stablecoin. What's the point? Monetary sovereignty. The logic is plain: if your citizens default to dollar stablecoins for their paychecks, savings, and payments, money leaks bit by bit out of the domestic financial system—this is "digital dollarization," a central bank's greatest fear. And there's a more concrete layer: a won-to-yen conversion today probably still detours through the dollar under the hood, adding a needless layer of cost and latency; issuing a local-currency stablecoin at least tries to save this "dollar toll." Put bluntly, you can decline to get on this rail—but declining means, by default, handing the monetary sovereignty on this newest financial rail over to the dollar. Even U.S. allies like Japan, Hong Kong, and South Korea are painstakingly laying their own narrow-gauge track beside the dollar rail; China's answer is more thorough still: rather than get on a public rail destined to distribute dollars, build the digital RMB (e-CNY) itself and pilot stablecoins firmly fenced inside regulation and capital controls. On the same newest rail, America has a tailwind, Asia a headwind, and China simply wants to lay a separate track—this is the sharpest divide that dollar thread casts on U.S.–China soil.

VI. Devil's Advocate: But China's Payments Are Clearly the World's Best—How Can It "Not Make a Stripe"?

As usual, let me steelman the counterargument in full before returning to the verdict.

The bull case, in earnest: when it comes to payments, China isn't behind—it's far ahead.

China's mobile-payment penetration is the highest in the world, and Alipay's and WeChat's QR codes make America's "swipe-a-card-plus-acquiring" setup look like a relic of the last era. Cross-border players with Chinese DNA are conquering territory worldwide, too—Airwallex can raise at an $11 billion valuation. Yu'e Bao, distributed via Alipay, was once the world's largest money-market fund. And Futu built a global brokerage with 30 million users and trillions in assets. So the line "China can't make a Stripe" even sounds a little arrogant: in scale and experience of payment infrastructure, maybe it's Stripe that should learn from China.

Let me offer one more analogy for the other side: Brazil's Google is still Google, but Brazil's "JPMorgan" is Nubank—in many markets, local players don't just fail to build the original; they slam the original onto the beach. China's payments are exactly this "the local version is stronger" specimen.

But back to the verdict: "can't make a Stripe" was never asking "is there a payment company."

It's asking: is there that one layer—the neutral, enterprise-facing infrastructure grown on enterprise willingness to pay and on an M&A/exit ecosystem.

China's payments really are stronger, but that's the consumer-facing wallet gateway (Alipay, WeChat)—the consumer-side faucet; whereas Stripe is the B-side pipe, neutral infrastructure a developer can call with a few lines of code. A country can have the best faucet in the world and still not have a single independent water utility—because the pipe layer is filled up by two platforms and the state team.

Someone will say: don't Alipay and WeChat do acquiring and settlement for merchants too? They do, but that's the faucet's built-in catch basin—rates squeezed below 0.6%, clearing routed through NetsUnion, and it only ever serves its own closed loop. The "function" of acquiring, China has in abundance; the "neutral-infrastructure business" of acquiring has never existed for a single day. Nothing illustrates the point better than Ant's own choice: it packed the two businesses most like Stripe—Antom (global merchant acquiring) and WorldFirst (cross-border collections)—into Ant International, headquartered in Singapore and independently operated since 2024, doing entirely offshore business.

More crucially: China's most successful financial-infrastructure innovations—Futu, Airwallex—without exception chose their battlefield offshore, even actively keeping their distance from "China," in order to grow. Far from being counterexamples, they're the most powerful positive examples: the very part of them that succeeded most is precisely the part they realized only by going outside China's soil.

VII. Closing the Loop: Stripe Isn't a Product, It's the Product of a Soil

Let me draw one line to bind the whole series together.

The unbundling of brokerages and asset managers finished in America's last decade, fees fell to zero, no meat left on the bone; so the smartest money pivoted to attack the banks; but most of the star companies going after the banks' consumer side were cyclical, thin-margined, cash-burning—Chime's IPO halved, Klarna was cut in half, consumer-finance exits met a cold reception; the ones who truly rode out the cycles and made the big money were the shovel sellers—Stripe clawed back from three years to $159 billion, Adyen holding a software-grade 53% profit margin.

And the shovel seller is the product of a soil. It needs enterprises willing to pay for software, a market whose gateways aren't monopolized by super-apps, independent distribution channels, and a road that lets a decade-long run of capital exit with dignity. America has all four; China lacks all four—what's missing isn't talent, it's soil.

So the final answer to "why there's no Chinese Stripe" is: Stripe isn't a product that can be copied; it's the fruit grown by a whole patch of soil. You can find engineers in China smarter than Stripe's, and a payment experience more usable than Stripe's, but you'll struggle to replicate, in China, the soil that lets "neutral infrastructure" grow up independently.

But this isn't a pessimistic ending. Futu and Airwallex have already pointed to a realistic path: since that soil is outside China, choose your battlefield where the soil is richer. Base yourself in Hong Kong, radiate across Southeast Asia, build B2B embedded finance and agentic financial rails, let the super-apps and virtual banks be your distribution, and in the end get absorbed with dignity by a bigger platform—the way Stash sold to Grab. This is an already-proven "shovel-selling" road that belongs to Chinese founders.

Just remember one thing: this road isn't "no big deal"—it's a seven-to-ten-year hard fight with licenses as the moat. Whether you can endure it is no longer a contest of product; it's a contest of capital, of timing, of whether you're willing to move a little closer to that richer soil. Wherever the soil is, that's where people must move—this is a hurdle no one can clear for you.

That is everything I set out to say in this series.

(Postscript: two pieces of unsolicited advice)

The series has said what it needed to; allow me to step outside the topic and add two pieces of personal advice—both about "choices."

First, about the industry: In the AI era, if you're not a founder with a CS/engineering background, I'd actually urge you toward fintech—or into a fintech company. What AI does is often understood backwards: it lowers the barrier to writing code—now anyone can get a model to bolt something together; but it hasn't lowered the barrier to starting a company, it's raised it. The barrier drops, people pour in, everyone's using the same batch of models, and what comes out is highly homogeneous, so competition only gets fiercer, and many arenas end up entirely as the giants' domain. AI itself is the most typical example: the real excess profit settles in the model layer, the compute layer, the chip layer, the underlying infra—all engineering-led hard fights (isn't that just "selling shovels," only swapped for AI's shovels?); whereas the alpha at the application layer will most likely be ground thin all the way down—the barrier keeps dropping, keeps getting more crowded; this week OpenAI throws out a new flagship, next week Anthropic and Google each follow, and you can't even keep up chasing; underneath it all lies the same batch of models, the only difference left is engineering implementation, and the moat breaks at a poke. Finance is exactly the reverse: its underlying logic has barely changed in a thousand years—credit, clearing, asset allocation, interest-rate transmission, risk management, from ancient money houses to modern Stripe, it solves the same set of problems; and the solution to that set of problems is not technology, but switching costs and trust—once it becomes a standard, the alpha is enormous and can be eaten for decades. I have to add a line to connect with the first three parts, lest you misread me: I'm not saying fintech doesn't need engineering—Stripe and Adyen are themselves top-tier engineering organizations; I'm saying that fintech's real moat isn't at the code layer, it's at the layer of trust, licenses, regulatory communication, business model, and structural design—and that layer is precisely where people with business and economics backgrounds have the edge over the purely technical. Engineering will be equalized by AI; trust and licenses won't.

Second, about foresight: Most people's judgment of "where to go" lags reality by several years. I did recruiting for a few years, and one painful observation is: college students with strategic foresight are exceedingly rare. To put it harshly, the student cohort can hardly be said to have "foresight" at all—early on, job-hunting relies on referrals within small circles of high-flyers, and which circle you're in determines where you go more than how the industry is actually doing. The result is that students' read on the market generally lags reality by 2–3 years: all the way to 2019, a whole crowd still thought Baidu was far better than Toutiao; all the way to 2021–2022, they still held that Tencent and Alibaba were far stronger than ByteDance; all the way to 2024, the gaming world still repeated "if there's Goose, pick Goose; if not, pick Mihoyo" (if there's Tencent, go to Tencent; if not, go to Mihoyo); all the way to 2023–2024, the sharpest AI upstarts—Zhipu, Kimi—still couldn't recruit the strongest campus hires. Behind this gap, half is the schools' constraints: internship time is capped, the bar for arranging your own internship is absurdly high; schools are full of people managing job placement, many of whom have never truly held a job, let alone gauged the real temperature of the market; everything leans toward "safe," while forgetting that college students are adults to begin with—and it's often the "disobedient" ones who have more foresight, yet schools happen to prefer the "obedient." How does a person's understanding change fastest? By seeing it with your own eyes—ideally after reality has "slammed" into you once. The information gap lags 2–3 years; the values gap lags 5 to 10.

These two pieces of personal advice are really the same thing: see where the soil is, then move there—a little earlier, a little earlier than everyone else. This, too, is the answer I gave myself after finishing this whole series.

(Note: this piece uses the latest available 2026 figures: the 13x software-spend figure is from Shixiang citing Battery Ventures (2022), China's SaaS share from IDC; the proposed penalty amounts for Futu/Tiger are from the CSRC's advance notice of 2026-05-22 and financial-media disclosures; the Stash→Grab $425 million acquisition is from Grab's investor announcement and SEC filings; the Brex→Capital One $5.15 billion, Cred→Meta $900 million investment, Yu'e Bao's scale, and Alipay/WeChat Pay shares are all from public media reports and company disclosures; Airwallex's license count and the "China backdoor" controversy are synthesized from public reporting, some of it allegation rather than settled fact. Data caliber for Chinese firms' ARPU, SME lifespan, and the like varies across sources and serves only as an order-of-magnitude reference. Progress on Japanese, Hong Kong, and South Korean local-currency stablecoins and the "99% pegged to the dollar" figure are from public reporting (the HKMA Stablecoins Ordinance took effect in 2025-08, etc.); the closing "personal advice" judgments about student perception and the recruiting market are the author's own professional observations, not statistical data.)