Original · Unique Research · 2026-04-21
Editor's note: The first-person report and its judgments belong to the original Chinese author. This English rendition retains the opening essay, three narrative sections, and the full panel transcript including the audience question. Product claims, KOL pricing, payback-period figures, market-size claims, and cultural comparisons are source or speaker claims, not independently audited findings. Company, personal and product names are transliterated where official English forms remain unverified. The source is dated April 21, 2026.
Unique Awards
Global-Expansion Payback Dropping Toward 360 Days? Four Insiders Advise You to Abandon the US Market
When everyone is talking about disruption, AGI, and traffic dividends, what truly determines whether you survive is often the payback period, KOL pricing, and whether you found the right market from the start.
"
Once upon a time, finding a wild tech blogger on YouTube to do an AI tool review cost a mere $200. Today, that figure has soared to $10,000, and you still have to pay a $5,000 "queue fee" upfront.
This is the naked, sweat-and-blood truth of current AI full-domain global expansion.
While everyone is shouting about disruption, AGI, and traffic dividends, in this underwater world, user-acquisition costs have been stretched to an extremely distorted extreme. At a hardcore session titled "Full-Domain AI Global Growth," four frontline experts deeply versed in overseas traffic playbooks and monetization—Monica, founder of AGI Villa; Zimu, SaaS growth hacker; Elaine, overseas user consultant; and Jingling, founder of DeerAPI—told a pile of extremely discouraging but unavoidable truths.
If you are about to charge into overseas markets with your code written, this conversation may be useful.
The Unclear Ledger and "Atavistic" Tactics
The earliest batch of people who fully reaped the mobile-internet global-expansion dividends know the concept of "three-month payback" by heart. Over the past few years, a decent global-expansion tool product could roughly recoup its advertising spend in 90 days.
"But now, it's getting closer and closer to the heavy-gaming industry." Jingling unsparingly exposed the bottom line of AI app ad spending. He has worked with a large number of clients doing GenAI computing-infrastructure access, and what everyone talks about most in private is pain: the ad-spend payback period has been stretched to 180 days, 270 days, and in extreme cases even 360 days to recoup capital.
This means that if you don't design how to earn your first dollar on Day 1, your company simply won't survive long enough to reach PMF.
The truth Jingling offered is very brutal: how do you set your initial free tier? How many dollars do you price the Pro version? Set it too high and conversion is poor; set it too low and you can't afford to buy traffic. Under an extremely long payback period, every user drop-off in the funnel is rapidly bleeding cash.
Interestingly, in this track supposedly representing humanity's most cutting-edge technology, user-acquisition tactics have instead shown an "atavistic phenomenon." Zimu discovered that when everyone is crowded onto the single-plank bridge of influencers and paid traffic, some of the "dumbest" methods are instead the most effective. For example, doing substantive content exchanges in specific communities, or deeply cultivating the most traditional SEO. Moreover, because Google Search has added the AI Overview (AI summary) feature, many informational searches have become "earning exposure but not clicks." How to write content that can both be indexed by AI and guide conversion has become an extremely meticulous manual payback exercise.
Avoid the Meat Grinder: The Hidden Cash-Burning Hotspots
When talking about global expansion, everyone's first reaction is often to lock their sights firmly on the US market. But in the view of frontline operators, charging straight into Tier-1 European and American red oceans from the start is undoubtedly hitting a Gundam with a wooden stick.
Elaine gave everyone a very impactful new awareness: shift your gaze away from the US, and you may not be able to imagine who is fanatically paying for AI.
"Two years ago when doing overseas growth, everyone still thought Brazil was an extremely distant lower-tier market." Elaine said. But in this year's actual backend data, they unexpectedly discovered that for various AI SaaS tools, Brazil has extremely counterintuitively become one of the countries with the most paying users outside the US.
Similarly well-funded and in a period of super-infrastructure explosion is the Middle East. Monica just completed deep research in the six Gulf states before Spring Festival. "Local per capita GDP is three times that of China, and the 2030 visions launched by Saudi Arabia and the UAE have poured over $100 billion into supporting the AI ecosystem."
More critically, emerging markets don't need a "mute software" that stiffly translates English into a local minor language. Monica pierced the cultural cognitive gap of AI products: "Americans see AI as a smart friend; the Japanese see it as a polite assistant; but in the Middle East, they more hope it to be an exclusive system that can understand Arabic and their own religion and culture." Whoever can knead this cultural sensibility into the large model can quietly make a fortune in corners others can't see.
"CEOs Who Have Never Raised a Dog Shouldn't Make AI Pets"
If paid traffic has gotten expensive, how exactly should you break through?
These experts who truly manage traffic offered a solution that sounds extremely counterintuitive: founders, please take off that "low-key and restrained" shell and step onto the stage to brag.
This is Founder-led growth, which has become prevalent in Silicon Valley over the past two years.
"The narrative in Europe and America is different from China's past. They have a dream and immediately start telling it—they may still be a hundred thousand miles from making a product, but the story has already begun." Monica has deep experience on this point. She herself spends even just ten-plus minutes a day making overseas short videos telling stories of Chinese founders going global, and can easily hit 100,000 views in a week. This is not only an extremely cheap way to get free traffic, but also the core anchor providing emotional value to early seed users. For C-end applications, this emotional bond is even more effective than buying a few more KOL placements.
In this process, the most tragic are those "pseudo pain points" that go global for the sake of going global, and do AI for the sake of AI.
"There was a founder who wanted to make an AI collar to solve the problem of children sitting too close while doing homework causing myopia. I asked him if anyone on his team had children. He said no. He didn't even know that 'sitting too close while writing' is not among the top three causes of myopia." Monica has coached over a hundred companies in the past two years and can only laugh and cry at this phenomenon. Moderator Ark also added a jab: "There are even people who have never raised a pet wanting to make AR pet companionship through a pile of code."
You don't even know how much pain your target has, so what makes you expect them to pull out their credit card for your algorithm?
Traffic is expensive, stories are priceless. The wild era where translating code was enough to make foreigners pay has completely ended. In today's AI great voyage, you either calculate every last cent of retention in complex spreadsheets, or boldly step into the spotlight as a storyteller who hurts more than anyone and empathizes better than anyone.
More Conversation Details
Unique Awards · Hangzhou AI WEEK Trends Roundtable Panel
"Full-Domain AI Growth: New Logic for Cross-Region, Cross-Market Global Growth"
Guests:
AGI Villa & MonicaChuhai Shuo — Founder — Monica
OPC — Overseas Software Practitioner & Growth Hacker — Zimu
Independent Overseas User Growth Consultant — Elaine Li
DeerAPI — CEO — Jingling
Moderator: EPIC Connector — Core Member — Ark
Ark: Today's topic is mainly about growth and marketing, which is a very hot topic, so let's get started. Guests, please briefly introduce yourselves in order.
Monica: I'm Monica. I have now founded an organization called AGI Villa, which is a Venture Studio for early-stage AI founders. I myself was a serial entrepreneur for over a decade, founding roughly ten projects from B2B e-commerce, social e-commerce, to AI programming youth education. Four projects are still alive, and some have reached unicorn status. I am also a mother of three children, so I have always paid attention to the AI education track. I am now focusing on helping Born Global companies with their global expansion. Everyone can follow my self-media called "MonicaChuhai Shuo," with weekly live streams about global expansion every Wednesday; Zimu has also participated in my live streams before.
Zimu: Hello everyone, I'm Zimu—Zimu is my alias. I've been in the SaaS industry for over a decade, and started doing SaaS global expansion about three years ago. I happened to do SaaS global expansion without AI, and after AI arrived, I did AI-related SaaS global expansion. I come from a technical background and now also write every day, enjoying creating content. Content is the foundation of marketing, so I also know a bit about Google SEO—that's roughly my situation.
Elaine Li: Hello everyone, I'm Elaine Li. I work as an overseas user growth consultant. For the past 10 years I've been working in overseas markets, now mainly doing influencer marketing, KOL, UGC, and also ad placement-related business.
Jingling: Hello everyone, my name is Jingling. I'm the founder of DeerAPI. We provide one-stop access to 500 large models globally, offering a one-stop API aggregation. I myself have done growth for many years. Before founding this company, I always did overseas market growth—ASM, ASO, and Facebook and Google ad placement. So much so that now when I serve clients, although we do GenAI computing infrastructure, many clients keep chatting with me about SaaS and SEO-related knowledge.
Ark: Finally, let me briefly introduce myself. I'm Ark. I am also an AI entrepreneur myself, currently based in both Shanghai and Silicon Valley. We are mainly doing AI digital employees, mainly for e-commerce and Social Media applications, helping automate employees' business-process time. Let's officially begin today. The first question is: over the past year, what do you guests think is the most obvious change in the global growth of AI products? Also, compared to traditional internet SaaS global expansion, what do you think is the biggest difference in the global growth logic of AI products? Starting with Monica in order.
Monica: Regarding this year's global expansion trends, I currently have several judgments. The first is the rise of new markets. I came back from the Middle East before Spring Festival, stayed there for a long time, and did detailed research on the relevant markets of the six Gulf states. Although there is currently some uncertainty, I am still very optimistic about such an emerging market. First, local per capita GDP is three times that of China, overall English proficiency is very high, basically reaching 100%. Including local people's demand for AI products, and at the government level—today we also see Arab countries, such as Saudi Arabia and the UAE, launching their 2030 visions, in which the entire country will invest over $100 billion in AI startup projects over the next three to five years. We see a series of support, including Abu Dhabi establishing the world's first AI-specialized university, whose leading core professor was also poached from CMU, with very many Chinese students studying for master's and doctoral degrees there. We see that their level of emphasis for AI talent also shocked me greatly. So whether it's the Middle East, Southeast Asia, or Latin America, these are very important enterprise global-expansion trends I see over the next two to three years. The second trend is that Chinese global expansion has shifted from past commodity exports, to brand exports that everyone talked about in the past two years, and this year the key is founder IP exports. This also resonates deeply with me—I started doing my own overseas IP not long ago. Every day I may only invest about 10 to 20 minutes, with no other cost input, and in a week I can achieve 100,000 views, simply telling stories of Chinese founders. Here we see that overseas ordinary people have a very large information gap regarding Chinese manufacturing and Chinese tech companies going global, which is why we see such an opportunity. China has always talked about Chinese people liking low-key and moderation, always saying you should have a presentable product before telling your story. But you will find that this narrative is completely different from Europe and America. In Europe and America, they often say that once you have a vision, a dream, you start telling it. They may still be some time from making the product, but they have already started doing narrative and storytelling. This is also what I believe is very key going forward—everyone should step onto the stage to tell their vision.
Ark: What Monica said resonates with me too—Founder-led growth, where the founder themselves steps out to tell stories and resonate with users. Next, Zimu.
Zimu: Because I'm an introvert, I prepared what to say. I'll cover the first and second topics together. First, on the first topic, what Monica just said was excellent. I'll only talk about AI SaaS global expansion, which I've always liked to follow—because I have a habit of checking every month which AI tools grew fast and had high revenue last month. For this new change, I think it changes very fast, and the timing of product selection is increasingly important. Back then, doing images and video, you just had to rush in; now video remains the track with the highest ceiling. Recently, around OpenClaw doing application global expansion, you can do foolproof one-click deployment, you can do Agent-hiring platforms, or you can do AI payments—as long as you have the capability, these are all very hot directions, but it depends on whether your capabilities match. The second point: I think traffic costs will get higher and higher. Early on, doing an AI YouTube review blogger cost only $200; now the quote is $10,000, and they also require paying $5,000 upfront to queue. Overall traffic is still declining, but you have to invest because they are the top AI review bloggers. Another point is that traffic entry points are also changing. Early traffic had SEO, which everyone knows well—do good SEO and you'll have a certain stable volume. Now it's not that everyone immediately invests fully in ads and influencers, but more refined. For example, if your ICP audience is in a certain community, you can cooperate with that community, or go to the community to produce content for exchange. The SEO track I follow long-term, in 2024 and 2025, comprehensively added an AI overview. Everyone who uses Google Search will know that there is an AI summary at the top. This AI summary has caused many informational search pages to have completely no clicks, only exposure. Everyone has also been discussing how AI creates SEO content to drive traffic for various products. This is a big project, because you first need to understand your product and your audience before you can create good content that also fits the SEO direction. Regarding the biggest difference in global growth logic between traditional global expansion and AI products, I understand the core is all about finding the right customers, the right scenarios, the right channels, and wherever users are and which channel suits them, you go to that channel to do growth. Although channels are changing, essentially there's no big difference. If there really is a difference, it's actually that the AI track now changes too fast. Everyone is getting more homogeneous—early on someone made an AI digital human, later made it into All-in-one; others did the video track, also made it into All-in-one, homogeneity is getting more and more severe. In this case, the biggest difference is seeing who can faster capture user value. For example, when new models and new scenario demands like OpenClaw come out, whoever can get into position first can eat the first wave of dividends. This means the track is more competitive—I understand it's roughly these two aspects.
Ark: Just like the 4Ps we once learned in college, or the people-goods-scene of e-commerce, now traffic entry points are also changing—everyone needs to better judge what channel to start from and better do GTM. Next, Elaine.
Elaine Li: Let me first echo what Zimu just mentioned—cooperation prices are getting higher and higher, which is quite obvious on my end. Every day I'm negotiating various cooperations, influencer prices are rising, and traffic prices are also rising particularly fast. Secondly, I find that domestic global expansion, especially for AI SaaS tool products, has an increasingly thorough understanding of the Brazilian market. Two years ago when doing it, everyone might have thought the Brazilian market was a very distant lower-tier market. But now we find that, especially for AI SaaS, everyone's understanding of the Brazilian market is that it is likely one of the countries with the most paying users outside the US, and everyone is also willing to understand this market. These are the two points I feel quite obviously.
Ark: Sometimes when we talk about markets, we either act like there's only the US or China in the world, but actually that's not the case. There are some overlooked high-value markets, such as Japan, South Korea, Hong Kong, even Singapore. And the Middle East that Monica mentioned earlier. It specifically depends on what kind of user your product targets, before deciding what market to hit. Finally, Jingling.
Jingling: I particularly agree with what Brother Zimu said—subjectively, the current KOL costs are rising in a straight line. In terms of performance ads, it's obvious that the payback period for Facebook and Twitter placements is getting longer and longer. Our clients' subjective feeling is that initially, a product a few years ago might have had 90-day payback as decent; now it's getting closer to the gaming industry, reaching 180 to 270-day payback. In extreme cases it might become 360-day payback. This further requires clients' products to have their data in retention, conversion, and all aspects iterated thoroughly, to support the upfront traffic-buying costs. The second subjective feeling is that now it feels like all promotion methods have returned to the PC-era tactics of 20 years ago—this is also the biggest subjective feeling of the AI era. In the mobile era before, we witnessed the great wave of gaming global expansion; the seven regions mentioned by everyone, Latin America, North Africa, the Middle East, etc., were very large and high-return markets in the mobile era. But now the tactics have returned to PC-era logic, and the cost structure and promotion models have changed very significantly.
Ark: As Jingling mentioned earlier, now whether KOL or paid traffic, these are huge costs. This is also why more Founders choose to step out, standing on stage like Monica doing Founder-led growth. In fact this is relatively cost-effective. For example, a Korean guy made an AI note-taking software, and the Founder himself stepped out to tell the Story, providing empathy and emotional value—sometimes more effective than paid traffic or hiring KOLs outside. Monica seems to want to add something?
Monica: I want to add, because I was just asked two questions and only answered the first—the second should be the difference between SaaS products and AI products. I myself have indeed done both SaaS products and AI products. There is a very big difference here: in the past, Chinese-made software products, when placed in different markets, often just needed a language version translation to generate natural paying users. But for AI products, you find this trick no longer works. The biggest problem here is that people of different races, countries, and cultures have different understandings and cognitions of AI products. Regarding artificial intelligence, for example, the Japanese more see AI as a polite assistant. Americans may more see AI as a smart friend. But in the Middle East, they more believe AI is an individual that understands my religion and my culture. So you will find AI products have a huge opportunity in the Middle East, which is to do Arabic localization. Arabic is different from other languages—it is written from right to left. So the entire user's understanding logic is different, and at this time whoever's AI product can better understand Arabic large models will have very good opportunities locally.
Ark: This also resonates deeply with me. In the past few years, the AI entrepreneur ecosystem often first made a Product, holding a hammer looking for a nail, taking the product and trying to sell it, feeling like forced buying and selling. In fact you should first understand users, find real needs, and conversely develop the Product—that is a smoother process. The next question is: if an AI startup team starts doing global expansion (GTM), what do you think is the first priority to do well? What are the reasons many Growth Hackers or Startups fall into growth bottlenecks? Starting with Monica in order.
Monica: What to do on Day 1, I still believe two things are most important. The first is still to find your first 100 paying users in the most suitable market you choose. This is far more important than doing ad placement or optimizing the product—find your 100 seed community users. The second thing is still what I just talked about: do founder IP, step onto the stage to tell stories. Regarding growth bottlenecks encountered, because I have coached over 100 companies in the past two years, I found the most common have several problems. I summarized a PPA model: the first P is Passion, the second P is Pain point, and A is AI Leverage. Passion means whether the product the founder is making is based on personal passion, or a field where you have very strong pain points. We talk about the ten-thousand-hour rule—you must be willing to put in 10,000 hours to polish and explore. Let me give an example I once coached: there was a founder who wanted to make a hardware product solving the pain points of runners, but he himself never ran. I think from the underlying logic it simply doesn't hold. The second is Pain point—whether the product you are making is a real pain point, a real demand? This may have been talked about many years ago, but in the AI era this problem is getting more and more severe. Another example: a classmate told me he wanted to make a hardware product solving children's myopia—a neck ring worn around the neck that can monitor how close the child is to the desk while doing homework, and if too close will emit an alarm. It sounds very make sense, but does everyone know what the top three causes of children's myopia are? First is excessive eye-use time, second is insufficient outdoor time, third is genetic inheritance. The top three causes have nothing to do with sitting too close to the desk while doing homework. I asked him a question: does anyone on your entire founding team have children? He said no. This is a deviation in pain-point cognition. I have always believed that a good founder must find pain points from the scenario you are in, and deeply understand the primary audience. The third is AI Leverage. Many people think adding a GPT interface to an existing product makes it an AI product—definitely not like that. The most important thing about AI is that on the basis of your existing scenario pain points, it improves efficiency by at least 10 times, or can provide a completely different experience. The hardware product that impressed me most last year was a product called Sticker Box. A child says to this red box that they want a butterfly with bull horns, and within a minute it prints out such a sticker, which the child can then color on. This is the completely different logic and experience that AI brings to children's painting creation. Products like this we think represent the innovative products that the AI era will produce.
Ark: When Monica just talked about that myopia neck ring, it reminded me that there was another team doing AR pets before. I asked them a question: what pets are you all raising now? They actually told me they don't raise pets. So very confused—they simply have no idea about pets, yet they are doing this thing, very inconsistent. How can they have enough empathy and sympathy to think of their users? If you can't stand in the user's perspective to think about this pain point, you won't know how much pain the user actually has, and naturally can't understand why users only use you for free and don't pay you. Next, Zimu.
Zimu: I also echo—Sticker Box this product I love too much. My child really likes coloring books, but going to the store to choose there are only those few. This is the era AI brings—he directly tells AI what he wants to draw today, it prints it out, colors it and can stick it wherever he wants. This is a very good product AI brought, very hot half a year ago and raised a lot of money. Back to this topic, I will first look at the overall country. Everyone likes to choose Tier 1 countries because payment awareness is relatively strong. But in the past year and a half everyone has been staring at the US, and doing early product validation while staring at the US market to compete will be a bit difficult, because the track may already be particularly competitive. Actually there are still many Tier 2, Tier 3 countries, such as Japan, South Korea, Europe, Latin America, etc., also with various demands. Choosing a suitable market for early validation would be better. Additionally, for example, if you are doing AI-version tools for Shopify independent sites, you should choose countries with rich seller ecosystems and many sellers. You should focus on considering where your customers are, where I am. To start doing global growth, the goal Monica just mentioned is very good, and I can break the goal into three points. First, first clarify your product's ICP. ICP is not thought up, it's tested out—you can roughly determine ICP through competitor analysis and user interviews. You can list the Top 3 audience groups, such as AI self-media, content creators, etc. Second, pick a Top audience group for rapid validation, choosing only one most suitable growth channel. For example, the fastest SEM search ads, which are relatively precise and can show user search intent; or find an AI self-media training camp to cooperate with, where everyone is your ICP, to validate whether your product can make them pay. To validate whether it's a problem of a half-finished product, or the product is already finished but people just won't pay. Finally, in the process do well the product's value and experience to validate, to see whether this is truly a real demand.
Ark: The answer just now was very spot-on, hit the nail on the head. Ultimately it's four words "user demand," or three words "real demand." You can quickly judge through interviews, or launching an MVP: first whether users use it, second whether users pay after using it, third how much users pay. Through these three soul-searching questions you can torture out whether this is truly a real demand. Next, Elaine.
Elaine Li: I also think user research and user interviews are the most important first step. Now I see many Founders don't like to do it themselves, because this is indeed a relatively tiring task. But especially early on, I particularly recommend every Founder to personally chat one by one with early users. Only by personally chatting can you dig out those very detailed pain points and demands; finding other classmates to help chat, they may not be able to help you chat to that level. This is my most obvious recent insight. Also customer service is very important—after you have users come in, the feedback from customer service particularly needs everyone to pay extra attention. Many Founders feel customer service can just find someone to help watch, but the feedback coming in through this channel is also early user feedback, and even some very subtle opinions are worth our deep digging.
Ark: User feedback after ad placement, including sending questionnaires, interviews, or data after product launch—users basically vote with their feet. Next, Jingling.
Jingling: I think in the first stage of global expansion, the clichéd PMF we won't talk about—everyone goes from 0 to 1 first to get PMF done. Here I'd like to add that pricing is actually also very important at this stage. You have to think about how to earn your first dollar from Day 1. If making AI hardware, should pricing go high-end, low-end, or mid-range? Good hardware means you have a clear business model from the start. If making SaaS products, how should your initial Freemium free tier be set? How should the second-tier Pro package and later Enterprise package be priced? This directly determines conversion rate—set too high and conversion is relatively poor, set too low and you can't afford to buy traffic. So in the early stage, I think pricing strategy should be established in advance.
Ark: Pricing is also the headache point for many SaaS. Set too low doesn't work, too high doesn't work either, so before that you need to do some interviews with users. I've also been following a word recently called "user premium"—how high a price is he willing to pay for your thing. I've always wanted to find the point where we can provide higher added value. Similarly everyone is connecting large model APIs, but we are not earning the Token spread fee—we provide added value, making applications and SaaS on top of the underlying API, truly solving user pain points that are painful enough and rigid enough, then they will be willing to pay a higher price. Our prescribed questions for today end here. Because this is the last Panel, those who have stayed until now are all warriors, and also truly love Growth and Marketing. I want to ask if anyone in the audience wants to ask the guests questions? If so, we can satisfy everyone's opportunity to ask questions.
Audience: Hello four teachers, thank you very much, we are all carefully listening to everyone's very professional sharing. I want to ask a question: although China and the US are currently the leading players in the entire AI track, where do you think our gap is? Especially in computing costs for running programs etc., Chinese institutions or individuals may have a slight gap. What aspects might widen the gap? How should we catch up with this gap?
Ark: Let's see which guest wants to answer this question.
Monica: Let me throw out a brick first. The gap between China and the US in AI—because I stayed in Silicon Valley for quite a long time last year, I have some feelings. First let's talk about the gap: what the US does better, first is capital. The entire AI venture-capital atmosphere in Silicon Valley is definitely much better than domestically, whether in the participation of private capital, or capital activeness and flexibility—it is indeed much better, this has to be admired. But on one hand talking about the gap, on the other hand for Chinese enterprises, I think an important direction for the future is to actively embrace international capital. In the past we saw more dollar funds, but next I think there are many opportunities, whether European capital or Middle Eastern capital, must also be very actively embraced. Especially these relatively neutral regional funds will be very helpful for future Chinese enterprises to resist geopolitical risks. Another aspect is talent. The US has the top engineers—although there are quite a lot of Chinese people in OpenAI, still a large portion are white engineers. But China's strongest is actually medium and upper-medium talent. When I gave speeches in Silicon Valley introducing China's WaytoAGI, DataWhale, they were very shocked. I said China now has 3 million college students using AI and learning AI—this thing is unimaginable in the US. So China's AI in the next few years has very strong staying power. So where is a gap now? This year everyone is talking about C-end product global expansion, so many college students coming out want to do C-end, but the biggest problem is scenarios. They are too far from scenarios. So this year I have also partnered with various cities' industries, hoping to dig out scenario pain points within industries, then plug in AIGC technical capabilities, acting as a bridge inside, truly promoting the landing of innovation.
Ark: What Monica just said resonates deeply with me. The US has a better entrepreneurial ecosystem than domestically; in acquisitions, the dream of many US SaaS companies is not IPO, but being sold to Salesforce or Microsoft, and investors also have clearer and faster exit paths. In this way, the entrepreneurial ecosystem and financing ecosystem also become more prosperous.