An AI Marketing Company Pulled In ₩6.6 Trillion in Deposits — Madup Lists Tomorrow (July 1) at a 3,305:1 Subscription Ratio

A single AI marketing company drew ₩6.6 trillion

Here's the deal: AI marketing firm Madup lists on Korea's KOSDAQ tomorrow, July 1. And the pre-listing retail subscription went wild — a 3,305:1 ratio and ₩6.6 trillion (~$4.8B) in deposit money. By the company's account, that's the highest among new listings in the past four years. Against 500,000 allotted shares, 1.65 billion shares were subscribed, and the offer price was fixed at ₩8,000 — the top of the indicated range. Lead underwriter: Mirae Asset Securities.

By the numbers it's "another hot IPO," but what makes it interesting is what Madup actually is. It's an ad agency that also calls itself an AI agent company. Its in-house AI agent, LEVER Xpert, claims to automate 90%+ of marketing data collection, and that solution's revenue reportedly jumped 285% in a year — from ₩1.7B in 2024 to ₩4.9B in 2025. In other words, the market isn't betting on an "ad agency" but on a "company that automates advertising with AI."

Why does that matter? AI marketing has been drowning in the "we use AI" modifier, but whether it actually makes money — whether it's a business the public markets will validate — was unclear. Madup's hot debut is the moment Korea's capital market first answered that question loudly: at least at the subscription stage, "yes." Whether the post-listing stock price holds up that expectation is an entirely separate test.

So here's what we'll unpack: what Madup does, why LEVER Xpert is its core weapon, what ₩6.6T in deposits means, and what it signals to the whole AI marketing industry. Three players: Madup (and founder/CEO Lee Joo-min), the investors who piled into the subscription, and the competing AI marketing field.

The players — Madup, investors, and the AI marketing market

First, Madup. A company that started in ad performance marketing. It has two core assets: data, and the AI that runs on it. The company touts the ad-performance data accumulated by actually deploying a trillion won of ad spend as its core edge. Why does that matter? An AI model runs on data as fuel, and real-deployment performance data — the kind you can only get by actually spending the money — isn't something just anyone can manufacture. CEO Lee Joo-min has steered the company on a strategy of layering an AI agent on top of this data asset.

Next, the investors who piled in. ₩6.6 trillion in deposits isn't just money — it's the size of expectation. Institutional book-building also drew 2,392 institutions at a 1,396:1 ratio. Institutions and retail rushing in simultaneously means the story landed with both experts and the public. Still, everyone knows IPO subscription heat and post-listing price are different things — a subscription is a signal of "I want in," not a promise to "hold for the long haul."

Third, the AI marketing market as a whole. Marketing is one of the first fields where AI got deployed for real — ad copy generation, targeting optimization, performance analysis, budget allocation. AI is rapidly replacing what humans did by hand. Madup's listing is one of the first major cases of a company foregrounding "AI automation" getting a public-market verdict, so the whole industry is watching it as a gauge of "can we be valued like this too?"

One sentence ties it together: a company that layered an AI agent on top of a trillion-won pile of real ad-spend data (Madup) drew ₩6.6T in deposits from institutions and retail, and the entire AI marketing field is watching the result as a gauge of its own future. That's the skeleton.

What the numbers say about Madup's IPO

Talk scatters, so here are the confirmed facts in a table.

Item	Detail
Listing date	July 1, 2026 (KOSDAQ)
Retail subscription ratio	3,305:1
Subscription deposits	~₩6.6 trillion
Allotted shares	500,000 (1.65B subscribed)
Offer price	₩8,000 (fixed at top of range)
Institutional book-building	2,392 institutions, 1,396:1
Lead underwriter	Mirae Asset Securities
Core AI	LEVER Xpert — automates 90%+ of marketing data collection
Solution revenue	₩1.7B (2024) → ₩4.9B (2025), +285%
Core asset	₩1 trillion of real ad-spend performance data

Line by line. First, the "offer price at the top + hot subscription" combo is the crux. Pricing at the top of the range means it was well-valued in institutional book-building, and a 3,305:1 retail ratio piled on top. So both price (institutional valuation) and popularity (retail demand) came in strong — a rare case. When both run hot together, day-one interest tends to be intense.

Second, the 285% solution-revenue growth is the story's engine. Ad-agency revenue swings with the economy and carries thin margins, but AI solution revenue is closer to software — different margins and scalability. Nearly 3x in a year, ₩1.7B to ₩4.9B, reads as evidence the company is shifting its center of gravity from "ad agency" to "AI software company." What investors bet on is the speed of that shift.

Third, the ₩1-trillion data pile is the real moat. AI itself is something anyone can grab a similar model for. Differentiation ultimately comes from data, and performance data you "can only get by actually spending a trillion won" is hard for a latecomer to replicate quickly. Madup's bet rests on the logic that "the more common AI becomes, the more the side with proprietary data to feed it wins." Whether that moat is real is the key to its long-term value.

Who gets what

Start with Madup's win. First, capital: pricing at the top of the range secured plenty of ammunition for expansion — funds for AI solution upgrades, data infrastructure, and overseas entry. Second, a trust asset: KOSDAQ-listed status raises credibility when dealing with large advertisers. Third, talent: listed-company shares and stock options are a powerful card for recruiting scarce AI engineers — and people are exactly what AI companies lack most.

Investors' expected win is clear too. First, a growth story: AI solution revenue growing 285% lets you imagine "room to grow further." Second, scarcity: pure AI marketing players are rare on the public market, so money wanting exposure to the sector funneled into Madup. Third, a data moat: a hard-to-replicate data asset gives a sense of "this company won't get pushed aside easily." But note all of this is "expectation" that only becomes real if post-listing results back it up.

On the flip side, there are risks to view coldly. First, a hot subscription isn't a rising stock price — deposits piling in means short-term-profit money is mixed in too, so post-listing volatility can be high. Second, an "AI premium" bubble risk: in a phase where the market pays richly for the word "AI," expectation can outrun results and inflate the valuation. Third, intensifying competition: AI marketing automation has low entry barriers aside from data, so if big tech or large agencies ship similar features, the differentiation can wobble.

Precedents — what worked and what didn't

There are several cases of marketing companies leading with AI and data. The success analogy is "a platform that defended its data moat to the end." Companies that accumulated proprietary data, ran it with AI to deliver better results to advertisers, and built that virtuous cycle grew steadily even through economic swings. The key was a structure where the data got thicker over time. For Madup to walk this path, its trillion-won data has to be a growing asset, not a frozen one.

The failure analogy is "companies whose AI premium drained right after listing." The subscription was a blowout, but results didn't keep up with expectations and the stock fell below its offer price post-listing — not rare. Especially for names where the "AI" label was priced too richly, the correction was harsh when the gap between expectation and reality showed. A hot subscription doesn't guarantee a good investment — that's the lesson from this side.

So the key isn't the "subscription scorecard" but "post-listing results." Whether Madup's vaunted 285% solution-revenue growth continues as a trend rather than a one-off, whether the data moat actually fends off rivals, how fast it reduces ad-agency dependence to become a software company — those three answer within the next few quarters. The hot subscription is only the starting gun; the finish line is far off.

How rivals counter

The first to react will be large ad agencies. Seeing Madup's debut, they may strengthen their own "AI automation solutions" or counter by acquiring similar AI marketing startups. A large agency that already holds massive ad-deployment data, with AI properly layered on, could even exceed Madup on data scale. Madup's moat has to live not just in "volume of data" but in "the know-how to run it with AI" to be defensible.

Big tech ad platforms (Google, Meta, etc.) are a more fundamental threat. They hold the source of ad-deployment data and are increasingly automating targeting and optimization with their own AI. If the trend that "advertisers can just use the platform's AI without a middle solution" strengthens, the reason for a middle layer like Madup to exist could wobble. Madup's counter is to grow the value a single platform can't provide — integrated optimization across multiple platforms.

For other AI marketing startups, Madup's listing is a double-edged sword. On one hand it proves "AI marketing can exit via IPO," pulling capital into the whole sector. On the other, with a "leading stock" now existing, investors start comparing directly against Madup, and latecomers that can't differentiate may get overshadowed. Madup's success or failure becomes the reference point that sways investor sentiment for the entire sector.

So what actually changes

If you're interested in investing, Madup is worth seeing as "a thermometer for the AI marketing sector." The subscription was hot, but the real thing to watch is whether post-listing results back the expectation. Checking the solution-revenue trend, the data moat, and the software-transition speed each quarter is the core — better to look calmly at the business fundamentals than get swept up in the subscription buzz. (For the record, this isn't investment advice — the call is yours.)

If you're a marketer or ad professional, this is a signal of shifting industry terrain. In an era where AI automates 90% of data collection and analysis, the human role moves from "gathering data" to "building strategy and making judgments with that data." Cede routine operations to AI and grow value in the creative, strategic, and relational areas AI can't handle — that's the way to survive. Madup's LEVER Xpert shows that shift is already reality.

If you follow Korea's AI ecosystem, this is a meaningful milestone. Korean AI has mostly drawn attention on the "hard" side — semiconductors, infrastructure, foundation models. Madup is a case where an "applied service that makes real revenue with AI" got a big public-market valuation. It signals AI moving past the lab and the chip fab into earning money in an everyday industry like advertising. It could be an inflection point where the monetization of applied AI kicks into gear in Korea too.

One step further — "hot subscription" and "company value" are different stories

The easiest thing to confuse here is the equation "hot subscription = good company." The two are related but not the same. ₩6.6T in deposits is the size of short-term demand — "this many people want in at this price." Company value, by contrast, is a function of long-term results — "how much this company will earn and return to shareholders." However hot the subscription, if post-listing results cool, the stock falls. So rather than getting excited by debut news, the real analysis is calmly testing whether the numbers the company touted (285% growth, ₩1T data) stay valid going forward.

The other thing to watch closely is the durability of the data moat. Madup's trillion-won data is clearly powerful, but a data moat isn't "build it once and you're done" — it's an asset you have to keep refilling to maintain. If the ad market's rules change (cookie deprecation, tighter privacy rules), if big-tech platforms block data access, or if a rival accumulates data faster, the moat can shallow out. So Madup's long-term value rests not on "the data it has now" but on "the ability to keep growing the data and running it better with AI."

In the end, the question Madup's IPO poses isn't "did the subscription succeed." It's "is AI marketing a sustainable business model, not a passing fad." Korea's capital market, at the subscription stage, bet ₩6.6T on "yes." Whether that bet was right, Madup has to prove over the coming quarters post-listing. July 1 is the starting line of that long proof, not the finish.

🥄 Three Things You're Probably Wondering

— So should I buy this? I can't call that for you — the investment decision is yours. Just remember: a hot subscription and the post-listing stock price are separate things. Checking each quarter whether the solution-revenue trend and the data moat translate into actual results matters far more than the subscription buzz.

— Is LEVER Xpert really that impressive? By the company's claim, it automates 90%+ of marketing data collection. Solution revenue growing 285% in a year is a clear result. But how much "data-collection automation" actually lifts ad performance itself needs more verification. Efficiency of automation and improvement in outcomes are different stories — best confirmed by post-listing numbers.

— AI marketing — isn't this a bubble? Some of it may be a bubble and some of it real — too early to call. It's true we're in a phase where the "AI" label gets richly priced, so there are certainly names where expectation outran results. Whether Madup is bubble or substance comes down to whether the growth it touted continues as a trend. Time will answer.

Sources

• Madup draws ₩6.6T in deposits, 3,305:1 retail subscription ratio — VentureSquare
• Madup posts 3,305:1 retail subscription ratio — Newspim
• Madup IPO subscription day 1 closes — CBC News
• Madup raises ₩6.6T in retail subscription — IT AI Totality
• IPO schedule — Korea Exchange KIND

Numbers and criteria are as of announcement and may change. Investment calls are yours to make!