The record behind the move is worth reading closely. G Squared launched in 2012, grew to more than 70 people, and added three agencies over the past four years, per Campaign Brief. That is a company built by accumulation, then tidied by structure — a sequence many startups in services businesses will recognise. For related coverage, see Klarna's IPO Filing Shows a Smaller Company Than the $45 Billion Peak: What the Document Says.

For founders weighing the same step, the Mercantile Co. announcement offers four concrete, checkable moves. None of them require a transaction. All of them require deciding something.

What does the Mercantile Co. structure actually look like?

Before extracting lessons, get the facts straight. Mercantile Co. comprises four businesses, each with a defined lane:

• G Squared — digital performance across paid media, SEO, performance creative and data, led by George Photios.
• Burrow — a 24/7 social media monitoring, listening and community management agency, launched in 2022, staffed by an Australian team, and led by CEO Manuel Kambos, who joined in 2025.
• Wordie — designs, builds and supports WordPress solutions for organisations.
• GammaDX — an AI-native digital experience agency building platforms and commerce at scale, launched in January this year, led by George Pappas.

That per-agency detail matters for the lessons below, because each one rests on a documented choice the founders made, not on general agency theory.

Move one: name a leader for every business — including the ones you founded

The first move is the least glamorous and possibly the most load-bearing. Mercantile Co. did not leave leadership ambiguous. Photios leads G Squared. Pappas leads GammaDX. Kambos, who joined Burrow as CEO in 2025, heads that agency. Every business in the group has one named person accountable for it.

Founders consolidating multiple businesses often resist this step, because naming a leader for an agency they personally ran means giving something up. The record here shows the opposite pattern: the founders kept leadership of two businesses between themselves and installed a hired CEO at the third. The lesson is not "step back" — it is that every entity needs exactly one accountable name, and that name does not have to be a founder.

Move two: define the specialism before you need the structure

Each Mercantile Co. agency has a one-line specialism you could recite from memory: performance, social monitoring, WordPress, AI-native experience. The founders framed the parent company as giving "each greater clarity around its specialist offering," according to Campaign Brief.

The practical test for any operator: can you state each business unit's discipline in one sentence without overlap? If two units both claim "digital," the structure is not ready. Mercantile Co.'s four lanes — performance, reputation, technology, experience, as the group describes its coverage — are deliberately non-overlapping. Founders can run this test today, on a whiteboard, before any legal restructuring.

Move three: sell the group, not the merger, to clients

The founders' stated client pitch is worth quoting exactly. Pappas said the group structure lets them "bring the right Mercantile Co. agencies together, combining their capabilities to solve the problem without compromising what makes each agency distinctive." Photios framed it from the buyer's side: "Clients don't need agencies trying to be everything to everyone, now more than ever they need deep expertise."

That is two claims founders can borrow directly: specialism as the default, cross-group assembly as the option. The client list — including Endeavour Energy, P&G, Bunnings, the University of Sydney, Monash University, Mitsubishi Electric, the NRL and Mirvac, among others, per Campaign Brief — suggests the agencies already sold at enterprise scale individually. The structure formalises cross-selling that presumably already happened informally.

Move four: time the formalisation to a milestone, not a crisis

The announcement lands as G Squared approaches its fifteenth year, and Photios tied the timing to it: "As we approach our fifteenth year the opportunity is much bigger." The group had four years of acquisitions behind it before formalising.

The sequencing is the instructive part. The businesses grew first, developed "their own identities, client bases and areas of expertise" — the founders' words — and only then got a parent company. Founders who restructure first and figure out identity later tend to produce structure in search of a reason. Here the structure followed the operating reality. A useful rule from the record: formalise when the units have distinct clients and distinct offers, not when the org chart feels messy.

What has to be true for this to work

A sceptical read, since announcements write themselves favourably. The Mercantile Co. model depends on three things the record asserts but cannot yet prove: that the four agencies genuinely keep their independence in practice, that cross-group referrals actually flow, and that the specialisms stay distinct as AI-native work like GammaDX's overlaps with G Squared's performance offering. GammaDX launched in January this year; whether it and the performance agency stay in separate lanes is the fault line to watch. This connects to our earlier piece, How IPO Pricing and Lock-Up Periods Actually Work.

None of that is a criticism — it is the checklist. Founders copying this structure should write down, today, how they will know in twelve months whether their version kept the lanes clean. The founders themselves said the point was "building sharper expertise and a clearer role in our clients' success," per Campaign Brief. Sharpness is measurable. That is the part worth copying.

Why it matters: Klarna is the test case for whether the 2021 fintech cohort can clear public markets at prices private investors paid three years earlier, and the filing is the first full look at the underlying business.

What does the filing actually show?

Revenue of about $1 billion for 2023, up 25 percent from 2022's roughly $800 million, per the filing, with the company reporting its first annual profit in that document after years of losses — net income helped by aggressive cost cuts, including a workforce that fell from over 5,000 employees to roughly 3,800 across 2022-2023, per figures the company has disclosed. Credit losses fell as the lender tightened underwriting through the 2022 rate shock.

The filing also shows the AI-leverage narrative the company has pushed publicly: marketing cost per employee and support-cost figures that management attributes to automation, per the company's own statements — company-claimed operational metrics, labeled as such, not independently verified here.

How did the valuation get to $45.6 billion and back?

Old-fashioned momentum. Klarna raised at $5.5 billion in September 2020, then at $10.6 billion in February 2021, then $45.6 billion in June 2021 — a valuation that more than quadrupled in nine months on the back of pandemic-era e-commerce volumes and zero rates, per company announcements from the period. By 2022, rising rates blew out the credit-loss model and a down round to $6.7 billion was widely reported in mid-2022, per Bloomberg and Reuters coverage at the time — an 85 percent decline inside a year, the sharpest private fintech reprice of that cycle.

The comparison worth holding onto: the 2024 IPO filing's revenue run-rate supports a single-digit billions valuation at the multiples public consumer-finance companies trade at, not $45.6 billion. Where the IPO prices, not where it files, settles the question.

Why list in New York and not Stockholm?

The filing chooses a US exchange over the company's home market, following the pattern set by Arm, and the stated reasons are the usual ones — deeper pools of sector-specific capital and analyst coverage — per the filing. The Swedish investment community's public grumbling about home-market listings leaving for New York is a recurring subplot in European tech coverage; the practical answer is that US exchanges have taken the large European tech listings of this cycle.

What remains open before pricing?

The listing range, which the filing does not yet set. The state of consumer credit in the US and Sweden, where the filing reports most of its book. Regulatory exposure — buy-now-pay-later products face active rulemaking in both markets, with UK regulation of the sector legislated and in progress as of 2024. And the shareholder overhang from a 2021-vintage cap table marked at prices public buyers have shown no appetite for. The filing establishes a profitable, growing, cost-cut lender; whether that equals a 2021 valuation is a question the market answers at pricing, and this desk does not predict it.

What actually happens before a stock starts trading?

A company that wants to sell shares to the public first files a registration statement with the SEC — typically Form S-1 — which contains the prospectus, the disclosure document laying out the company's business, financials, and risk factors, Investor.gov says. The SEC staff reviews the filing for completeness of disclosure. That review is not an endorsement: "the SEC's declaration of effectiveness does not represent an approval of the merits of the IPO," the bulletin states.

Only after the registration statement is declared effective can shares actually be sold and trading begin.

Who sets the IPO price, and how?

The company and its underwriters — the investment banks running the deal — set the offering price together, and Investor.gov describes it as a mix of "market conditions, analysis and negotiation," not a formula. Underwriters gather "indications of interest" from prospective institutional buyers during a roadshow and use that demand data to recommend where to price the deal.

The interests in the room are not aligned. A company generally wants a higher price, because it raises more capital for the same number of shares sold. Underwriters often favor a price attractive enough to generate strong first-day demand for the clients they're allocating shares to — and because underwriter compensation is typically a percentage of the amount raised, Investor.gov notes the arrangement carries a built-in conflict of interest.

The bulletin is direct about what that pricing exercise does and doesn't tell investors: "the offering price may bear little relationship to the trading price of the securities" once the stock opens.

Why don't individual investors get IPO shares at the offering price?

Because the allocation happens before the stock trades publicly, and underwriters decide who's on the list. Investor.gov states that the bulk of shares in a typical offering go to "institutional and high net-worth clients, such as mutual funds, hedge funds, pension funds, insurance companies." Retail investors without an existing relationship to an underwriting bank overwhelmingly end up buying — if they buy at all — once the stock is already trading on the open market, at whatever price the first session sets.

That timing gap matters. The people setting the initial price are, in most cases, not the same people buying at the opening bell.

What is a lock-up period, and why does it exist?

A lock-up agreement is a contractual restriction — arranged by the underwriters — that bars a company's existing shareholders, including founders, employees, and pre-IPO investors, from selling their shares for a set stretch after the offering, commonly around 180 days, per Investor.gov.

The mechanism exists to manage supply. If every early shareholder could sell on day one, the stock would face a wave of selling pressure right when trading volume and public information about the company are both thin. Restricting early holders from cashing out is meant to let a market for the stock establish itself first.

What happens when the lock-up expires?

The restriction lifts, and shareholders who were barred from selling become free to do so. Investor.gov frames the moment plainly: "the lock-up expirations give these early investors the opportunity to sell their shares to the extent they weren't able to do so as selling shareholders." That's it — the bulletin does not claim the stock necessarily falls, only that a new pool of sellers becomes eligible to enter the market at that point. What any individual stock does around that date depends on company-specific news, broader market conditions, and how many locked-up holders actually choose to sell, none of which the mechanism itself predicts.

Why do freshly public stocks swing so much in the early weeks?

Trading volume right after an IPO is often thin relative to the company's total share count, since most shares are still held by locked-up insiders and institutional allocees rather than freely traded in the market. Investor.gov warns that this "limited trading volume ... can operate to drive the trading price of an issue steeply up" — and that underwriters sometimes provide temporary price support in the earliest sessions that can end abruptly, contributing to sharp moves once it does.

The bulletin's own framing of the category is blunt: "IPOs can be risky and speculative investments." Every company's prospectus includes a risk-factor section — management's own list of threats to the business — that Investor.gov treats as required reading before the mechanics of pricing or allocation matter at all.

The mechanism, not the pick

None of this tells you whether any specific offering is priced right, whether a particular lock-up expiration will move a particular stock, or whether an IPO is a good buy — those are calls this desk does not make. What the public record does establish is the sequence: registration and disclosure, underwriter-negotiated pricing built on institutional demand rather than public trading, an allocation process that routes most shares to institutional and high-net-worth buyers first, and a lock-up period that defers — but does not eliminate — the supply of shares held by insiders. The gap between the negotiated offering price and whatever the market does afterward is the space where all of the "pop" and "crash" headlines live.

Frequently Asked Questions

Does the SEC set or approve the IPO price?

No. The SEC reviews a company's registration statement for disclosure completeness, not for whether the price or the investment is sound — Investor.gov states explicitly that effectiveness is not an approval of the offering's merits.

How long is a typical IPO lock-up period?

Investor.gov describes roughly 180 days as the common length, though the exact term is set by contract between the company and its underwriters and can vary by deal.

Can retail investors buy shares at the IPO price?

Rarely, without an existing brokerage relationship tied to the underwriting banks. Most retail buying happens after the stock is already trading publicly, per Investor.gov.

Does a stock always drop when its lock-up expires?

Not necessarily. The lock-up expiration makes previously restricted shareholders eligible to sell; whether they do, and what that does to the price, depends on company-specific and market conditions the mechanism itself does not dictate.

Where can investors find the risk factors for a specific IPO?

In the company's prospectus, filed as part of its registration statement with the SEC — Investor.gov identifies the risk-factor section as required reading before investing.

Why does the discount exist?

The explanations with the strongest evidence, in the order the literature settled them. Winner's curse: IPO allocations are asymmetric — informed institutions get the good deals disproportionately, so the average buyer needs a discount to stay in a market where they systematically receive the leftovers; underpricing is the compensation that keeps uninformed order flow in the book. Information production: the book is built on investors' private research about demand; the discount pays them for producing that information honestly. Insurance and litigation: underpriced deals do not get sued — the pop is partial protection against liability for the filing's optimism. And signaling: leaving money on the table signals confidence about future raises — companies planning to return to the market rationally underprice more. No single theory wins outright; each explains a documented fragment, and the discount persists because every fragment pays someone.

Who actually pays for the pop?

The arithmetic's unambiguous part: the pop is paid by selling shareholders — the company and the insiders who sold at the offer price — and collected by allocation recipients. Circle's 2025 debut moved roughly $3 billion of value from what the issuer's side received to what allocated institutions held by the closing bell. Figma's roughly $19 billion pricing versus its doubled debut mark told the same story at software scale. The subtlety: 'paying' is only knowable in hindsight — the counterfactual is not 'priced at the closing level' but a smaller, less certain book; underpricing buys demand quality, which is why the practice survived six decades of founders complaining about it.

Why do some pops run extreme?

The documented correlates of extreme first-day returns: scarcity — small floats in hot categories (the 2020-2021 SPAC-and-tech windows averaged extreme pops precisely because supply was thin and retail access broad); sentiment windows — the well-documented correlation of underpricing with bull markets and hot-issue periods; and asymmetric information — deals where valuation is genuinely uncertain (new categories: crypto infrastructure in 2025, AI infrastructure ahead) clear wider because the book is softer. The pattern that repeats across cycles: the average discount is stable, but its variance is a market thermometer — the 1999-2000 and 2020-2021 peaks in average first-day returns both marked cycle tops in new issuance within quarters.

Can a company avoid leaving money on the table?

The documented attempts and their outcomes. Auction IPOs — Google's 2004 modified Dutch auction being the landmark — narrow the discount but do not eliminate it, and the format has stayed niche: distribution is the banks' value-add, and auctions trade certainty for a narrower spread in ways issuers have mostly declined. Direct listings solve the pop by not pricing one — the opening auction clears everything, and no money is 'left' because no allocation was sold — at the price of raising no capital and losing the underwriters' stabilization. The practical middle: the greenshoe and aftermarket support manage the week; honest range-setting — pricing at or above a raised range, as the 2025 class mostly did — recovers some value while keeping the book real. The six-decade verdict stands: some discount is the market's toll for first-day liquidity, and the founders who fight it entirely fight the machine that makes their listing trade.

What should founders and readers take from it?

Founders: budget the pop as a cost of the raise, like the gross spread — the negotiation is about its size, not its existence; price with the book's quality, not against the hype, because the 2021 vintage's lesson — pricing into euphoria — is repriced within quarters. Readers of debut headlines: the pop measures the deal's scarcity and the market's mood, not the company's worth — the closing price on day one is the first honest estimate, and even it will move. And everyone: the discount that looks like banker carelessness is a half-century-old equilibrium that pays the investors, protects the deal, and funds the lawsuits that never come — the machine's quiet lubricant, printed in every listing's first candle.

What did the filing actually show?

The F-1 documented one of the sector's most dramatic operating turnarounds. The 2024 numbers: revenue of roughly $2.8 billion, up about 24 percent; net income reported for the first half of 2025 following a full-year 2024 loss near $85 million — compressed from 2023's roughly $244 million loss; 93 million active consumers and 675,000 merchant partners. The cost story was the document's spine: Klarna's headcount fell from over 5,500 at the 2022 peak toward roughly 3,000, with AI deployed across customer service — the company's publicly stated framing, that AI now does work that would have required hundreds more staff, made its filing a reference text in the AI-and-jobs debate. The disclosures also printed the risk register plainly: consumer-credit cyclicality, regulatory fragmentation across markets, and the interest-rate sensitivity of its funding model.

How does the price compare with the private marks?

The honest ladder: $45.6 billion at the 2021 peak; $6.7 billion at the July 2022 raise — the vintage's emblematic markdown; and roughly $15 billion at the 2025 IPO. Below the peak by two-thirds, above the trough by more than double — the geometric mean of everything the company learned. The pricing sequence itself: an initial range of $37 to $41, an intra-roadshow moment when reports suggested the range might be cut, then pricing at $40 with a first-day close modestly above. A disciplined debut by the standards of the window — no euphoria, no break, a real company clearing a real market.

Why the NYSE and why then?

The documented logic: the U.S. listing chose Klarna's largest growth market — American BNPL volume was the company's fastest-growing line while European maturity set in — and the September timing followed the window's strong summer (Circle, Chime, Figma had all priced) with a backlog of European fintech watching Klarna as the test case. The filing choice itself — an F-1 as a foreign private issuer — kept disclosure requirements lighter than a domestic S-1, a decision analysts noted as deliberate.

What did the debut prove for the sector?

Three documented takeaways. That the turnaround template works: growth companies that cut deeply and rebuilt profitability accessed the public market — Klarna's path from $244 million losses to profitability is the cycle's cleanest case study. That AI's operating story is now an equity story: Klarna's AI-driven cost structure was priced into the deal explicitly, the first major listing where the AI-labor narrative appeared in the investor materials as a core thesis rather than a footnote. And that private marks resolve downward: the $45.6 billion 2021 price was a zero-rate artifact; the $15 billion public price is a clearing price — and the gap between those two numbers is the 2021 vintage's lesson stated once more, in F-1 typography.

What are the open questions?

The record's honest list: credit performance through a full consumer downturn — the losses in a real recession, not a modeled one; the funding-cost cycle as rates fall; competition from Affirm in the U.S. and bank-offered installment products everywhere; and the regulatory consolidation of BNPL rules across the U.S. and EU, moving from improvisation to code. Each is printed in the filing, and each is now the public market's to price quarterly.

Klarna's debut was the fintech cycle's closing argument: the 2021 peak was a mirage, the 2022 trough was overdone, and the public market — unglamorously, correctly — priced the company at what the record supports.

What actually opens a window?

Four documented preconditions, roughly in order. Rates: the cost of capital sets the discount rate on every future cash flow — the 2020-2021 window opened at zero rates, and the 2025 opening tracked the Federal Reserve's easing cycle from late 2024. Comparables: newly priced deals trade against public peers, and a software IPO cannot price when public software trades at depressed multiples — the window for a category opens when its comparables recover, which is why windows differ by sector in the same macro. A successful bellwether: the first big deal of a window sets the risk appetite of every institutional buyer — one strong debut (a 2025 Circle or Figma) unlocks the calendar behind it, because allocators who missed the pop chase the next one. And supply-demand balance: windows close when the queue of issuers exhausts institutional cash — the 2021 experience, where a record year ended in broken deals and December cancellations as supply overwhelmed demand.

What slams them shut?

The shutdown triggers are faster than the opening ones. Macro shocks: a market correction above roughly 10 percent closes risk appetite in weeks — the March 2020 freeze and the 2022 tightening are the documented textbook cases, with the 2022 closure triggered by rate increases repricing growth stocks. The broken deal: one high-profile failed pricing — a deal cut, postponed, or trading below offer in week one — makes every subsequent issuer's buyer more cautious; underwriters pull deals preemptively to protect the calendar. And volatility itself: VIX above the mid-20s is the market's practical 'no-visibility' zone, because pricing a book requires a stable tape. The asymmetry matters for planning: a window opens over quarters and slams in days.

How should a company time against a window?

The documented discipline of the companies that actually got out: file when the window is merely visible, not open — the S-1 process takes a quarter, and the filing queue is the window's waiting room, with the SEC confidential-submission path letting companies stage filings without exposure. Price when the comparables and the bellwether align, and accept that perfection is unavailable — the 2025 class priced at good-not-great multiples relative to 2021, and every one of them is glad it did. And prepare the company, not just the filing: the audit, the public-company hires, the quiet-period discipline — the documented pattern is that readiness, not market genius, separates the listed from the stranded in every window; the 2021-2022 cohort that filed early in 2021 mostly got out, and the equally qualified cohort still preparing when the window shut waited three years.

What does the current cycle's record show?

The 2025 reopening's documented shape: quality-first — the market priced profitable or near-profitable companies (Circle's reserve income, Chime's first profit, Figma's retention) and punished concepts; dispersion after the pop — the window rewarded retention metrics with durable trading and reverted fintech multiples toward bank comparables within quarters; and a heavy pipeline staged behind it, with the mega-private companies — Stripe, Databricks, Canva, the AI labs themselves in some form — positioned as the window's second act, each one's filing decision partly a function of the last one's aftermarket. The window's second-year test, historically, is whether the second act prices as well as the first: the 1999 and 2021 precedents both failed it; 2026's verdict was still being written.

Windows are market weather: not predictable in detail, but seasonal in pattern, observable in real time, and survivable only with preparation done in advance. The companies that list are almost never the ones that timed the window perfectly — they are the ones ready when it opened.

Where are the real numbers?

Start with the selected financial data and the full statements: revenue, gross margin, operating loss, and — decisive for the modern IPO — the reconciliation of any company-defined metric to GAAP. The core red flags in the numbers: revenue growth without gross-profit growth (pricing power eroding while the top line buys itself); 'adjusted' profitability that excludes stock-based compensation at software-scale magnitude (a real, recurring cost — when SBC exceeds 20-30 percent of revenue, the adjustment is the story); and customer concentration above roughly 20 percent in a single account, which the risk factors will disclose and the roadshow will not mention. Figma's S-1 was the 2025 masterclass in reading carefully: the headline net loss was an accounting artifact of the terminated Adobe deal's one-time equity charges, visible only in the notes.

What do the growth metrics actually measure?

The Metrics section is where definitions do the work. The flags: metrics defined so loosely they cannot fail — 'engaged users' where engagement is undefined; ARR that includes contracted-but-unstarted amounts; net retention above 130 percent that on inspection mixes expansion accounting with a shrinking base. The craft is reading each metric's definition and asking what a dishonest company would do to inflate it — the answer is always printed in the definition. When a company changes a metric's definition year-over-year, the change is disclosed in the notes; find it, because the reason for the change is usually the story.

What do the risk factors confess?

Risk factors are boilerplate by design — but the specific ones are confessions. The flags: a single named customer, supplier, or platform dependency stated as a risk (Chime's banking-partner structure; any startup dependent on one cloud or model provider); regulatory risk naming a pending rule that would restructure the business (Circle's S-1 leaned on stablecoin legislation that arrived weeks after listing); insider-control language — dual-class shares, voting agreements, founder control provisions — which tells you governance is not yours to influence; and the litigation section, which lists the suits the company must disclose and hopes you weigh lightly. The craft: risk factors are written by lawyers paid to disclose without alarming; the ones with operational detail are the ones management actually fears.

What does the cap table and insider section reveal?

The ownership and related-party sections answer who this IPO is for. The flags: selling stockholders outnumbering primary proceeds — a liquidity event for insiders rather than capital for the company; pre-IPO investors with liquidation preferences stacked above the offer price, which the notes will reveal and which means common holders' economics are worse than the headline; option pools sized aggressively with the pool-shuffle mechanics loading dilution onto post-IPO holders; and related-party transactions — the company buying from the founder's other venture, executives' interests aligned in ways the summary omits. In AI-era filings, add one: compute commitments and supplier concentration — the S-1s of the model-dependent companies disclose multi-year cloud obligations that are, in substance, the company's real balance sheet.

What should a reader's thirty-minute routine be?

The routine the evidence suggests: financial statements first — revenue, gross margin, operating loss, cash, SBC; then the metric definitions and any changes; then the concentration and related-party notes; then the specific risk factors with numbers in them; and only then the narrative at the front, now inoculated against it. Compare against the last private round: the disclosed valuation history versus the IPO price is the honest record of the private market's pricing (Chime's $25 billion private mark against an $11 billion IPO was printed in its own documents). And read the underwriters' lockup and greenshoe terms, which tell you the supply calendar before you ever look at a chart.

Every S-1 is two documents: the story the company tells and the arithmetic the accountants certified. The second one is always available, always at the back, and always the one that ages well.

What is actually in a lockup agreement?

The standard terms: a 180-day lock for company insiders and pre-IPO holders; earlier release triggers, increasingly common, that free some or all shares if the stock trades above a threshold (typically 120-130 percent of the IPO price) for a set window after earnings; and staged releases for early investors in some deals. Underwriters want the lockup — unrestricted insider selling in month one would break the aftermarket they stabilized — and companies accept it because underwriters price certainty. The document is boilerplate; its calendar consequences are not.

What does the documented pattern around expiry look like?

The academic and market record, consistent across decades of studies: lockup expirations are associated with abnormal trading volume — several times normal — and modest average negative returns in the surrounding weeks, with the drift beginning before the date as markets anticipate. The average effect is small; the variance is enormous. The determining variables, per the same literature: how much insider equity is unlocking relative to float, whether the stock is above or below the IPO price, and — the dominant factor — whether the company's post-IPO fundamentals gave insiders a reason to hold. Expired lockups on well-performing stocks routinely pass without damage; lockups on broken stocks cluster with the selling everyone expects.

Why do companies stage or release lockups early?

The 2021 vintage innovated aggressively, and the record explains why: with hundreds of IPOs competing for attention, companies used early-release triggers to reward employees locked out of liquidity while prices ran, and to defuse the day-180 cliff by spreading sales. The documented failures of the same era cut the other way — releases that dumped supply into weak aftermarkets accelerated declines, and several companies that waived lockups in 2021 watched their stocks never recover. The lesson the market drew: staged releases are a tool for strong aftermarkets, not a rescue for weak ones.

What happened at recent expirations of note?

The 2025 class provided the test cases. The consistent documented pattern across the reopened window's big listings — including the fintech and crypto debuts: elevated volume at expiry, pressured prices in the weeks around it for stocks trading below their offer, and non-events for the winners. The largest structural factor, documented in deal documents for every recent listing: IPO floats remain small — often under 15 percent of shares outstanding — so the expiry can release several multiples of the float at once. Small float flattered the debut; the same small float concentrates the expiry risk. The two events are the same mechanism read twice.

What should founders and employees plan for?

Founders: negotiate lockup terms at IPO with the same care as pricing — staged releases, employee carve-outs for tax obligations, and communication planning for the expiry week, because a silent calendar date becomes a narrative event when the volume spikes. Employees: the honest planning frame is that lockup expiry is when your options become spendable, and the documented pattern — supply pressure into the date — argues for a pre-decided selling plan rather than a week-of decision made while watching the tape. Both should remember what the data shows: the expiry amplifies the underlying story; it does not write a new one.

Day 180 is the IPO's second act: the day the market learns whether the people who know the company best are buyers of the story they sold. The calendar prints the date months ahead — which is exactly why the drift starts early and the preparation should too.

What did the S-1 show?

The documented profile: roughly 8.6 million active members (as of Q1 2025, company-defined), 2024 revenue of about $1.67 billion, up roughly 32 percent from 2023, and net income of about $25 million for 2024 — first full-year profit after a $203 million loss in 2023. The revenue engine is interchange: Chime makes money when members swipe its debit card, a model dependent on theDurbin-exempt small-bank partner structure and on member transaction volume. Growth had slowed from fintech's 2021 peak — the filing's own numbers show the user base and revenue growing but decelerating — and valuation context was the sore spot: Chime's last private round in 2021 valued it at $25 billion, meaning the IPO priced at well under half its private mark, the largest documented private-to-public markdown of the 2025 class.

How did the pricing sequence run?

The initial range of $24 to $26 had already been lifted from early expectations closer to $19, and the deal priced at $27 with the first-day close near $40. The pop was real but the aftermath told the story: within weeks the stock settled back toward the $30s as the market repriced it against consumer-bank economics — deposit costs, interchange sensitivity, and a rate cycle turning against deposit-funded models. The pattern matched the window's other fintech debuts: warm welcome, then bank multiple.

What does Chime's model look like post-IPO?

The documented strategy in the S-1 and roadshow: deepen revenue per member — Chime's reported revenue per active member was near $250 and rising — via products beyond interchange, including an earned-wage-access product (MyPay) that grew fast through 2025, plus secured credit building toward a fuller product shelf. The risk register the filing prints: reliance on two banking partners for the charter structure, interchange regulation as a perennial Congressional topic, and competition from Cash App, PayPal's debit programs, and every incumbent's digital arm. Nothing in the post-debut record changed those facts; the market simply priced them.

What did the listing prove about the window?

Three documented takeaways for the 2025 class. That the window opened for profitability, not concepts: Chime listed on its first audited profitable year, as Circle had on reserve income — the market's price of admission. That private marks still don't clear: Chime's $25 billion 2021 round versus an ~$11 billion IPO printed the lesson the 2021 vintage had learned — late-stage private pricing was a different asset class from public equity. And that category labels fade: 'fintech' priced like finance within a quarter, exactly as the neobank skeptics predicted and exactly as Klarna's September debut would confirm again.

What are the open questions on the record?

Whether MyPay's growth compensates for interchange cyclicality at scale — its credit performance through a full rate cycle is unproven in public data. Whether the member count monetizes up toward the $300-plus revenue-per-member that would justify growth-stock pricing. And whether the banking-partner structure survives regulatory attention — a listed company is a more visible target than a private one. Each is printed in the filing's risk factors, and each is priced into a multiple that sits between a bank's and a software company's, which is where the honest reading of Chime has always lived.

Chime's debut proved the window works and the repricing is real: a profitable neobank, a warm first day, and a bank multiple by autumn. For founders, the lesson is arithmetic — your Series E price and your IPO price are answers to different questions.

What are the actual cost differences?

Gross spreads — the underwriters' fee on an IPO — run about 7 percent at deal sizes below $250 million and 3.5 to 5.5 percent on large tech IPOs, per filings and market data. On a $500 million raise, that is $20 to 25 million of fees, plus the underpricing itself: the first-day pop is a transfer to allocated institutions, frequently larger than the cash fee. A direct listing pays exchange listing fees, a financial adviser (typically a bank working for a flat retainer in the single-digit millions), and legal costs — materially cheaper, at the price of raising no money. Some issuers split the difference: Spotify and Slack listed directly in 2018 and 2019 raising nothing; the NYSE then won approval, and the SEC signed off, for direct listings with a primary raise — used rarely since, with the 2021 era's attempts (e.g., Squarespace considered; Coinbase listed via direct listing in 2021) spanning both forms.

What are the mechanics of each path?

In an IPO, the syndicate builds a book, prices the deal overnight, and allocates shares — the issuer knows its proceeds when it signs. In a direct listing, a market maker runs an opening auction from collected buy and sell interest; the opening price is where those curves cross, and early trading can be violent — Spotify's 2018 debut, the category's proof case, opened up sharply and settled; Coinbase's 2021 debut swung tens of billions in market cap within hours because no book had anchored placement. No lockup is required in a direct listing, though companies adopt voluntary ones — meaning employees and insiders can sell from day one, transferring the supply risk the IPO's lockup defers onto the opening auction.

Who has actually chosen each path?

The documented cases. Direct listings: Spotify (2018), Slack (2019), Palantir and Asana (2020), Coinbase (2021), and thereafter a trickle — the form has been rare since 2021's rate reset, because the companies for which it fits — cash-rich, brand-known, no primary need — have mostly been absorbed by private markets instead. IPOs: essentially everyone else, including the cash-rich — Stripe's contemplated listing is universally reported as a traditional IPO because employee liquidity at scale needs the underwriting machinery. The record's quiet conclusion: direct listing is a niche instrument for a specific company profile, not a movement.

How do the two paths compare on control and signaling?

Control: a direct listing involves no lead-left bank with aftermarket influence and no allocation politics; an IPO installs a stabilization agent whose research coverage and market-making matter for years — the lead bank is a long-term hire, not a toll. Signaling: pricing an IPO above a raised range is a marketing event a company can plan; a direct listing's open price is whatever the market says, with no narrative control on the day. For companies whose story needs explaining, the roadshow is not a cost — it is the product tour. For companies whose brand precedes them (Spotify, Coinbase), skipping it saved money and revealed price.

Which companies should choose which?

The decision matrix on the documented record. Choose a direct listing if: the company needs no primary capital; the brand is consumer-known enough that buyers arrive without a roadshow; insider supply is manageable and transparent; and volatility on debut is tolerable. Choose an IPO if: the company is raising money; institutional placement and aftermarket support matter; the story requires education; and certainty of proceeds — the priced deal — is worth the 7 percent and the pop. The 2025 window's evidence is unambiguous: Circle, Chime, Figma, Klarna — every major listing of the reopened window chose the underwritten path, because each needed cash or certainty or both, and the direct-listing alternative remained what it has been since 2021: available, cheap, and almost never chosen.

The toll booth is expensive, but it delivers passengers. The open road is free when your destination needs no cargo — and almost nobody's does.

What did the S-1 actually disclose?

The filing told a growth story with an AI asterisk. Documented figures from the S-1: 2024 revenue of about $749 million, up roughly 48 percent from 2023; a net loss of about $732 million for 2024 — but driven overwhelmingly by one-time stock-based compensation charges triggered by the terminated Adobe deal's repurchase of shares at the acquisition price, which mechanically re-priced employee equity; over 90 percent gross margins; net dollar retention reported at 120 percent or above; and roughly 830 customers paying more than $100,000 annually. Strip the one-time charges and the operating picture is a near-breakeven, high-retention software company — the distinction every analyst note on the filing made within a day.

Why did Adobe's failed acquisition matter to the listing?

The killed deal set the IPO's economics in three documented ways. First, the $1 billion termination fee Adobe paid funded the balance sheet. Second, the repurchase of vested employee options at the $20 billion deal price cashed out years of equity, resetting the internal wealth baseline and arguably removing urgency to go public — the company waited through a closed window instead of rushing. Third, the regulatory block that killed the acquisition — UK CMA and EU Commission concern over Adobe's dominance in interactive design — became Figma's competitive talking point as an independent: the file-format center of gravity for a whole design ecosystem, expanding toward whiteboard and AI tooling.

How did the pricing sequence run?

Textbook compression. Initial range of $28 to $30; demand during the roadshow led the company to raise the range to $32 to $34; the deal priced at $33; the stock opened materially higher and roughly doubled from pricing by the close. For the issuer, that first-day move was money on the table — but the strategic logic dominated the coverage: Figma needed no cash, having the Adobe fee and positive operating economics, and the listing was as much about employee liquidity, index inclusion, and currency for acquisitions as about proceeds.

What does the listing mean for the software IPO window?

Figma was the marquee software debut of the 2025 reopening, alongside Circle in June, Chime in June, and Klarna in September — the busiest listing calendar since 2021. The documented pattern across the class: price above range, pop on debut, then dispersion — the market rewarding retention and growth quality rather than category labels. Figma's post-debut trading held up materially better than the fintech names, consistent with its S-1 metrics: software gross margins and 120 percent net dollar retention against consumer-banking economics.

What are the open questions on the record?

Three, all printed in the filing's risk factors. AI disruption: generative design tools compress parts of Figma's value proposition, and the company's own AI features must monetize without cannibalizing seats. Concentration: the enterprise customer count is strong but the design-professional TAM is finite, and the expansion into developers and marketers is unproven at scale. Lockup expiry: the post-IPO share overhang from early employees and the Adobe-era repurchase mechanics arrives on schedule, and the float's behavior around it is unknowable in advance. Each is a listed risk, not a hidden one — the S-1 is unusually candid, which is itself part of the record.

Figma's debut validated the patient path: blocked from selling for $20 billion, it built the standalone public-company case for eighteen months and priced within reach of the dead deal's number on day one. The window reopened; the dispersion started immediately after.

What happens between the filing and the range?

The S-1 filing goes to the SEC, which comments; the company amends. Weeks later, the underwriter syndicate — the banks led by one or two 'lead-left' bookrunners named on the filing's cover — publishes a price range, typically 20 to 25 percent below what insiders expect to achieve. The range is an anchor, not an appraisal: it is set low deliberately, because raising the range during the roadshow reads as momentum while cutting it reads as failure. Figma's July 2025 process was the textbook case: an initial range of $28 to $30 was raised to $32 to $34 mid-roadshow, and the deal still priced at $33 — a sequence designed to generate the upgrade headlines it generated.

What is the roadshow actually for?

For about ten days, management presents the same 30-to-60-minute pitch to institutional investors across cities, and the banks' sales desks collect indications of interest: how many shares, at what price, from whom. This is book-building. Indications are soft commitments — investors can reduce or walk — but by the end, the lead bookrunner knows roughly how many times oversubscribed the deal is at each price. A book that is five times covered at the top of the range gives the banks confidence to price at or above it; a book that barely covers the deal forces a price inside or below the range, or a postponement.

How is the final price actually set?

On pricing night, the lead underwriter brings a recommendation to a call with the company's board and executives, and the number is negotiated and approved — a decision made in hours, on the basis of the book's shape, comparable-company trading that day, and the general market tone for the next morning. The company signs the underwriting agreement, and the shares are sold to the institutions at that price. The money the company and selling stockholders receive is fixed at this point. Everything that happens on the exchange afterward — the pop, the fade, the volume — redistributes value among other people.

Why does the stock open somewhere else entirely?

The opening price on the exchange is set by an auction among buyers who did not get allocation or want more, run by the exchange's designated market maker with the lead banker's guidance. If the book was heavily oversubscribed and allocation was scarce, the first trade can gap far above the IPO price — Circle's 168 percent first-day close in 2025 being the extreme recent example. That gap is called money on the table: shares sold by the company at $31 that the market valued at $83 the same day. It is not a malfunction. It is the fee, in the form of a cheap allocation transferred to institutions, that companies pay for the underwriters' distribution certainty.

Who wins and who loses in the process?

The map of interests: the company wants maximum proceeds and a stable after-market; underwriters want a successful deal, a client for the aftermarket, and happy institutional buyers who get the discount embedded in underpricing; institutions want the allocation and the pop; selling insiders want price; employees with options want any price above their strikes. These interests conflict directly, and the pricing night decision balances them imperfectly. Founders should also understand greenshoe stabilization: underwriters typically sell 15 percent more shares than the base deal, with an option to cover the over-allocation at the IPO price — a mechanism that lets them support the stock in early trading without naked risk, and one that quietly enlarges the deal when the stock works.

What should a founder take from the machinery?

Three practical notes. Choose the lead-left bank for the aftermarket commitment and research coverage, not only the valuation promise — the range is negotiable, the six months after listing are the bank's real job. Model dilution and proceeds at the midpoint, not the top of the range, because pricing below range happens to good companies in bad weeks. And treat the first-day pop as a marketing event, not value creation: the money on the table left at Circle's pricing was real, and the beneficiaries were the allocation recipients, not the issuer.

The IPO is not one price but three — the range that anchors, the deal that funds, and the open that trades. Knowing who sets each, and for whom, is the difference between reading a debut and understanding one.

What does Circle actually sell?

Circle's core product is USDC, the second-largest dollar-backed stablecoin, with a circulating supply in the tens of billions as of 2025. The documented business model, laid out in the company's S-1 filing, has an unusual shape: most of the revenue comes from interest on the reserve of Treasury bills backing the tokens — roughly $1.7 billion of total revenue in 2024, per the filing, almost all reserve income, of which the majority was shared with Coinbase under a distribution agreement. That makes Circle's revenue a levered play on interest rates: when the Federal Reserve cut rates in late 2024 and 2025, Circle's per-dollar income fell mechanically.

What did the S-1 disclose that mattered?

Three things. First, the Coinbase dependency: the distribution agreement pays Coinbase most of the reserve yield on USDC held on its platform, a term that surprised many readers of the filing. Second, the regulatory contingency: USDC's status depended on pending stablecoin legislation, which arrived weeks after the listing when the GENIUS Act was signed into law on July 18, 2025 — federal framework rather than state-by-state improvisation. Third, growth: USDC circulation roughly doubled year over year into 2025, from roughly $28 billion to over $60 billion, per company disclosures.

How did the pricing and pop actually go?

The company initially marketed at $24 to $26, raised the range, priced at $31, and watched the stock open near $69 and close the first day around $83. By late June the shares had touched $100-plus before settling; the float then behaved like the crypto-cycle proxy it is, swinging with bitcoin and with stablecoin headlines rather than with payments-peer comparables. An IPO pop of that size is standard underpricing dynamics at the extreme — money left on the table by sellers, captured by allocations — and it says more about scarcity of crypto equity exposure than about steady-state valuation.

What does the listing mean for the stablecoin market?

Circle's listing gave the stablecoin sector its first public financial statements, and the disclosures moved policy: legislators negotiating the GENIUS Act could price the reserve-disclosure and licensing requirements against a live public company rather than hypotheticals. The Act ultimately required issuers to hold high-quality liquid reserves and publish monthly attestations — close to what Circle already did, an alignment skeptics noted. For startups, the signal was simpler: the IPO window for fintech and crypto-adjacent companies, shut since 2021, was reopening.

What are the open questions on the record?

The rate dependency: every Fed cut compresses Circle's core revenue line, a risk the S-1 states plainly. The Coinbase share: the more USDC grows on Coinbase, the less Circle keeps. And competition: Tether, the larger rival, remains private and offshore with a reported treasury profit sharing no public statements; banks entering issuance post-GENIUS Act could compress fees further. Circle's answer on all three is transaction growth and new products — payments APIs and tokenized funds — where the record so far shows early traction, not proof.

Circle's debut worked as a listing: it priced, popped, and held a multi-billion public valuation through a rate-cutting cycle. Whether it works as a business is a question about interest rates and Coinbase's cut — and both are printed in the S-1 for anyone who reads past the ticker.