Bending Spoons (BSP): A Genuine Danaher-of-Consumer-Software Playbook, Reportedly Seeking ~$20B — Initiating at Hold, Fair Value $13-16B

Filing Snapshot

Bending Spoons S.p.A. — the Milan-based serial acquirer of digital businesses — filed its public F-1 on June 8, 2026, to list ordinary shares on the Nasdaq Global Select Market under the symbol BSP. The offering combines primary shares with a secondary component from existing holders; the price range, share counts, and founder voting percentages are blank in this first public filing and will be set in a pricing amendment. Goldman Sachs International, J.P. Morgan, and Allen & Company lead the book, supported by ten additional bookrunners spanning US bulge bracket and the Italian banking system (BofA, Wells Fargo, Jefferies, Evercore ISI, BNP Paribas, Mizuho, Société Générale, Crédit Agricole CIB, IMI–Intesa Sanpaolo, UniCredit).

The financial profile is unlike any consumer-software IPO in recent memory, because Bending Spoons is not organically a consumer-software company — it is a capital-allocation machine that buys neglected digital products and re-engineers their economics. Reported revenue grew from $387 million (2023) to $671 million (2024) to $1.306 billion (2025), an 84% CAGR, of which the overwhelming majority was acquired: organic growth was 7% in 2024, 13% in 2025, and 6% in Q1 2026. Q1 2026 revenue of $601 million grew 132% year-over-year. On a pro forma basis — treating the November 2025 Vimeo, January 2026 AOL, and March 2026 Eventbrite acquisitions as if owned from January 1, 2025 — FY2025 revenue was $2.608 billion, with pro forma operating income of $509 million.

Revenue Assessment

The reported growth rates are an artifact of acquisition timing and tell you almost nothing about the underlying engine; the three numbers that matter are 7%, 13%, and 6% — organic growth in 2024, 2025, and Q1 2026, respectively. Management is explicit that this is by design. The founder letter states the company favors "returns over organic revenue growth" and underwrites every deal "assuming we'll never sell the acquired business." The portfolio compounds through capital deployment, not through unit growth: businesses generating 100% of Q1 2024 revenue contributed only 24% of Q1 2026 revenue — despite growing in absolute terms — because $4.6 billion of new acquisitions (Vimeo $1.36B, AOL $1.45B, Eventbrite $505M, Tractive $900M, plus six smaller deals) landed on top of them in eighteen months. Subscription revenue is 84% of the Q1 2026 mix (93% in FY2025 before AOL's advertising business arrived), advertising 12%, other 4%. Roughly 65% of revenue is North American, and no customer exceeds 1% of revenue.

Profitability Assessment

Underwriting profitability requires choosing which income statement to believe. GAAP operating margin has hovered around 19-22% for three years; the company's Adjusted Operating Income margin — which strips amortization of acquired intangibles, transaction costs, reorganization costs, and "other items" — reached 47% in FY2025 and 51% in Q1 2026. The truth is in between and closer to the adjusted figure than skeptics will concede, but the gap deserves scrutiny because the add-backs are not incidental to the model: reorganization expense ($13.5M → $51.8M → $78.6M across 2023-25, and $75.8M in Q1 2026 alone) is the recurring cost of the playbook itself — every new acquisition triggers severance and restructuring, and the company intends to acquire indefinitely. We treat amortization of acquired intangibles (a non-cash purchase-accounting artifact) as a legitimate add-back and the reorganization/transaction costs as a semi-recurring cost of doing business, putting "steady-state" operating margin in the low-40s — still exceptional for consumer software.

Cash and Balance Sheet Assessment

The balance sheet is the most leveraged of any 2026 technology IPO: total debt principal of $4.48 billion at March 31, 2026 (up from $2.71 billion at year-end 2025) against $741 million of cash, with Term Loan B tranches priced at SOFR + 587.5bp (Euribor + 525-587.5bp for euro tranches) maturing March 2031 — leveraged-loan pricing, not investment-grade. Q1 2026 interest expense of $93.2 million annualizes to ~$373 million, an ~8.5% blended cost of debt. Against annualized Q1 Adjusted Operating Income of $1.23 billion, net leverage runs ~3.0x with ~3.3x interest coverage — manageable but tight enough that the IPO is functionally a deleveraging and re-arming event: proceeds are earmarked for "general corporate purposes and to invest in new acquisitions." Operating cash flow ($291M FY2025) badly trails Adjusted Net Income ($376M) because cash interest, cash taxes, and restructuring payments are all real even when adjusted earnings exclude them — a 77% conversion ratio that drops further once the $166 million Italian tax-transfer bill is considered. Goodwill plus acquired intangibles total $5.45 billion, 78% of total assets.

Offering Structure & Governance

Governance Reality Check

Three structural features deserve more attention than they will get on the roadshow. First, the dual-class structure: class A shares carry 5 votes against 1 for the ordinary shares sold to the public, and all class A stock sits with the four co-founders. The F-1 leaves the post-IPO voting percentage blank, but the design intent is unambiguous — public holders will own economics, not influence. Second, foreign-private-issuer status: BSP will not file 10-Qs (expect semi-annual or voluntary quarterly reporting), is exempt from US proxy rules and Regulation FD, and intends to follow Italian home-country governance rather than Nasdaq's independence requirements. For a company whose model depends on rapid, large, debt-funded acquisitions, the reduced disclosure cadence materially raises the cost of monitoring the thesis. Third, the underwriter web: Allen & Company is simultaneously a shareholder, a board seat (Leah Schwartz), a historical M&A advisor, and a lead underwriter; board member Paola Tagliavini is deputy chair of Intesa Sanpaolo, which is a lender, security agent under the credit agreement, and an underwriter. None of this is improper — it is disclosed and Rule 5121 is satisfied via J.P. Morgan as qualified independent underwriter — but investors should understand that nearly every adviser at the table has an economic stake in this deal pricing high.

Assessment: The governance package is aggressive but coherent: founders who consider the company a permanent capital vehicle ("assuming we'll never sell") are building a structure that lets them run it that way. The correct response is not to avoid the stock but to price the control discount explicitly — we apply one in the valuation section — and to size positions knowing that minority holders cannot course-correct capital allocation if the playbook degrades.

Company Overview

Business Model

Bending Spoons owns and operates a portfolio of roughly twenty digital businesses serving over 500 million monthly active users and more than 9 million monthly paying customers as of March 2026 (both metrics aggregated across products without de-duplication — treat them as scale indicators, not distinct-user counts). The portfolio's trajectory tells the story: 111 million MAU and 3 million payers in December 2023 became 389 million and 8 million by December 2025, then crossed 500 million and 9 million three months later — almost entirely through acquisition. The company does not build products; it buys established ones whose prior owners under-monetized them, re-engineers cost structure and pricing, and redeploys the cash into the next purchase.

Revenue is overwhelmingly subscription (84% in Q1 2026), sold direct and through the Apple App Store and Google Play, with advertising (12% — mostly AOL) and other revenue (4% — mostly Eventbrite ticketing fees) completing the mix. The operating chassis is a shared "Platform": 621 core employees ("Spooners," 27% of 2,284 total FTEs), proprietary internal tooling (including Minerva, an AI lifetime-value prediction system built in 2019), and centralized functions — talent, infrastructure, monetization experimentation — deployed across every acquired business.

Revenue by Business (Pro Forma FY2025)

The F-1 reports one segment and never provides a per-business revenue table — the single most consequential disclosure gap in the filing. The pro forma statements and acquisition notes allow a partial reconstruction:

Within the legacy block, the F-1 discloses only entry-point contributions (Evernote $95M in 2023; WeTransfer $61M in five months of 2024, implying ~$146M annualized; Meetup/StreamYard/Issuu $92M partial-year 2024; Loomly/komoot/MileIQ/Harvest $119M partial-year 2025). Tractive (closed May 18, 2026, $900M total consideration) appears in no financial statement — its revenue is undisclosed.

Revenue by Type

The mix shift is acquisition-driven, not strategic drift: AOL brought advertising from 5% to 12% of revenue in one quarter, and Eventbrite's ticketing fees built the "other" line from nothing. Each new business arrives with its revenue model intact; the portfolio mix is the weighted sum of what has been bought. Investors should expect the subscription share to keep falling as the company diversifies into marketplaces (Eventbrite), media (AOL), and hardware-attached subscriptions (Tractive) — a quality consideration, since subscription revenue carries the highest visibility.

Revenue by Geography

An Italian company whose home market is 1% of revenue: BSP earns 62% of revenue in the US (up from 49% in FY2025 — the AOL effect) and roughly 65% in North America, with billing in dollars against a partly euro-denominated Milan cost base. The geographic profile is the practical argument for the Nasdaq listing — and it is also why CFIUS treats this Italian issuer as a recurring counterparty (see Key Topics #9) and why FX swings ($34M net loss in 2025) will keep visiting the other-income line.

The Businesses: Cluster-by-Cluster Analysis

The Cash Engine — AOL (~24% of PF revenue)

A thirty-year-old consumer franchise (email, news portal, search distribution) bought from the Yahoo/Apollo complex for $1.45 billion — 2.3x revenue and 4.3x its $334 million of FY2025 carve-out operating income. AOL competes, nominally, with Gmail and free news portals; in practice it competes with its own users' mortality and inertia, and the 95% average net revenue retention disclosed for it is extraordinary for an asset universally written off two decades ago. Its Q1 2026 contribution ($141.8M revenue, $65.5M pre-tax income) covers most of the consolidated interest bill on its own.

Our view: the portfolio's best deal and its least repeatable. The decline curve is real (the user base ages out; the dial-up annuity shrinks), advertising adds cyclicality, and the undisclosed search-distribution terms are a single-counterparty risk. But at 4.3x operating income the purchase price amortizes itself before the decline can bite, and Bending Spoons' monetization toolkit has never had a base this large to work with. AOL is the balance-sheet anchor of the whole 2025-26 expansion.

The Video Stack — Vimeo + Brightcove + StreamYard (~24% of PF revenue)

Three acquisitions in twenty months assembled a vertically layered video business: Vimeo (prosumer/enterprise hosting and tools, ~$420M revenue), Brightcove (large-enterprise streaming infrastructure, ~$175M, 1,700 paying customers at ~$100K+ implied ACV), and StreamYard (browser-based live production for creators). All three were sub-scale challengers in markets dominated by YouTube on one end and bespoke enterprise platforms on the other; none was growing at acquisition; all three were bought for a combined ~1.5-2x revenue. The transformation logic is shared infrastructure and shared sales motion — and StreamYard's completed playbook run (-71% headcount, return to growth, 91% NRR) is the template Vimeo and Brightcove are now undergoing.

Our view: strategically the most coherent cluster — the only place in the portfolio where acquisitions compound each other rather than merely co-existing. Execution risk is highest at Brightcove (enterprise churn compounds: the F-1 concedes declining subscribers through 2025) and the competitive backdrop (YouTube, Twitch, Mux, Cloudflare Stream) is the toughest the company faces. If the video stack returns to growth as a unit by FY2027, it will be the proof that the playbook works on connected B2B assets, not just standalone consumer apps.

Events & Communities — Eventbrite + Meetup (~12% of PF revenue)

Eventbrite ($292M FY25 revenue, bought for $505M — 1.7x revenue for a marketplace that IPO'd at a $1.8B valuation in 2018) and Meetup (bought in the January 2024 three-deal basket) give the portfolio its only genuine network-effect assets: organizers bring attendees, attendees discover organizers. Both were shrinking under prior ownership; Meetup has already returned to subscriber and ARPU growth under the playbook (a disclosed top-3 organic contributor in Q1 2026), while Eventbrite — closed March 10, 2026 — is three months into its restructuring with an operating loss at entry.

Our view: the highest-beta cluster. Marketplaces respond non-linearly to product and pricing changes — in both directions. The Meetup result is encouraging and the take-rate toolkit (Eventbrite monetizes ~8-9% of gross ticket value) gives transformation levers consumer apps lack, but ticketing is competitive (Ticketmaster/Live Nation above, free tools below) and discretionary-events exposure adds cyclicality the subscription portfolio doesn't have.

Productivity & Prosumer Tools — Evernote, WeTransfer, Harvest, MileIQ, Loomly, Issuu (~25-30% of PF revenue, est.)

The model's heartland: decade-old utilities with habituated user bases — notes (Evernote, ~200M historical registrations, 7.2-year average customer tenure at acquisition), file transfer (WeTransfer, 58M MAU and ~$150M+ revenue), time tracking (Harvest), mileage logging (MileIQ), social-media scheduling (Loomly), digital publishing (Issuu). Each competes against larger free alternatives (OneNote/Notion, Google Drive/Dropbox, built-in OS features) and survives on workflow lock-in and switching friction. This cluster generated the disclosed organic growth in both 2025 and Q1 2026 (WeTransfer and Meetup the standouts) and contains the playbook's two complete case studies: Evernote (registrations from -20%/year to +20-29%, ARPMAU 2.5x, 99% NRR) and the Harvest integration (staffed by redeploying half the Evernote team).

Our view: reliable, harvestable, and the most exposed to the AI-substitution question — note-taking, file transfer, and mileage logging are precisely the utilities that OS vendors and AI assistants absorb first. The 8-year tenure base makes erosion slow; it does not make it stoppable. We model this cluster as the stable core through FY2028 with the long-run question deliberately unresolved.

Consumer Apps — Remini, komoot, Splice, Mosaic portfolio (~12-15% of PF revenue, est.)

The original Bending Spoons business: high-velocity consumer subscriptions distributed through the app stores. Remini (AI photo/video enhancement) is the showcase — subscriptions from 43% to 85% of its revenue, ARPMAU +50% on 5x MAU growth, seven viral spikes of which six were deliberately engineered — and also the cautionary tale, declining in Q1 2026 as MAU fell. komoot (outdoor route planning, bought March 2025, competing with AllTrails and Strava) was a top-3 organic contributor in Q1 2026; Splice (music sampling) declined through 2024.

Our view: structurally the lowest-quality revenue in the portfolio — hit-driven, app-store-dependent, fashion-exposed — and management appears to agree, given that no acquisition since komoot targets this profile. Expect this cluster to shrink as a share of revenue indefinitely; the question is whether Remini's decline is a viral-cycle trough (its history says spikes can be re-engineered) or the start of AI-native substitution (Topic 5).

Hardware-Attached Subscriptions — Tractive (excluded from all filing financials)

The newest direction: $900 million (including $119M deferred) for an Austrian pet GPS/health-monitoring business — collar devices plus a required subscription — closed May 18, 2026, three weeks before filing, with zero financial disclosure. Tractive competes with Whistle, Fi, and Apple-AirTag improvisations in a structurally growing pet-tech market.

Our view: strategically interesting (hardware attach creates churn resistance no pure app can match; the subscription is the product, the collar is the moat) and procedurally troubling — the second-largest acquisition in company history will hit the income statement two quarters after IPO buyers commit, sight unseen. The diligence burden lands entirely on trust in the published hurdle rates.

Customer Economics Snapshot

• Net revenue retention: 93% (2023), 91% (2024), 95% (2025), 94% (Q1 2026) — existing cohorts shrink in dollar terms every year; by business: Evernote 99%, AOL 95%, StreamYard 91%, Remini 87%.
• Tenure is the asset: 48% of Q1 2026 subscription revenue from customers with ≥5 years of tenure, 28% from ≥10 years; revenue-weighted average tenure 8.0 years.
• Distribution is nearly free: organic channels drove 79-83% of new-customer revenue every year since 2023; advertising expense fell from 9% of revenue (2023-24) to 3% (Q1 2026).
• Concentration is nil: no customer exceeds 1% of revenue.

The full customer and end-market analysis — including revenue-quality scoring — is in its own section below.

Assessment: Read as a portfolio, the revenue base is more diversified than any single-product software peer and lower-quality than its 84%-subscription mix implies: roughly a quarter of pro forma revenue (AOL) is in managed decline, another fifth (video stack) is mid-turnaround, and the highest-NRR assets are the most AI-exposed. What the breakdown makes visible — and the F-1's single-segment reporting obscures — is that Bending Spoons is always simultaneously harvesting some assets, transforming others, and incubating organic growth in a third group. The investor is buying the conveyor belt, not any position on it.

Company History & The Playbook

Origins: "0 for 1" in Copenhagen

The founding story matters here more than for most IPOs, because the company's only real product is its operating discipline. Luca Ferrari, Francesco Patarnello, Matteo Danieli, and Luca Querella — engineering graduates of the University of Padua and the Technical University of Denmark — first built Evertale, an AI-based "self-writing diary" startup, between 2010 and 2013. It failed. The team regrouped in 2013 as Bending Spoons (the name a Matrix joke about doing the seemingly impossible, chosen, in their words, because "a touch of irony seemed appropriate" for a company starting with "$40,000, a team of five, and a track record that read 0 for 1"), and pivoted from building products to buying them: the first acquisition cost $10,000. Thirteen years later the company deploys billions per quarter against the same thesis.

"Our Playbook (acquire, transform and optimize, then reinvest) has stayed essentially the same since 2013... Along the way, we learned about Henry Singleton at Teledyne and Tom Murphy at Capital Cities... In recent years, we studied Broadcom, Danaher, and TransDigm, each of which demonstrated how excellence in operating businesses can generate attractive financial outcomes through acquisitions." — Founder letter, F-1

The intellectual lineage is announced rather than implied — Singleton, Murphy, Danaher, TransDigm — and the letter closes with the tell of the genre: "perhaps our best acquisition opportunity will be the shares of our own company." This is a permanent-capital compounder pitch, executed in consumer software, headquartered in Milan, and staffed by what the company calls Spooners: 621 core team members (27% of 2,284 total FTEs) selected from approximately 800,000 job applications in 2025, of which 286 were hired — a 0.04% acceptance rate.

Key Milestones

M&A Track Record

M&A assessment: Eleven disclosed transactions, zero disclosed write-offs, two demonstrable full-cycle successes (Evernote, StreamYard cohort), one visible struggle (Brightcove), and four deals too recent to grade — with the four newest representing 69% of all capital ever deployed. The discipline signature is consistent: entry multiples of 1.2-3.5x revenue, no contingent earnouts before 2026, all-cash structures, and targets chosen for installed-base durability over growth. This is a genuinely strong record on which the only honest verdict is: proven at $200-500M deal sizes, unproven at $1.4 billion.

The Playbook, Quantified

The F-1 discloses the acquisition underwriting standard with rare precision: internal-rate-of-return hurdles of 65% levered / 25% unlevered applied to acquisitions completed from 2023 through Q1 2026, with financing modeled at the lower of 85% of enterprise value or the debt the target's own free cash flow can repay within five years. Capital deployed has scaled from $194 million in all of 2023 to $2.01 billion in Q1 2026 alone. Management concedes the hurdles will likely fall as scale grows — "there's plenty of room for returns to fall before we'd find them unattractive."

What "transform" means in practice, from the disclosed case studies:

Where We Are Now

The company arrives at the IPO mid-metamorphosis: the proven small-deal machine of 2013-2024 is eighteen months into becoming a large-cap consolidator, with the three biggest transformations in company history running concurrently, a debt stack at leveraged-loan pricing, and a stated intention to keep deploying "almost all available capital" indefinitely. History generates two forward-looking conclusions. First, the playbook is real and repeatable at small-to-mid scale — Evernote, StreamYard, and Remini are three independent demonstrations across three business types, and the consolidated financials corroborate the case studies. Second, the 2025-26 vintage is a different bet: AOL, Vimeo, and Eventbrite together cost 6.5x more than every prior acquisition combined, and the F-1's own framing ("only a few hundred to remain" of 1,830 acquired employees) makes clear the same surgery is being attempted on patients an order of magnitude larger, simultaneously, on borrowed money. The playbook's scale-invariance is the thesis question, and it is unproven by design — every cohort must be bigger than the last for the model to compound.

Investment Thesis

Bull Case

1. The transformation playbook is proven, codified, and margin-accretive at every demonstrated scale. Three completed case studies across three business models (Evernote, StreamYard, Remini) show the same signature — 70-85% headcount reduction, monetization re-architecture, return to output growth — and the consolidated numbers corroborate it: Adjusted Operating Income margin rose from 36% (2023) to 51% (Q1 2026) while gross margin gained nearly 700bp. If the same curve holds for AOL, Vimeo, and Eventbrite, FY2027 core operating income approaches $1.4-1.5 billion against a $2.6 billion revenue base.
2. Entry prices make the leverage safer than the headline stats suggest. The portfolio was assembled at 1.2-3.5x revenue — against public software comps at 4-10x — and AOL alone (bought at 4.3x carve-out operating income) contributes pre-tax income covering ~70% of the consolidated interest bill. The debt is expensive but the assets servicing it were bought cheap; that combination is how 65% levered IRRs happen.
3. The revenue base is an annuity on habit. 8.0-year average subscriber tenure, 48% of subscription revenue from 5-year-plus customers, no customer above 1% of revenue, and 79-83% of new-customer revenue arriving through free organic channels. Distribution costs 3% of revenue. This is consumer-staples economics wearing a software multiple's clothes.
4. AI structurally widens the arbitrage — for now. 90% of pull requests AI-(co)authored, revenue per core employee of $2.57M and rising, and a hiring funnel selecting 0.04% of 800,000 applicants. Every future acquisition gets transformed by a cheaper, faster machine than the last one did.
5. The reinvestment runway is effectively unlimited relative to deployment capacity. A disclosed universe of 1,000+ identified targets with ~$400B aggregate revenue means returns, not opportunity, govern the compounding rate for a decade.

Bear Case & Key Risks

1. The organic engine is a treadmill. Organic growth of 7% / 13% / 6% (2024 / 2025 / Q1 2026) against 91-95% net revenue retention means existing cohorts shrink ~6%/year before new customers arrive — and the flagship legacy app (Remini) is now declining outright. All reported growth above mid-single digits must be purchased, forever.
2. Vintage concentration: $4.2 billion across four concurrent mega-integrations. AOL, Vimeo, Eventbrite, and Tractive — 69% of all capital ever deployed — are being transformed simultaneously, while going public, under CFIUS review, on borrowed money. One slip extends deleveraging and breaks the "machine" narrative; there is no diversification across time.
3. Leverage is high, expensive, and structurally necessary. $4.48B of principal at SOFR+587.5bp (~8.5% blended, ~$373M annualized interest) against FY2025 operating cash flow of $291M. The model cannot stop borrowing without stopping growing.
4. Adjusted earnings flatter the economics. Transaction and reorganization costs ($164M in FY2025; $114M in Q1 2026 alone) are recurring cash costs of a perpetual-acquisition strategy, and OCF conversion of adjusted net income was 77% in FY2025 and 37% in Q1 2026. The 51% adjusted margin is not owner earnings.
5. Governance and disclosure are built for the founders, not for you. Class A shares carry 5x votes and elect the entire board; FPI status means semi-annual reporting, no proxy rules, no Reg FD; the debut interims were restated; and the adviser web (Allen & Company shareholder-underwriter-board seat; Intesa lender-underwriter-board link) is conflicted throughout.
6. Washington is watching. CFIUS is reviewing AOL and asking about Eventbrite — a systematic overhang for a foreign serial acquirer whose pipeline is dominated by US consumer-data assets.

Variant Perception

The roadshow debate will be fought over the wrong question — "is 6% organic growth acceptable for a software multiple?" — because both sides will price BSP against operating-company comps. Our variant view is that BSP is not an operating company with M&A on top; it is a capital allocator whose operating assets are working inventory, and the correct valuation anchor is the return on redeployed capital (demonstrated: 25% unlevered hurdles, met through Q1 2026) times the credibility of redeployment at the new scale (undemonstrated). That framing cuts both ways against consensus: it makes the organic-growth bear case largely irrelevant (the assets are bought to be harvested, not grown), and it makes the adjusted-margin bull case insufficient (margins on existing assets don't compound; redeployment does). The practical mispricing: if the deal prices off software-growth comps it will be too cheap on earnings power; if it prices off the AppLovin-adjacent "AI consumer platform" story some bankers will tell, it will be far too expensive on redeployment reality. Pre-pricing, the edge is having the framework ready while the market argues about labels.

Kill Criteria

Financial Profile

Historical Summary

Three patterns to fix in mind before the bridges below: GAAP operating margin has been flat around 20% for three years while the adjusted margin marched from 36% to 51% — the gap is the deal machine's cash and accounting cost. Net income is uninformative in every period (tax artifacts each year — see Key Topics #7). And the annualized Q1 column overstates run-rate net income but understates run-rate cash flow, because Q1 carried both peak reorganization spending and the heaviest interest quarter in company history.

Most Recent Quarter Snapshot (Q1 2026)

Quarter assessment: Q1 2026 is the first look at the post-AOL company and it carries the model's full signature in a single quarter: transformative inorganic growth (AOL contributed $141.8M of the $342M increase), record adjusted margins, peak restructuring spend, a five-fold interest bill, soft organic growth with the largest legacy app declining — and positive GAAP net income anyway. The quarter the company chose to go public on is, fittingly, the quarter that contains the entire debate.

Revenue Bridge: FY2024 → FY2025

Decomposing the +95% reported growth into its acquired and organic components, using the F-1's disclosed per-deal revenue contributions:

The bridge formalizes the model: roughly $548 million of the $635 million increase (86%) is acquisition timing, and the organic residual is itself heavily price-driven. The forward growth algorithm is therefore: reported growth ≈ capital deployed × (1/purchase multiple) + 6-13% organic. At the FY25 pace of ~$2-4 billion deployed annually at ~2-3x revenue, acquisitions alone add $700M-$1.5B of revenue per year — until they don't.

Margin Bridge: FY2024 → FY2025 (GAAP Operating Margin)

The bridge isolates the analytical disagreement an investor must resolve: the three sustainable components (+798bp combined) are the transformation engine working as advertised, while the G&A drag (-569bp) is the recurring toll of running the engine. Bulls will note the toll is discretionary — stop acquiring and margins snap toward the adjusted figure. Bears will note the company has promised never to stop acquiring. Both are right; that tension is the stock.

Cash Generation & Earnings Quality

FY2025 operating cash flow of $291 million against Adjusted Net Income of $376 million is a 77% conversion ratio; Q1 2026 conversion was 37% ($75.7M restated OCF vs. $206M Adjusted Net Income). The wedge is structural, not suspicious: cash interest (~$93M/quarter at current debt), cash restructuring payments, and Italian cash taxes are all excluded from adjusted earnings but not from the bank account. Working capital is a tailwind — $479 million of deferred revenue means subscribers pay in advance — and capex is negligible ($13M of PP&E on a $7B balance sheet; the company owns laptops, not factories). Free cash flow before acquisitions is effectively OCF, and acquisitions are the entire investing line ($1.85B in FY2025, $1.65B in Q1 2026 alone).

Assessment: Treat Adjusted Net Income as an upper bound on owner earnings and OCF as the floor; true distributable earnings power on the current portfolio sits around $500-700 million annualized once the 2025-26 reorganizations roll off but interest stays. The equity story requires that gap to close from both ends — reorg costs falling, AOL/Vimeo/Eventbrite cash earnings rising — through 2026. The semi-annual FPI reporting cadence means investors will get exactly two looks per year at whether it is happening.

Capital Allocation

There is exactly one use of capital here, by explicit design: acquisitions. Deployment scaled from $194 million (2023) to roughly $760 million-plus (2024: WeTransfer $476M, the Meetup/StreamYard/Issuu basket $281M, plus the undisclosed Mosaic asset deal) to ~$2.28 billion (2025: Vimeo $1.36B, the four-deal basket $701M, Brightcove $218M) to $2.01 billion in Q1 2026 alone (AOL $1.45B, Eventbrite $505M) — with Tractive's $900 million following in May. No dividends, no buybacks, no debt paydown beyond contractual amortization; the founder letter frames even the IPO through a Singleton lens ("perhaps our best acquisition opportunity will be the shares of our own company" — the only hint that buybacks ever enter the calculus, and only at the right price). For investors, capital-allocation diligence here replaces product diligence entirely: the dividend policy is the deal pipeline.

Return on Capital

The company's claim: 65% levered / 25% unlevered IRRs on acquisitions completed 2023 through Q1 2026. The outside check: roughly $6.1 billion of disclosed consideration deployed since January 2023 now produces ~$1.09 billion of annualized core operating income (our normalized definition) — an ~18% pre-tax unlevered yield on gross invested capital, with the largest deals (69% of that capital) less than three quarters old and pre-transformation. Given the demonstrated 2-3 year maturation curve (Evernote's economics tripled between years one and three), the disclosed 25% unlevered claim is consistent with the consolidated arithmetic rather than contradicted by it — a rare and checkable virtue in an IPO prospectus. The spread over the ~8.5% cost of debt and our ~10% cost of equity is the entire economic engine: every year it persists at current deployment scale adds roughly $300-500 million of annual earnings power. Track it deal by deal; it is the single number the whole thesis compounds on.

Industry Deep Dive & TAM

Bending Spoons' "industry" is not a product market — it is the market for buying digital businesses, with the underlying app/SaaS markets mattering only as the weather its assets live in. Both layers need sizing.

Market Size & Growth

The honest TAM statement: supply of targets is effectively unlimited relative to BSP's balance sheet for the next decade. The binding constraints are integration bandwidth, debt capacity, and — increasingly — competition for assets.

Industry Structure: Who Sells, Who Competes to Buy

The sellers are structural and replenishing: venture-backed companies that plateaued below venture-scale outcomes (Evernote, Eventbrite, Vimeo all fit this profile), corporate orphans shed by strategics (AOL out of the Yahoo/Apollo complex, Mosaic out of IAC, MileIQ originally out of Microsoft), and founder-owned utilities facing succession (komoot, Tractive). Every venture cycle mints a new cohort: the 2020-21 funding bubble alone created hundreds of subscription apps now stranded between growth expectations and reality.

The buyers BSP competes against: software-focused private equity (the Thoma Bravo/Vista complex and hundreds of smaller funds), strategic consolidators (Ziff Davis in media, Constellation and its spawn in vertical SaaS), and special situations desks. BSP's differentiation in auctions is operational, not financial — it underwrites deeper cost cuts and faster monetization lifts than a PE operating partner model can, which lets it pay a "full" price that is still cheap against its own transformation math. The early evidence is the 82-91% premiums it paid over Eventbrite's and Vimeo's pre-deal trading prices while still entering at 1.7x and 3.3x revenue.

Secular Trends

1. App-economy maturation produces orphans faster than it produces winners. The consumer-software industry is two decades old; its long tail of habituated-user, low-growth products grows every year, and public markets increasingly refuse to hold sub-scale software (Eventbrite and Vimeo both traded below 2x revenue before BSP arrived). Tailwind for the buyer of record.
2. AI collapses the cost of operating mature software. The lean-ops thesis BSP pioneered is becoming industry consensus — which both validates the model and, over time, arms every competing buyer with the same spreadsheet. Expect entry multiples across the category to rise as transformation math becomes standard underwriting.
3. App-store economics are loosening. Regulatory pressure (EU DMA, US litigation) is gradually reducing platform take rates and steering restrictions — a direct gross-margin tailwind for a portfolio that monetizes heavily through Apple and Google payment rails, and one BSP does not control.
4. Cross-border deal scrutiny is tightening. CFIUS attention to consumer-data assets (already visible on AOL and Eventbrite) adds a tax — in time, certainty, and occasionally structure — on exactly the US-heavy pipeline BSP targets.

Cyclicality & Macro Sensitivity

The subscription core (84% of revenue) is consumption-staple-like: small-ticket, auto-renewing, habit-anchored. The acquired diversification adds cyclical surface area — advertising (12% of mix, AOL) and discretionary events (Eventbrite) both lean into the consumer cycle — and the model's funding side is genuinely rate-sensitive: floating-rate TLB debt at SOFR+587.5 means every 100bp of policy easing adds ~$30-40M to pre-tax income, and vice versa. The deepest macro exposure, though, is deal-market beta: tight credit shuts off the growth engine even when the portfolio performs.

Regulatory Landscape

Four live fronts: CFIUS (AOL under review, Eventbrite queried — the recurring toll on US acquisitions); app-store/platform regulation (cuts both ways: cheaper payments, but also AI-agent discovery disrupting the organic channels that supply ~80% of new-customer revenue); EU digital regulation (DSA/DMA/GDPR — a Milan-headquartered issuer gets no home-field exemption); and consumer-protection scrutiny of subscription dark patterns — auto-renewal, cancellation friction, and usage-limit paywalls are the exact levers of the monetization playbook, and the FTC and EU consumer authorities have both been tightening rules around them.

Competitive Landscape: The Comparable Set

No public company does exactly this. The comp set triangulates from three directions — serial-acquirer compounders (what BSP aspires to be priced as), consumer-internet operators (what its assets look like individually), and the levered/declining roll-ups (what skeptics will call it):

Trading data as of June 13, 2026; compiled from public market-data aggregators, cross-checked across at least two sources. Constellation figures converted from CAD.

Assessment: The spread between Constellation (18.6x EBITDA) and Ziff Davis (5.0x) is the entire BSP debate compressed into one axis: serial acquirers earn compounder multiples only while the market believes redeployment returns persist. BSP's margins argue for the top half of the range; its organic profile, leverage, and disclosure regime argue for the bottom half. Where it lands inside that spread — initially set by IPO pricing, then re-set by two semi-annual reports a year — will matter more to returns over the next three years than any operational outcome.

Competitive Position & Moat Assessment

Bending Spoons does not compete in a product market in the conventional sense — each portfolio app has its own competitors (Notion against Evernote, Dropbox against WeTransfer, YouTube against Vimeo) — but the company-level competition is for assets: against private equity firms, strategic acquirers, and other software consolidators bidding on the same neglected digital businesses. The moat question is therefore: what lets Bending Spoons pay more than rivals for a target and still earn 25% unlevered IRRs?

Transformation capability (the process moat)

The demonstrated ability to run acquired software businesses with 18-30% of their prior headcount at higher output (Evernote: -82% FTEs, product releases doubled) is the economic engine. It rests on accumulated procedure — proprietary deployment/experimentation tooling, monetization data from 1,000+ pricing experiments, and a hiring funnel selecting 0.04% of applicants. This is the closest analogue to the Danaher Business System in consumer software: not patentable, but embodied in people and practice, and compounding with each integration.

Durability: Stable-to-strengthening — each deal adds data and trained operators; the AOL/Vimeo/Eventbrite triple integration is simultaneously its largest test and largest training set.

Underwriting discipline (the capital-allocation moat)

Published hurdle rates (65% levered / 25% unlevered), a co-founder running M&A, and a 13-year record of paying 1.2-3.5x revenue for assets others wouldn't touch (AOL at 4.3x operating income). Discipline is a moat only as long as it survives scale and auctions; the company itself warns hurdles will fall.

Durability: Eroding at the margin — moving from proprietary small-cap deals to public-company auctions (Eventbrite, Vimeo) necessarily compresses the entry-price advantage.

Installed-base inertia (the asset-level moat)

The portfolio's 8.0-year average subscriber tenure and 48%-of-revenue-from-5yr+ cohorts represent habit, stored data (notes, videos, routes), and workflow lock-in. This is the moat Bending Spoons buys rather than builds — and the company monetizes it more aggressively than prior owners dared (sub-100% NRR is the toll).

Durability: Stable but depleting per asset — inertia harvested through price increases is a one-way withdrawal; the model replenishes it by acquiring new installed bases, not by deepening old ones.

What is NOT a moat here

• Network effects: essentially none portfolio-wide; Meetup and Eventbrite have local marketplace dynamics, but there is no cross-product graph. The "over 500 million MAU" figure is an aggregation, not a network.
• Cross-sell: the F-1 claims little and the products' audiences barely overlap; we model zero.
• AI capability per se: the same frontier models Bending Spoons uses (Anthropic, Google, OpenAI — disclosed as swappable vendors) are available to every PE operating partner. The current edge is organizational adoption speed, which is a lead measured in quarters, not years.

Competitive threats

• Private equity convergence: Thoma Bravo, Vista, and a hundred smaller funds are building AI-era lean-ops capabilities. As they do, auction prices for "neglected software with loyal users" rise, compressing BSP's entry IRRs — the single most important long-term threat.
• AI-native substitution: the same technology that lets 60 people run Evernote lets 6 people rebuild it. Utility apps with thin feature sets (file transfer, photo enhancement, mileage tracking) are the most exposed; Remini's Q1 2026 MAU decline may be the first data point.
• Platform dependency: Apple App Store and Google Play control distribution and payments for much of the consumer portfolio; AWS and Google Cloud host it; AI-agent search could disintermediate the organic discovery channels that supply 79-83% of new-customer revenue.
• Reputation accumulation: each price-increase cycle (Evernote, WeTransfer) generates user resentment that transfers to the parent brand; at some portfolio size, "acquired by Bending Spoons" itself triggers churn — a self-limiting dynamic the company must manage asset by asset.

Customer & End-Market Analysis

Customer Profile

End-Market Exposure

Aardvark estimates from disclosed acquisition contributions and pro forma figures; the F-1 reports a single segment and provides no end-market split.

Customer Dynamics

Revenue Visibility

High at the portfolio level despite churn at the cohort level: 84-93% subscription mix, $479M of deferred revenue (customers pay in advance), and the 8-year tenure base mean the next twelve months of revenue are substantially in hand at any moment. What visibility does not extend to is the growth line — sub-100% NRR plus mid-single-digit organic means the existing book reliably tells you what shrinks, and acquisitions decide what prints.

Pricing Power — Demonstrated, and Demonstrably Resented

The playbook's central empirical claim is that acquired user bases tolerate dramatic price increases: Evernote's ARPMAU 2.5x in three years, WeTransfer monetization up on both price and subscribers, Remini ARPMAU +50%. The tolerance is real but not free. The price-hike cycles generated documented user backlash and migration (Evernote's free-tier restriction to 50 notes was widely covered; WeTransfer's co-founder publicly criticized the new ownership and announced a competing service in late 2025), and the 87-91% NRR at Remini and StreamYard is what harvested goodwill looks like in the numbers. Pricing power here is a withdrawal from a finite account of accumulated user trust — large balances, real interest, but withdrawals nonetheless.

Concentration & Stickiness

Atomized customers mean no negotiating leverage against BSP — no renewal cliff, no single contract that matters. Stickiness is data-gravity-dependent: a decade of notes in Evernote or a video library in Vimeo is a real switching cost; a file-transfer habit is not. The portfolio's stickiest revenue (Evernote 99% NRR) and its loosest (Remini 87%) differ by exactly this dimension, which is why the acquisition mix shifting toward data-accumulating and hardware-attached assets (Eventbrite, Tractive) is the right strategic read.

End-Market Health

The healthiest exposure is the events recovery (Eventbrite/Meetup ride a structural live-experiences tailwind); the sickest is legacy media (AOL's user base ages out at a knowable rate); the most uncertain is consumer utilities, where AI-native substitution risk is live (Remini already declining). Nothing in the portfolio is hostage to a single end-market — which is precisely the design.

Leadership & Key People

Executive Team Overview

Note the pattern: every operating executive except the GC is either a co-founder or an internal promote, most are under 40, and none has held an officer role at a public company. The fourth co-founder, Luca Querella, holds no executive role but remains a 5%+ shareholder and class A holder.

CEO Deep Dive: Luca Ferrari

Luca Ferrari, co-founder and CEO since June 2013, is the architect of the playbook and its public voice. The formation is atypical for a consumer-software founder: engineering degrees from the University of Padua and the Technical University of Denmark, two years as a McKinsey associate (2010-2012), and a failed AI startup (Evertale, where he overlapped with all three co-founders) before Bending Spoons. The arc explains the company. From McKinsey: the returns-first capital-allocation framing, hurdle-rate discipline, and the habit of writing strategy in public (the founder letter reads like an owner's manual, closing "Letter done, back to building."). From the Evertale failure: the pivot from building products at zero to buying products with installed bases — the foundational insight that distribution and habit, not code, are the scarce assets in consumer software.

The execution record under his tenure is the strongest argument for the stock: $40,000 of seed capital in 2013 compounded into $1.31 billion of FY2025 revenue ($2.61 billion pro forma) with no outside CEO, no pivot away from the 2013 playbook, and an explicitly Singleton/Murphy-styled philosophy now backed by thirteen years of consistent practice. The communication style — quantified, self-critical in specifics ("0 for 1"), allergic to adjectives without numbers — reads more like an owner's letter from the capital-allocation canon than a tech-IPO prospectus, and it sets up unusually testable public commitments (published hurdle rates, named practices investors are told they may dislike). The same record contains the governance bet: Ferrari has never run a public company, reports semi-annually as an FPI, and will control the company through super-voting shares alongside three co-founders. Key-person concentration is extreme and structural — the playbook is the founders' judgment, applied deal by deal.

Is this the right CEO for this phase? For the machine as designed — founder-controlled, acquisition-led, indifferent to quarterly optics — yes, almost definitionally: the company is an expression of his framework. The honest caveat is that the next phase introduces constituencies the framework has never had to serve (public minority holders, credit markets in stress, US regulators), and the F-1's governance choices systematically minimize their leverage over him. Investors are not hiring a CEO; they are subscribing to one.

The Engine Room

• Francesco Patarnello (co-founder, vice chair, head of business acquisitions): runs the deal machine — sourcing, underwriting, and the IRR-hurdle process the entire thesis depends on. That M&A is led by a co-founder rather than hired bankers is, in our view, a feature: incentive horizon matches the "never sell" underwriting assumption.
• Co-CFO structure: Scarpazza (financing and tax) plus Martinelli (accounting and reporting) — splitting the CFO seat between a financing specialist and an internally-promoted engineer two quarters before an IPO is unconventional. It mirrors the company's talent-over-experience doctrine — and it means the finance organization, like everything else here, is run by people in their thirties doing the job for the first time at this scale. The Q1 cash-flow restatement (Key Topics #10) is the first small test of that bet.
• Francesco Mancone (CTO): owns the proprietary-technology layer — including Minerva, the AI lifetime-value engine underwriting acquisitions since 2019 — and the AI deployment behind the 90%-of-code productivity claims.
• The talent machine: ~800,000 applications in 2025 for 286 hires (0.04% acceptance); a centralized Talent team with final authority over all hiring and separations; no more than three management layers between CEO and individual contributor; "nearly all of our businesses and functions are led by people in their twenties or thirties," per the founder letter.

Board of Directors

It is a genuinely strong board on paper — five independents of nine, an audit committee (Sinwell chairing, with Mylod, O'Neal, Tagliavini) that would grace a company many times this size. The structural caveat empties much of that strength: under the bylaws, the entire board is elected by the class A shareholders — the four co-founders — and Italian FPI status exempts the company from Nasdaq's compensation-committee, nominating-committee, and executive-session requirements, all of which it intends to use. The independents serve at the founders' pleasure, in substance advisory rather than supervisory.

Compensation & Alignment

Total 2025 compensation for all executive officers and directors combined was approximately $10.8 million — $2.2 million cash, $8.6 million equity — strikingly lean for a company of this scale (many US software CEOs individually out-earn this entire C-suite). The deeper alignment mechanism is the company-wide "equity acquired, not granted" doctrine: employees buy equity at a discount by forgoing cash salary rather than receiving automatic grants. In 2025, 84% of eligible team members converted an average of 28% of cash compensation into options, and dilution from equity compensation averaged just 1.5% per year from 2023 through Q1 2026 — a fraction of the 3-6% annual dilution typical of venture-formed software companies.

Insider ownership percentages are blank pending pricing, but the architecture is disclosed: the four co-founders hold all class A super-voting shares; the named 5%+ holders are Galileo Quattordici S.à r.l. (a Luxembourg vehicle), Baillie Gifford, and co-founder Luca Querella; and Selling Shareholders will take some chips off the table in the offering itself — the sizes, when the final amendment lands, will be worth reading closely.

Key Topics from the F-1

1. Organic Growth Is 6-13% — And Management Wants You to Underwrite It Anyway

"An acquisition target's projected organic growth is embedded in our assumptions and influences the price we're willing to pay. But we ultimately make investment decisions based on expected returns, regardless of the organic growth profile and assuming we'll never sell the acquired business." — Founder letter, F-1

The single most important disclosure in the filing is the organic growth series: 7% (2024), 13% (2025), 6% (Q1 2026). Note the definition is generous — it includes the estimated pre-acquisition revenue of acquired businesses in the base, so it captures the post-acquisition monetization lift Bending Spoons engineers. Even so measured, the engine decelerated to 6% in the most recent quarter, and the detail is more sobering than the headline: Q1 2026 organic growth was carried by komoot, Meetup, and WeTransfer while Remini — the largest legacy consumer app — declined on falling MAU, and Brightcove declined throughout 2025 on subscriber losses.

The math investors must hold in their heads: at 94% net revenue retention, the existing book sheds roughly 6 points of revenue per year before new customers arrive. Organic growth of 6-13% therefore implies new-customer acquisition contributes 12-19 points annually — respectable, but it means the portfolio's organic engine is a treadmill, not a flywheel. The growth that prints in the income statement comes from capital deployment.

Assessment: This is not a hidden flaw; it is the stated strategy, and the company prices acquisitions accordingly. But it has a direct valuation consequence: BSP cannot be valued on a revenue multiple calibrated to organic-growth software peers. It must be valued as a leveraged compounder — on earnings power, return on incremental capital, and reinvestment runway — which is exactly how we approach it below.

2. The 2025-26 Step-Change: $4.6 Billion Deployed in Eighteen Months

From Evernote (January 2023) through Harvest (July 2025), Bending Spoons' disclosed acquisition spend totaled roughly $1.9 billion across ten-plus deals — measured, escalating, self-funded plus moderate debt. Then the tempo changed: Vimeo for $1.36 billion (closed November 24, 2025), AOL for $1.45 billion (January 2, 2026), Eventbrite for $505 million (March 10, 2026), and Tractive for $900 million including deferred consideration (May 18, 2026) — approximately $4.2 billion in six months, funded by a debt stack that grew from $2.71 billion to $4.48 billion of principal in a single quarter. Three of the four targets were public companies or carve-outs (Eventbrite was NYSE-listed; AOL was carved out of the Yahoo/Apollo entity; Vimeo was Nasdaq-listed), meaning Bending Spoons is now competing in auctions against private equity rather than picking off neglected assets quietly.

The prices paid remain disciplined on revenue multiples — AOL at ~2.3x revenue and ~4.3x carve-out operating income, Eventbrite at ~1.7x revenue, Vimeo at ~3.2x, Brightcove at ~1.2x — but the integration load is unprecedented: 1,830 employees arrived in two quarters, more than doubling the company, with the stated intent to retain "only a few hundred."

Assessment: The vintage-risk concentration is the underappreciated risk in this filing. Every prior transformation was sequential; AOL, Vimeo, Eventbrite, and Tractive are concurrent, and they happen while the company is also going public, also managing a CFIUS review, and also carrying leveraged-loan debt at an 8.5% blended cost. If the 2025-26 vintage transforms on schedule, FY2027 adjusted earnings power approaches $1.5 billion and the IPO will look cheap at almost any plausible price. If integration slips even one of the four, the deleveraging math extends and the "machine" narrative takes its first public dent. There is no diversification across time here — the bet is singular and simultaneous.

3. AOL: The LBO Inside the IPO

The AOL acquisition deserves standalone treatment because it is the largest deal, the most profitable asset, and the clearest expression of the model's logic. Bending Spoons paid $1.45 billion for a business that produced $633 million of revenue and $334 million of operating income (a 53% margin) in FY2025 on a carve-out basis — roughly 4.3x operating income for an asset in structural decline. In its first quarter under Bending Spoons ownership (January 2 to March 31, 2026), AOL contributed $141.8 million of revenue and $65.5 million of pre-tax income — by itself covering roughly 70% of the company's quarterly interest bill.

The strategic read: AOL is not a growth asset; it is the cash engine that services the debt that funds the next acquisitions. Its ~95% NRR (remarkable for a legacy email/portal franchise) and decade-tenured user base are exactly the profile the playbook targets — maximum-inertia installed base, minimum competition for the asset, deeply discounted price because everyone else sees a melting ice cube.

Assessment: We view the AOL deal as brilliant on price and risky on optics. Brilliant: at 4.3x operating income, even a -10%/year revenue decline yields a sub-4-year payback against the purchase price, and Bending Spoons' Q1 evidence (annualizing well above the underwrite) suggests monetization upside on top. Risky: AOL's economics depend partly on advertising (a newly volatile 12% of consolidated mix) and on search distribution arrangements whose terms the F-1 does not disclose — and the asset sits under an open CFIUS review (Topic 9). The bull case for the whole IPO is, in miniature, the bull case for this single deal.

4. What "Transform" Actually Means: Price Up, Headcount Down, Users Sideways

Strip the language of "reimagining businesses from scratch" and the disclosed mechanics of transformation are: (a) cut 70-85% of acquired headcount within 12-24 months; (b) re-architect pricing and gate features behind usage limits (Evernote's usage-limit paywall drove 51% of all 2025 conversions); (c) shift one-time buyers to subscriptions (Remini: 43% → 85% subscription mix); (d) raise average revenue per user materially (Evernote ARPMAU 2.5x in three years; Remini +50% in four). The result is structurally higher margins and — in the disclosed cases — a return to registration growth driven by product investment after the cost reset.

The consolidated income statement corroborates the pattern: gross margin rose from 61.1% (2023) to 67.9% (Q1 2026), R&D fell from 13.4% of revenue to 9.2% (FY25), advertising spend fell from 9% of revenue to 3%, and revenue per Spooner reached $2.57 million in 2025 — roughly 3-4x the revenue-per-employee of typical consumer-software companies.

Assessment: The playbook monetizes user-base inertia, and the customer-sentiment cost is real — price increases at Evernote and WeTransfer generated visible user backlash, and sub-100% NRR plus Remini's Q1 MAU decline are the quantitative residue. The model's genius is that it doesn't need loyalty, only inertia plus an 8-year average tenure. The model's limit is that inertia is a depleting resource: each portfolio company can be harvested aggressively only once. Sustained compounding therefore requires perpetual acquisition — which is precisely why the IPO exists.

5. The AI Productivity Claim: 90% of Code AI-Authored

"The share of pull requests authored or coauthored by AI at Bending Spoons grew from less than 10% in Q1 2025 to more than 90% by the end of Q1 2026, with around 70% authored by AI alone. In part helped by progress in AI, revenue per full-time equivalent Spooner increased from $1.12 million in 2023 to $2.57 million in 2025." — Founder letter, F-1

This is the most aggressive AI-productivity disclosure we have seen in an SEC filing — 70% of code authored by AI alone, deployed across a portfolio of 20+ products, by a 621-person core team running businesses that employed thousands under prior owners. The company uses models from Anthropic, Google, and OpenAI (API or self-hosted), with architecture designed for provider switching, plus proprietary narrow models. Management's thesis: AI raises the ceiling on how few people can run an acquired business, which widens the transformation arbitrage and improves the scalability of the entire model.

Assessment: Directionally credible, precisely unverifiable. The revenue-per-Spooner trend is consistent with the claim, but the metric conflates AI productivity with the arithmetic of buying revenue while capping core headcount. The deeper point cuts both ways: if AI makes lean operation of mature software assets dramatically cheaper, Bending Spoons is the best-positioned acquirer in the category — but the same technology lowers the moat around its own portfolio (AI-native competitors can rebuild an Evernote or a Remini feature set faster than ever) and will eventually compress the price of the "neglected software" assets it buys, as every PE firm adopts the same tooling. AI is simultaneously the model's accelerant and its long-run competitive threat; the F-1 only advertises the first half.

6. The GAAP-to-Adjusted Gap: Recurring "Non-Recurring" Costs

FY2025 Adjusted Operating Income of $613 million reconciles from GAAP operating income of $278 million via four add-backs: amortization/impairment of acquired intangibles ($151M), transaction costs ($85M), reorganization costs ($79M), and other items ($21M). In Q1 2026 the gap widened: $308 million adjusted versus $120 million GAAP, with reorganization expense alone running $76 million in the quarter — nearly equal to all of FY2025 — as the AOL/Eventbrite/Vimeo restructurings hit.

Two of the four add-backs deserve different treatment. Amortization of acquired intangibles is a non-cash purchase-accounting artifact; adding it back is standard and, for a serial acquirer, essential to seeing the economics. But transaction and reorganization costs are cash costs that recur by design: the company has done at least one acquisition in every year since 2021, intends to "allocate almost all available capital to acquisitions for many years to come," and restructures every target. Treating these as one-time is like an airline adjusting out fuel.

Assessment: Our normalization: take Adjusted Operating Income, subtract a through-cycle allowance for transaction + reorganization costs (~$120-160M annually at the current deal tempo, roughly their FY25+Q1 26 run-rate). That yields "core" operating earnings power of roughly $1.05-1.1 billion annualized off Q1 2026 — a low-40s percent margin. Excellent, but materially below the 51% headline, and the distinction matters at the valuation stage. Watch whether reorganization expense actually rolls off in H2 2026 as the three big transformations complete; that is the cleanest near-term test of management's framing.

7. The Tax Story: A $166 Million Move to Italy and a 100% Effective Rate

FY2025's near-zero GAAP net income is a tax artifact, not an operating outcome: the company recorded a $166 million tax expense from "the tax implications of a transfer of certain acquired businesses to Italy," partially offset by a $40 million valuation-allowance release — producing a ~100% effective tax rate against $111 million of pre-tax income (and an $84 million Q1 2025 tax expense against a pre-tax loss). Conversely, FY2023's $161 million net income was inflated by $103 million of net tax benefits from basis step-ups and group transactions.

Assessment: The economics of onshoring acquired IP to Italy are rational — Italy's IP regimes allow amortizing stepped-up intangible bases against future income, lowering cash taxes for years — and the one-time charge buys a structurally lower go-forward rate. But the three-year GAAP net income series ($161M → $89M → $0M) is analytically useless as a result; investors must build earnings power from operating income down, applying a normalized mid-20s cash tax rate. The volatility also previews life as a BSP shareholder: an Italian holding structure executing cross-border transfers will produce lumpy tax lines indefinitely, reported semi-annually under FPI rules. Model accordingly.

8. The Debt Stack: $4.5 Billion at Leveraged-Loan Pricing

Total debt principal stood at $4.48 billion at March 31, 2026: roughly $2.18 billion of USD Term Loan B (including the $950 million fourth-amendment tranche funded January 2026 for AOL), $737 million of EUR TLB, $660 million of USD Term Loan A, ~$500 million of Facility A2 tranches, and bilateral Intesa loans. The TLB prices at SOFR + 587.5bp (1% floor) with euro tranches at Euribor + 525-587.5bp — single-B-style pricing — amortizing 1.25% per quarter, maturing March 2031, with a 25bp margin step-down upon IPO. Against this: $741 million of cash, fully undrawn revolvers of €976 million + $195 million, and Q1 annualized interest expense of ~$373 million.

On annualized Q1 Adjusted Operating Income of $1.23 billion, net leverage is ~3.0x and interest coverage ~3.3x. On our normalized core earnings (Topic 6), coverage is closer to 2.9x. The remaining Tractive deferred payment ($119 million, May 2027) and any new deals add from here — unless IPO primary proceeds retire debt, which the use-of-proceeds section conspicuously does not promise.

Assessment: The leverage is rational for the model (the founder letter explicitly frames debt as the lever on acquisition IRRs — 65% levered versus 25% unlevered is the spread leverage creates) but it is the model's binding constraint. At an 8.5% blended pre-tax cost, every $1 billion of new debt needs ~$85 million of incremental operating income just to break even on the carry. The IPO's most underrated function is refinancing optionality: a public equity currency plus a likely post-IPO repricing of the TLB (the built-in 25bp step-down is a start) could cut $40-80 million from the annual interest bill. Watch the post-IPO capital-structure actions more closely than the first revenue print.

9. CFIUS Is Reviewing the AOL Deal — and Asking About Eventbrite

The F-1 discloses that Bending Spoons filed a voluntary CFIUS notice for the AOL acquisition in March 2026 (review ongoing, expected to take "several months"), and that in April 2026 the Treasury Department's non-notified-transactions unit asked questions about the Eventbrite acquisition, to which the company responded — with a formal filing possibly to follow. The filing acknowledges the theoretical outer bound: mitigation conditions, restrictions on ownership or integration, or in the extreme, divestiture.

Assessment: A forced unwind is a tail case — Italy is a NATO ally and the assets are consumer software, not critical technology — but the realistic middle cases matter: data-handling mitigation agreements on AOL's email franchise (50M+ US consumer inboxes) could constrain exactly the kind of cross-portfolio data usage the "proprietary data" platform pitch depends on, and a pending review may chill the company's appetite for US targets during the months it runs. The disclosure that Treasury proactively inquired about Eventbrite suggests Washington is now watching this acquirer's US shopping spree systematically. Expect CFIUS-related risk language to be a permanent feature of BSP filings, and treat any mitigation agreement's terms as thesis-relevant when disclosed.

10. The Restated Interims: Small Error, Awkward Timing

The Q1 2026 interim cash-flow statements in this F-1 are marked "as restated": $48 million of restricted cash acquired with Eventbrite was originally misclassified as an operating inflow (and partially netted against FX effects) rather than as part of investing activities. The correction has no income-statement or balance-sheet effect, but it cut originally-stated Q1 operating cash flow from ~$124 million to $75.7 million — a 39% reduction in the headline OCF number.

Assessment: Substantively minor, symbolically not. A company asking public investors to trust semi-annual FPI reporting, heavy non-GAAP framing, and complex multi-jurisdiction purchase accounting cannot afford classification errors in its debut filing — and this one happened to flatter the metric (operating cash flow) where the company's GAAP-versus-adjusted gap is already widest. We flag it as a credibility data point to monitor, not a thesis driver: one restatement is noise; a second would be a pattern.

11. The $400 Billion Target Universe — Reinvestment Runway as the Real TAM

"We've identified more than 1,000 digital businesses (both private and public) that we expect could be attractive acquisition targets over the next few years. These businesses generated nearly $400 billion in aggregate estimated annual revenue in 2025." — F-1, Addressable Market

Bending Spoons defines its TAM not as a product market but as an acquisition universe — the correct framing for a capital allocator. The study (internal, built on PitchBook data, explicitly not reviewed by PitchBook) implies a pipeline ~150x the company's current pro forma revenue base. The more useful arithmetic: at the company's demonstrated 1.5-3x revenue purchase multiples, the identified universe represents roughly $600B-$1.2T of potential consideration — effectively unlimited runway relative to any plausible balance sheet for a decade.

Assessment: The runway is real but the number is marketing. The binding constraints on deployment are not target availability — they are integration bandwidth (Topic 2), debt capacity (Topic 8), and competition from PE firms armed with the same AI-era operating thesis. The honest version of the claim: the supply of neglected digital businesses comfortably exceeds Bending Spoons' capacity to buy them, so returns on capital — not opportunity scarcity — will govern the compounding rate. That is a good problem, and it is also why hurdle-rate discipline (the disclosed 65%/25% IRRs) is the single most important thing to monitor as the company scales into bigger auctions.

Value Creation Framework

The compounding engine is not organic growth — it is the spread between the return on acquired capital and the cost of that capital, multiplied by the rate of deployment. The framework a long-term holder must underwrite:

What Breaks the Compounding

• Entry-multiple inflation: if auction competition pushes purchase prices from ~2x toward 4x+ revenue, the unlevered IRR math collapses toward PE-normal returns and the premium multiple thesis dies with it. Monitor disclosed prices on every future deal.
• Integration overload: a visible stumble on AOL/Vimeo/Eventbrite — revenue attrition beyond plan, reorg costs persisting into 2027, key product breakage — would mark the playbook's scale ceiling.
• Debt-market dependence: at SOFR+587.5 the model already pays junk pricing; a credit-market shutoff doesn't kill the company (covenant-light, 2031 maturity, $1.3B undrawn revolvers) but freezes the deployment engine that drives the growth rate.
• Hurdle erosion by choice: management explicitly reserves the right to lower return hurdles as scale grows. Watch the language in each annual letter; the day "65%/25%" becomes "attractive returns" is the day to reassess.

Valuation: A Framework for the Day the Range Prints

The F-1 contains no price range, so the honest output of this section is not a target price — it is a fair-value band and a pre-committed reaction function. We value BSP on normalized owner earnings (the metric a permanent-capital compounder should be judged on), cross-checked against revenue and operating-income multiples.

Step 1: Normalized Earnings Power

The FY2027E column assumes: Vimeo and Eventbrite reach 35-45% post-transformation operating margins on a pruned ~$650M combined revenue base (versus ~breakeven today — the Evernote/StreamYard precedents support this); AOL monetization holds against mid-single-digit legacy decline; reorganization expense reverts to a deal-tempo baseline; Tractive contributes its first full year; and — critically — zero credit for acquisitions not yet announced. Deployment upside is real but belongs in the multiple, not the base.

Step 2: What Multiple Does This Deserve?

The comparable set splits into the aspirational and the cautionary. Premium serial acquirers (Constellation Software-type) earn 25-35x owner earnings on decades of 100%+ NRR and unlevered balance sheets — BSP qualifies on process, not on either metric. Levered consumer-internet roll-ups and declining-asset harvesters (Ziff Davis-type) trade at 6-10x earnings — BSP's demonstrated redeployment returns argue it is decisively better than that cohort. Quality-compounder-with-leverage analogues (TransDigm in its growth phase) sustained high-teens-to-20x owner earnings with far stronger asset-level moats. Netting BSP's strengths (process moat, hurdle discipline, 0.04%-selectivity talent engine) against its specific frictions (sub-100 NRR, 3x leverage at junk pricing, FPI semi-annual disclosure, founder control, single-vintage concentration risk), we anchor at 15-20x normalized FY2027E owner earnings.

Translating the comparable set (full table in the Industry section) into implied BSP values, using annualized Q1 2026 adjusted EBITDA of ~$1.25 billion (Adjusted Operating Income plus residual D&A) and $3.74 billion of net debt:

The EBITDA-lens range of ~$11-16 billion brackets our owner-earnings-derived band of $13-16 billion — the two methods agree that fair value sits in the mid-teens billions, and that the reported ~$20 billion target (see Market Context) requires granting BSP the Constellation multiple on its first day as a public company. Note also what the revenue lens says: our band equates to 6.4-7.6x pro forma revenue, a premium to Constellation's 3.9x that is justified arithmetically by BSP's roughly double margins — investors hearing "cheaper than AppLovin at 27x revenue" on the roadshow should remember the relevant comp set trades at 3-6x.

Step 3: Fair-Value Band and Reaction Function

What's Priced In — Reverse-Engineering Candidate Valuations

Working backwards at three illustrative marks (equity value; EV adds ~$3.7B net debt):

• At $11B (the October 2025 private-round mark, T. Rowe Price-led): EV ~$14.7B = 5.6x PF FY25 revenue, 13.5x current-portfolio core operating income, ~13.5x FY27E owner earnings. The market would be paying roll-up multiples for compounder evidence — the gap is the opportunity, and we would be buyers.
• At $15B: EV ~$18.7B = 7.2x PF revenue, ~17x current core operating income, ~18x FY27E owner earnings. The market prices the 2025-26 vintage transforming on schedule, with deployment optionality roughly free. Fair — a Hold with a constructive bias.
• At $20B — the reported target: EV ~$23.7B = 9.1x PF revenue, ~22x current core operating income, ~25x FY27E owner earnings. The market pre-pays for both flawless integration and several years of future deployment at undegraded hurdles, sight unseen, on semi-annual disclosure. That is faith, not analysis.

Sensitivity: Equity Value ($B) — FY27E Owner Earnings × Multiple

The grid spans $8.5-19B — wide, because both axes are genuinely uncertain pre-pricing. Two asymmetries are worth noting. First, the downside is partially floored by asset value: the portfolio was assembled for ~$6.1 billion of disclosed consideration since 2023, AOL alone annualizes $260M+ of pre-tax income, and the franchise of the playbook itself has demonstrated private-market bidders. Second, the upside case requires no heroics on any single input — 21x on $900M is reachable on schedule-keeping alone — which is precisely why the bankers will try to price it there.

Catalysts & Event Path

What They're NOT Saying

1. Per-business P&L — the big one. Ten businesses spanning a 53%-margin legacy portal, breakeven enterprise video, and declining consumer apps are reported as a single segment. Investors cannot see which assets are growing, which are being harvested, and which are failing. The aggregate organic-growth figure (6-13%) is doing an enormous amount of concealing work — Brightcove and Remini are disclosed as declining only in narrative asides.
2. How much of "organic growth" is price. The case studies celebrate ARPU up 50-150% with subscriber counts flat or down; the organic figure is never decomposed into price × volume. Given sub-100% NRR, we infer volume is roughly flat-to-negative portfolio-wide — but the F-1 never says so directly.
3. Tractive's financials. A $900 million acquisition — the second-largest in company history, closed three weeks before filing — appears in no financial statement, no pro forma, and no operating metric. Investors are being asked to underwrite it sight unseen.
4. AOL's search and advertising economics. AOL is 24% of Q1 revenue and its 53% margin depends partly on search-distribution and advertising arrangements whose counterparties, terms, and durations are not disclosed. A renegotiation of a single partner agreement could move consolidated earnings materially.
5. Subscriber counts and gross churn. "Monthly paying customers" (9M+) is explicitly non-de-duplicated and is the only volume metric provided. No per-business subscriber counts, no gross churn, no cohort retention curves — only the blended NRR series.
6. When reorganization expense normalizes. $76 million in Q1 2026 alone, with no guidance on the glide path. The adjusted-earnings story depends on this number collapsing in H2 2026; the F-1 carefully avoids committing to it.
7. Individual purchase prices for half the portfolio. Meetup, StreamYard, Issuu ($281M combined), and Loomly, komoot, MileIQ, Harvest ($701M combined) are disclosed only in baskets; Remini's price is absent entirely. Per-deal return math — the company's core claim — cannot be independently reconstructed.
8. Founder voting percentage post-IPO. Blank pending pricing, as is the identity and size of each Selling Shareholder's stake. Investors will learn who is cashing out — and how much control they're ceding — only in the final amendment.
9. The pre-IPO private marks. The F-1 contains no reference point for what crossover investors (Baillie Gifford et al.) paid or what the last private round implied — context every IPO buyer will want when the range prints against it.
10. AI model dependency economics. The 90%-AI-authored-code claim implies material and growing inference costs (token-based pricing from Anthropic, Google, OpenAI is disclosed) — but AI spend appears nowhere as a quantified cost line, making the durability of the productivity margin unverifiable.

Street Perspective

Debate: Compounder Premium or Roll-Up Discount?

Bull view: This is the closest thing to an early Constellation Software or Danaher the consumer-internet sector has produced: published hurdle rates, founder-led M&A, a 13-year playbook with three audited years of 84% revenue CAGR, and a $400 billion reinvestment universe. Serial acquirers with demonstrated process moats compound for decades and command 20-30x earnings throughout. Buying at the IPO is buying Constellation in 2006.

Bear view: The pattern-match is flattering but the mechanics differ: Constellation buys hundreds of tiny, sticky vertical-market monopolies with 100%+ NRR and zero leverage; BSP buys a handful of large, contested consumer assets with 91-95% NRR, funds them at SOFR+587.5, and books "growth" that is 86% acquisition timing. The closer historical analogues — debt-fueled consumer roll-ups from media to apps — mostly ended with multiple compression once deal flow slowed and the market noticed organic reality. Roll-ups deserve discounts, not premiums, until organic growth proves out.

Our take: The truth is asset-specific. The playbook evidence (Evernote, StreamYard, Remini) clears the bar for "process moat, narrow"; the leverage, vintage concentration, and sub-100 NRR cap the multiple we will pay well below compounder royalty. We'd put BSP between the archetypes: a real operating system for software assets, with a balance sheet and disclosure regime that demand a discount until the AOL/Vimeo/Eventbrite vintage proves the system scales. The market's eventual label — compounder or roll-up — will be decided by two H1 reports, not by the roadshow.

Debate: Is the 51% Adjusted Margin Real?

Bull view: Amortization of acquired intangibles is purchase-accounting fiction; every serial acquirer adjusts it out, and on that basis alone the margin is mid-40s. The trajectory (36% → 45% → 47% → 51%) tracks the disclosed, verifiable operational changes — headcount cuts, infrastructure re-engineering, advertising spend down 600bp. AOL arrives at 53% GAAP-ish margins before Bending Spoons touches it.

Bear view: Reorganization and transaction costs are cash, recurring, and intrinsic to a company that promises perpetual acquisition; capitalizing them into "adjusted" earnings overstates owner earnings by $150M+ annually at the current tempo. OCF conversion of 77% (FY25) and 37% (Q1 26) is the tell — cash doesn't lie.

Our take: Both, precisely. Strip intangible amortization (legitimate), expense a through-cycle deal/reorg allowance (necessary), and normalized operating margin lands in the low-40s with operating earnings power around $1.05-1.1B annualized on the current portfolio — the number our valuation actually uses. The cash-conversion gap should close visibly as the 2025-26 reorganizations complete; if it hasn't by the FY2026 report, the bear reading wins.

Debate: Is AI the Tailwind or the Termite?

Bull view: AI is the structural gift to this exact model: it collapses the cost of operating mature software (90% of code AI-authored, revenue per Spooner up 2.3x in two years), widens the transformation arbitrage on every future deal, and panics sellers into exiting at better prices ("an environment of greater uncertainty could provide opportunities to acquire businesses at more favorable valuations," per the founder letter).

Bear view: AI is termites in the portfolio's foundations: thin-utility consumer apps (photo enhancement, file transfer, mileage logs, note-taking) are precisely what AI-native tools and OS-level features absorb first. Remini — the portfolio's AI showcase — is already declining. Meanwhile the operating edge is rented from the same three model vendors every competitor uses.

Our take: Near term (2-3 years), the bull case is simply true — the cost side compounds today while substitution erodes slowly against 8-year-tenure user bases. Long term, the threat is real and the company knows it, which is why the portfolio is visibly rotating toward assets with structural moats beyond habit: marketplaces (Eventbrite, Meetup), enterprise infrastructure (Brightcove, Vimeo), hardware-attached subscriptions (Tractive), and media franchises (AOL). Watch the acquisition mix — it is the company's own revealed preference about which of its assets AI threatens.

Market Context

Reported Deal Parameters (Press, Not Yet in the Filing)

• Target valuation: ~$20 billion (Italian financial press cites a $20-25 billion range) — approximately 1.8x the October 2025 private mark, eight months later.
• Primary raise: ~$1.5 billion; total offering including Selling Shareholder secondary reportedly above $4 billion (~20% of post-IPO equity).
• Timeline: confidential filing March 13, 2026; public F-1 June 8; institutional book reportedly opening mid-June; listing expected late June or early July 2026.

Private Funding History — The Marks Beneath the Target

The crossover roster (T. Rowe, Fidelity, Baillie Gifford, Durable) is the quality signal bulls will cite; the velocity of the marks is the caution bears will. Note the October 2025 round's $440 million secondary component — insiders have already taken meaningful chips off the table at $11 billion, context for reading the Selling Shareholder line when the final amendment prints. Tamburi Investment Partners' ~3.2% stake (held via StarTIP since 2019) repriced its parent visibly on filing day — Milan, at least, is celebrating.

The IPO Window: Hot Opens, Weak Follow-Through

BSP arrives into a window that rewards branded names on day one and punishes them afterward. Circle opened +168% (and still trades well above its IPO price); Chime popped +37% on debut and now sits roughly 38% below its IPO price; Figma's +250% first day gave way to trading ~44% below issue. The pattern — selective institutional demand, retail-driven pops, then multiple compression as lockups and reality arrive — argues two things for BSP: the print itself may well "work" at the reported $20 billion regardless of our fair-value math, and the 6-18 month after-market is where valuation discipline gets paid. We rate the stock, not the first trade.

The Reputation File

The press record around the playbook is now substantial and adversarial in places: documented post-acquisition workforce reductions far beyond tech norms (WeTransfer ~75%; Vimeo's video team; Filmic in its entirety); the Evernote free-tier restriction and multi-year price escalations that seeded a genre of "zombie apps" commentary; a WeTransfer co-founder publicly criticizing the new ownership and announcing a competing product; an investigative feature from a Dutch outlet treating the acquire-cut-raise cycle as systematic; and a Delaware Chancery suit by Eventbrite stockholders over the take-private's voting mechanics. None of this is undisclosed risk — it is the playbook's externalities, accurately reported. The investment-relevant question is whether accumulated reputation begins to (a) raise churn at acquisition announcements, (b) move sellers toward other buyers at the margin, or (c) attract regulatory attention to subscription practices. We watch all three; the F-1's own NRR-by-business table (Remini 87%) is where the answer will first appear.

Assessment: At the reported ~$20 billion target, the IPO asks public investors to pay 1.8x what the world's most sophisticated crossover funds paid eight months ago — for a company that has since added AOL, Eventbrite, and Tractive, levered up to $4.5 billion, and entered a CFIUS review. Some of that step-up is earned (the AOL deal alone arguably added $2-3 billion of value at the price paid); the rest is IPO-window optimism. Our $13-16 billion fair-value band implies the earned portion, not the optimism.

Thesis Scorecard — Initiating Coverage

Bottom Line

What Would Change Our Mind

Key Signposts