AI enables each person to generate $2.57 million annually, with revenue tripling in two years! "Startup version of PE" Bending Spoons prepares for IPO

Bending Spoons applies for a U.S. stock IPO, with an estimated valuation of $22 billion. The company has acquired software like Evernote and others, transforming through AI and layoffs, resulting in a 3.4-fold revenue increase in two years.

This summer’s IPO market is very lively, with SpaceX, Anthropic, and OpenAI lining up one after another, but perhaps few notice that a company from Milan, Italy, with 500 million global monthly active users, quietly filed an F-1 registration statement with the U.S. Securities and Exchange Commission (SEC) in June.

This company is called Bending Spoons, named after the scene in the 1999 movie "The Matrix" where a spoon is bent with the mind. It’s okay if you haven’t heard of this company, but you’ve probably used its products: Evernote, Vimeo, WeTransfer, Remini, Eventbrite, StreamYard. Any of these are digital tools used daily by millions.

According to Bloomberg, Bending Spoons aims for a valuation of $20 billion to $22 billion (about NT$640 billion to NT$704 billion) in its IPO, with plans to list on NASDAQ under the ticker BSP, with underwriters including Goldman Sachs, Morgan Stanley, and Allen & Co.

What kind of company is Bending Spoons?

Bending Spoons was founded in 2013 in Milan by five Italian students, including co-founders and co-CEOs Luca Ferrari and Francesco Patarnello. They had previously co-founded a company that failed due to lack of funding, but this time they changed their approach.

Image source: Matteo Danieli’s personal LinkedIn, Bending Spoons co-founder Matteo Danieli

Luca Ferrari stated that relying on “doing whatever the market lacks” was easily influenced by external factors, so they chose to let the market filter products, only taking over digital assets that already had a verified user base and were willing to sell, then improving them further.

“Acquiring digital businesses, deeply transforming and continuously optimizing to expand profits, then reinvesting those profits into new acquisitions—this cycle compounds,” is how Bending Spoons describes itself in the F-1 registration.

In simple terms, Bending Spoons is somewhat like a “startup version of private equity (PE),” with the difference that traditional PE usually exits after restructuring for arbitrage, while Bending Spoons aims for permanent holding and compound reinvestment.

All the familiar brands in Taiwan—Evernote, WeTransfer, Remini, Vimeo, StreamYard, Eventbrite—are under its umbrella. As of the IPO application, Bending Spoons has completed over 50 acquisitions.

Image source: Evernote official website

Revenue tripled in two years

This acquisition strategy appears to be working quite smoothly so far.

According to Bending Spoons’ F-1 filing, revenue skyrocketed from $387 million in 2023 to $1.31 billion in 2025, with a compound annual growth rate (CAGR) of 84% from 2023 to 2025. In the first quarter of 2026, quarterly revenue reached $601 million (about NT$12.4B), a 132% increase from the same period last year.

Profitability is equally impressive. In 2025, adjusted operating profit margin reached 47%, rising further to 51% in Q1 2026. The same quarter’s net profit was $27.5 million (about NT$880 million), compared to a loss of $112 million in the same period last year.

In terms of user scale, monthly active users grew from 111 million in December 2023 to over 500 million in March 2026, while paying monthly customers increased from 3 million to 9 million. Revenue is highly subscription-based, with subscription income accounting for 84% of revenue in Q1 2026.

The most impressive figure: in 2025, each core full-time employee (called “Spooner” internally) contributed $2.57 million in revenue, which is of course related to AI.

90% of code contributions involve AI

As of the end of Q1 2026, over 90% of code submissions (Pull Requests) within the company are written or co-written by AI, with about 70% generated entirely by AI independently.

For Bending Spoons, AI’s value is concentrated in three specific scenarios.

• First is legacy code migration. After each acquisition, they transfer the acquired company’s years of old, outdated code onto Bending Spoons’ own technical infrastructure, including data infrastructure (Pico, Lumen, Abacus systems), payment management system (Juno), and more. These repetitive, low-creativity tasks are largely replaced by AI, reducing engineers’ labor.
• Second is rapid commercial iteration. The company’s self-developed A/B testing frameworks (Janus and Orion systems) enable teams to quickly generate test versions and continuously optimize conversion rates. The time from hypothesis to validation is significantly shortened thanks to AI assistance.
• Third is user lifetime value (LTV) prediction. The company’s self-developed AI model Minerva predicts each user’s LTV, directly influencing marketing budget allocation, allowing Bending Spoons to maximize ROI on every marketing dollar with limited resources.

Thanks to AI, the company maintains a flat management structure—no more than three layers—and received about 800k job applications in 2025, ultimately hiring 286 people, with an acceptance rate below 0.04%.

Layoffs are part of the business model, not a side effect

Beyond AI’s “open-source” contribution to revenue, Bending Spoons’ cost-cutting policies deserve more discussion.

The company openly admits in the F-1 filing that layoffs are part of their approach: After acquiring Evernote, the dedicated team shrank from 341 to 60 people within a year, an 82% reduction; StreamYard’s dedicated team shrank by 71%.

This pattern is almost applied to every acquisition: buy, cut staff, migrate onto their own tech infrastructure, run business tests with AI, and wait for the numbers to improve.

Risks for Bending Spoons: high leverage and AI as a double-edged sword

However, AI remains a double-edged sword, and Bending Spoons’ operating model carries significant risks.

While AI has drastically lowered labor costs, its products like Evernote and WeTransfer face the threat of functionality erosion from general AI models like ChatGPT. If users find that general AI can now meet needs that previously required these tools, paid retention rates could face structural pressure.

Another risk stems from the operational model itself. Bending Spoons mainly operates on borrowed money, currently carrying $4.36 billion (about NT$139.5 billion) in debt, with nearly every acquisition financed through borrowing. As long as interest rates stay stable and the market continues to offer cheap, attractive targets, the flywheel keeps spinning.

But if borrowing costs rise or no more worthwhile companies can be found, the entire model could stall, because it’s not designed to slow down.

In its F-1 filing, Bending Spoons has identified over 1,000 potential acquisition targets, estimated to total nearly $400 billion in annual revenue. This asset pool sounds large, but in an era where AI is rewriting the valuation of every software category, the line between “selling quickly before being eliminated by the market” and “long-term holding and optimizing” is rapidly blurring.

This article is reprinted with permission from: “Startup Gathering”
Original title: “AI makes each person generate $2.57 million annually, revenue doubles in two years, ‘startup PE’ Bending Spoons prepares for listing”
Original author: Zeng Linghuai