Liblib, what's your dream?

@吴睿睿

A few days ago, Liblib — oh wait, Evoken (Yanyu Technology), announced a new round of funding.

Over the past three years, this AI application company has cycled through at least eight names: Liblib, Lovart, LibTV, Xingliu, Zaoci, Shakker, Evoken... Before Evoken, there was even a transitional parent company name. The profusion of names stems from an equally profusive product lineup. And every time one of those products hit, the company's valuation climbed another rung.

Evoken, officially, evokes creativity. But it sounds an awful lot like "Evoke Token" — a true token awakener.

Yet as one of the top three Chinese-founded AI application companies by valuation (the other two being Manus and Genspark), Liblib may also be the most controversial.

It's already "China's largest AI creation platform," but many people's first association remains token arbitrage: users come here for the fastest, cheapest, most comprehensive access to models. Hence the industry lore that Liblib subsidizes its way to negative gross margins — a figure that should normally sit above 70%.

Even within the investment world, whether you invested or passed is itself a statement of conviction. But by my read, few companies provoke such intense bewilderment, even outright attacks, from those who passed. At one fund that did invest in Liblib, another partner told us: "This company will blow up sooner or later."

That hasn't stopped the chase. This round's post-money valuation exceeds $2 billion — four times its Series B valuation just eight months prior.

These are merely the surface-level contradictions. A more precise framing might be: an AI company that seems to be following the internet playbook; a product company whose every product carries the whiff of imitation; a startup burning cash to compete with giants.

In the AI era, this is a company that is not very AI, yet has succeeded enormously.

To exist is to be justified. We have no intention of putting a startup under the microscope with a magnifying glass, but we are genuinely curious how such a contradictory story came to be. elsewhere spoke with more than a dozen people close to Liblib to try to answer that question.

Given that Evoken hasn't yet achieved any real name recognition, we'll use its original name throughout: Liblib.

A Manufacturing Company

When you think of an AI company, what's the first thing that comes to mind?

Perhaps something like DeepSeek, finding a new path for large models, or Manus, first defining what a general-purpose agent could be. These imaginings likely share a common thread: creation.

And this, precisely, may be what Liblib possesses least of all.

Judging by the company's three most important products today: Liblib is China's version of Civitai; Lovart was inspired by Manus; LibTV and TapNow are pixel-for-pixel similar.

In other words, aside from Lovart as the "world's first creative agent," this is a company with little originality.

When many technical hires joined Liblib, their two most important tasks were: one, ensure API stability; two, integrate new models. Veteran employees would tell them: "Whatever product features you build, none will bring in as many new users as integrating the next new model. Just wait for the next model drop."

The most aggressive department for model integration wasn't product or engineering — it was marketing, racing to push out press releases about being "first to integrate XX model." The two technical teams, based in Beijing and Shanghai, would even compete for marketing's praise based on whose model integration performed better.

A former Liblib employee responsible for model integration told me that every time a new model dropped, Liblib's speed ranked in the top three — sometimes even faster than Dreamina when integrating ByteDance models.

To burn less cash, he added, Liblib scoured extensively for arbitrage API channels. Marketing staff were tasked with sourcing: maybe a major client's bulk protocol package going unused, resold cheap; maybe from specialized suppliers like FAL.

Most frequently cited was their bulk purchase of discounted Nano Banana accounts. To secure better pricing, many of the accounts they procured were student accounts — since Google grants eligible students over a year of premium model credits.

A thousand threads to gather, then the technical team executes: XX channel has a 25% discount, cheaper than our current 20% off, should we migrate some traffic over?

So some say Liblib is essentially a wholesale-to-retail business.

And thus inevitably encounters this industry's common problems: counterfeit goods. Recently Liblib faced significant controversy over adulterated models, with users complaining that cheaper substitutes were being used in place of Nano Banana and Seedance 2.0. A former Liblib technical team member told me the actual issue likely stemmed from arbitrage APIs — often they couldn't trace the source themselves.

There was one incident where a major supplier delivered fakes, sending Chen Mian into a rage, hunting for someone to blame across the company.

Every day, in every work chat, Chen Mian drives this nearly 200-person company: competitors launched a new discount package, we need to match fast; one approach isn't working, pivot 180 degrees immediately; a conversion funnel issue arises, commercialization and R&D both get pulled into overtime. From its founding, this company has operated on a six-day workweek.

When directives go out and aren't advanced same-day, that's too slow. Chen Mian's follow-ups ping in the group. He favors short messages, each less than two seconds apart, popping up to fill the screen: I said this before — why hasn't it changed — just leaving it — letting users see this? Then @A, @B, @C.

Hard to imagine: the operating principle of a top 3 Chinese AI application company seems far from creation, yet close to moving bricks.

It resembles something more like a traditional industry. I once asked a Liblib executive what that might be. After long consideration, three words: manufacturing.

Indeed, manufacturing's underlying code is: product innovation doesn't matter; what matters is cost control, worker management, and supplier relationships.

An employee once remixed a Japanese salaryman meme featuring Chen Mian, red with rage, shouting: At this rate the company will go bankrupt! Many employees loved using it; Chen Mian himself would send it too.

Revolving CTOs

When you think of an AI company, what's your second assumption?

I suspect: technology?

But nearly everyone will tell you — even current Liblib employees will tell you — this is not a technical company.

A former Liblib technical staffer told me that in just three years since founding, Liblib has gone through seven CTOs/technical leads. In his view, Lovart's slowing growth since late last year stemmed directly from the CTO's diminishing authority.

Yet interestingly, from an investor who backed Liblib, the narrative became: (that) CTO wasn't top-tier, has been replaced.

Liblib's breakout moment was Lovart: this company once had an AI-native team, and proved it could build innovative products.

Manus made Chen Mian realize that next-generation products needed to be built on algorithms, so for Lovart, Liblib assembled a more AI-focused technical team in Shanghai: partly from Tencent, ByteDance, and Alibaba's AI labs; partly from Doubao and MiniMax product teams.

With different DNA and distance from headquarters, this team briefly developed a new collaboration model. The traditional "product writes PRD, submits requirements to R&D" structure was dismantled. Ideas no longer flowed only from product; R&D would proactively approach product to find good interaction forms for their technology. The entire team finalized the solution before functional roles took over execution.

A Shanghai-based executive at the time believed this was the most important reason Lovart achieved such good results so quickly.

But soon, the relationship between Beijing and Shanghai teams grew delicate.

Chen Mian and the Shanghai team held diametrically opposed views on "how to build an agent." To the Shanghai team, an agent is a person; what it needs isn't correction of its actions, but exploration of how to manifest its intelligence. The director leading Shanghai's technical team, from Tencent's YouTu Lab, wanted to continue the integrated product-R&D model, encouraging R&D side projects for product innovation.

But what Chen Mian needed most now was stable user delivery, continued revenue growth for Liblib and Lovart. Whether it was an agent or a generator mattered little.

Those who joined Lovart in late 2025 came with aspirations of exploring image AGI. One frontend Twitter legend broke his never-commute rule for three months; another rejected a ByteDance offer, taking a 50% pay cut to join Lovart. The brick-moving work was dirty work to them; product writing PRDs for R&D to implement was regressing from AGI.

Several times, Chen Mian's directives couldn't be executed by this team — unacceptable to him.

In a 70-person group chat, in all-hands emails, Chen Mian attributed Lovart's later performance to "the technical lead's misalignment with my vision," forcing the other party to pledge self-docking their year-end bonus.

Multiple Liblib technical staff told us that, to date, Lovart's original technical team has almost entirely departed.

The replacement came from Doubao, internally codenamed Stone, with many years at Taobao. Upon taking over, his directive to the team was — abandon all side projects, follow orders.

"None of our original team understood technology," a Liblib executive told me, and in areas he doesn't understand, Chen Mian struggles to trust others, making CTOs "very difficult to land."

Some investors believe Chen Mian's iteration speed is remarkable, and money serves as a moat; when the next phase demands stronger technology, he'll certainly recruit a stronger CTO. Some Liblib executives have expressed to us their need for someone like Peak Ji, combining technical and engineering capabilities.

Indeed, Peak's role at Manus is clearly irreplaceable. Only, if Peak had joined Liblib instead, could he have successfully landed?

Sense of Rhythm

But on the other hand, if we look only at growth data, Liblib's report card is genuinely impressive.

Multiple investors who reviewed Liblib's data pack during fundraising told us they "could hardly resist."

The numbers released in this round's publicity are eye-catching: ARR exceeding $300 million; LibTV daily revenue surpassing $1 million; Liblib AI cumulative users over 30 million, Xingliu (Lovart's domestic version) users exceeding 10 million.

But notice: new products announce revenue data; old products talk user numbers.

A person close to Liblib's core team told us that with every new product launch, the company tilts all resources to pump up the numbers, crafting the narrative accordingly. Pretty data always excites investors; the only anxiety is, "What if investors ask why the previous product's numbers dropped?"

Old products withering when money and resources stop flowing — of course that's not right. But this seems to be a rhythm Liblib has gradually mastered: build new product, show new growth data, raise more money. At least for the past three years, Chen Mian has proven this playbook fits the current technology cycle. Before model convergence, all intermediate products are disposable fuel.

In some supporters' framing, this is like surfing: fundamentally about catching waves; painstakingly crafting products risks being swallowed by models.

More than one investor told elsewhere that LibTV's borrowing from TapNow was a "good signal": since "world's first XX" products can't always come from Liblib, let smaller startups validate first, then crush.

"Surpassing the original in a week or two — then when Chen Mian wants to enter other markets in the future, he'll be the winner too."

If money is key to improving the odds of the next surf, he should have more. An investor who described himself as "lucky enough to get it right" told us: "Occupying this multimodal stronghold, getting more money, more attention, recruiting better people, leaving nothing around you — that's how you hope to survive to the end."

Internal sources told us Chen Mian also considered acquiring TapNow.

But note: Manus and Liblib are quite different. Manus grows with cleverness, exits quickly while ahead — guerrilla warfare against giants. Liblib burns cash for growth, fighting giants and small companies alike. In a sense, you could say: Liblib itself is a giant — just relative to smaller companies.

Late last year, Liblib began more aggressively recruiting ByteDance senior staff, offering compensation matching ByteDance's levels. The company expanded rapidly from 40-plus to nearly 200 employees, adding over a dozen monthly. Like ByteDance, hire massively, cull swiftly.

Awakened by Manus

Many call Liblib a club deal. That's not quite fair. Until late October last year, Chen Mian was hardly a star CEO in primary markets.

In 2023, Liblib's first round valued the company at just $18 million. For Gaorong Ventures, Source Code Capital, and GSR Ventures, there was little to quibble about at that price. Chen Mian later complained many times: he didn't understand fundraising then, letting three funds enter simultaneously at low valuation.

The second round was harder. Money was nearly exhausted by subsidy wars, and the product had been taken down. Baidu's strategic investment threw a lifeline — but with conditions: only $5 million, and Liblib had to partner with Baidu's mobile app.

A Liblib early executive once told us that some investors even told them to "just shut it down."

Sequoia, Hillhouse, IDG — all looked repeatedly, passed repeatedly. Sequoia had invested in "Toast," founded by Shen Zhenyu, a Liblib competitor; they looked every round, never pulled the trigger until Lovart emerged. IDG internally pushed Liblib six times, ultimately still passed.

Many investors then believed model aggregation platforms were too thin a model; heavy subsidies and community building weren't worth much. The aesthetic then favored big-company senior PM founders like Leon Ming and Zhang Yueguang, not commercially-oriented founders like Chen Mian.

Chen Mian originally started this venture inspired by CapCut, seeing images as a commercially valuable space with ready user awareness, where he wouldn't have to clash head-on with giants. Simply put, a lane for quiet money-making.

Until Manus appeared.

Chen Mian has expressed his shared vision with Manus on various occasions. Multiple core Liblib members told me that around late 2024, Liblib had been internally conceptualizing a more autonomous form — but hadn't landed on "agent."

Many interpret Manus's success as a "timing gap strategy": at the critical inflection point of underlying model capabilities, wrapping it with the fastest engineering into a well-executed concrete product.

Thus, product capability matters less; strategic judgment and execution matter immensely.

On Lovart, Chen Mian indeed demonstrated opportunistic agility and Manus-like qualities: also began preparation a year ahead; strong execution — hearing gpt-image-1 would launch in three weeks, immediately pulled the company's best people to Shanghai for closed development; Manus emailed Silicon Valley influencers, Lovart got liked by Elon Musk; also cleverly rode the GPT-4o Ghibli trend, even with product imperfections and some hardcoded capabilities, determined to seize "world's first creative agent" mindshare.

In May 2025, Lovart opened beta; the waitlist exceeded 100,000 in five days — the company's first viral product. Perhaps because it faced overseas markets, Lovart wasn't folded into the current fundraising entity Yanyu Group.

But Lovart was indeed the inflection point that turned this company's fortunes.

"After that he understood the rules of this game," said the aforementioned investor who "got it right." "Many founders still haven't understood today; some who understood can't execute."

Three months after Lovart's official launch, Sequoia finally had another partner push Liblib through investment committee, co-leading $130 million with CMC and Ant Group. Liblib's valuation leaped from $90 million to $500 million.

An investor at another firm even listed it as his "2024 biggest regret." "If I'd known they were building Lovart, definitely would have invested."

Yet some investors still passed after Lovart. Alibaba, reportedly, didn't bite at this round's $800 million valuation, because Liblib refused to disclose its ESOP.

Mian and Win

Chen Mian's English name is Malvin — the first half phonetically echoes "Mian," while the second half happens to be "win" in English. Chinese means win (Mian's meaning), English also means win — win upon win.

But in Chen Mian's decade-long career before this, he hadn't experienced true winning.

Born 1992, entered the workforce in 2014, Chen Mian's résumé includes Tencent, 360, Baidu, DiDi, Mobike, Missfresh, ByteDance. He may have worked at the most major tech companies of any founder in this AI wave. Yet for long stretches, he seemed ill-timed — joining these famous companies either after core businesses had wound down, or before he was ready to lead.

Until 2020. At ByteDance working on Guaguolong, Chen Mian fought his "most satisfying" battle. Guaguolong's playbook was simple: blanket advertising, lowest prices. But soon, education's "double reduction" policy arrived. Transferred to CapCut, Chen Mian never entered the core power circle.

Many see Liblib as similar to Genspark: scarce originality, but with its own strengths in growth and commercialization. But Chen Mian hates such comparisons; the person he aspires to be — Yiming Zhang.

Call it ambition, call it partly trauma. His friends, his investors, have received late-night calls — sometimes anxious, sometimes weeping.

This chance to win is too precious for Chen Mian.

To this day, most co-founders who followed Chen Mian into entrepreneurship have left him, willingly or not. When they disagreed, Chen Mian would emphasize: "You need to be brain-synced with me." Brain-synced, roughly: think like me.

This March, Roi — who left Liblib to start his own venture — announced a new funding round, calling himself "Liblib co-founder." Chen Mian posted that he had "no true co-founders." This shocked those still at the company. Someone once showed Chen Mian the original investor materials, clearly listing co-founders' names. Roi among them.

Chen Mian said (paraphrased): that was then, this is now.

At this company known for high turnover, the shortest tenure I know of — four days. This former ByteDance 3-2 was recently poached; on day four, finding the work mismatched what was promised, had a major blowup with Chen Mian. That afternoon, her Lark was wiped clean.

Over three years, Liblib may offer a growth miracle for an AI company. Yet perhaps because it's not AGI enough, because of the apparent lack of idealism, it hasn't yet earned the most applause.

But the growth Chen Mian believes in may have more silent, devout followers. In an atmosphere of technology-first, innovation-first, Liblib and Chen Mian continue applying a mobile internet success playbook; while VCs proclaim they invest in AI-native talent, they direct the most money to Chen Mian.

Many founders at smaller companies have told me that AI applications have no core technology to speak of, models are breathing down their necks, and Chen Mian building something this large on such thin ice already commands their respect.

As for after model convergence, perhaps better technical teams will be needed, or perhaps it's a new game altogether. Who knows? But you have to survive first.

This is why investors choose them.

As Liblib announced this funding round, I'm told its next round — at $3 billion valuation — is already underway.

Cover image: Paul Delaroche, The Execution of Lady Jane Grey, 1848, Louvre-Lens / Walker Art Gallery