From Dependency to Independence: Valuation Soars to Hundreds of Billions

Written by / Meng Huiyuan

Edited by / Li Jinlin

Layout by / Annalee

In June 2026, a funding announcement sent shockwaves through the venture capital community: Kuaishou's AI video generation business, Kling AI, officially launched its first funding round post-spin-off, with a pre-money valuation of $18 billion (approximately RMB 122 billion).

How exaggeration (extraordinary) is this figure? As of early June, Kuaishou's parent company had a market capitalization of around $27 billion, meaning Kling, a subsidiary, was valued at approximately 67% of the entire Kuaishou.

A business established just two years ago, with a valuation nearing 70% of its parent company's market cap, is akin to dropping a bombshell both within and outside the industry.

Crucially, Kling is not an isolated case.

ByteDance's Doubao AI is quietly laying the groundwork through an "Independent Equity Incentive Plan," while Alibaba has merged its Tongyi Large Model Business Unit and Future Living Lab to form the Token Foundry Business Unit. Baidu's Kunlunxin has submitted an IPO application to the Hong Kong Stock Exchange... Clearly, a wave of AI asset spin-offs is underway, transitioning from the "safety of major company incubators" to "independent capitalization."

But why are major companies rushing to spin off their AI businesses? Does the video generation sector offer room for independent listings and growth?

Costs, Valuations, and Strategy: The Three Hard Calculations Behind Major Companies' AI Spin-offs

Undoubtedly, the rate at which AI businesses burn through cash is challenging traditional internet companies' financial statements.

Take Kuaishou as an example. Its Q1 2026 financial report showed total revenue of RMB 33.7 billion, up 3.4% year-on-year; adjusted net profit was RMB 3.4 billion, a significant 26.3% decline year-on-year. Meanwhile, Kuaishou's full-year 2026 capital expenditures are expected to reach RMB 26 billion, an increase of approximately RMB 11 billion from 2025, with nearly all of the increment (increase) going toward AI computing power, with Kling being the top priority. Q1 research and development expenditures reached RMB 3.6 billion, up 9.8% year-on-year, with the financial report explicitly attributing this to increased AI investments.

The issue lies in the fact that Kling's commercial returns are far from keeping pace with the growth in costs.

In Q1 2026, Kling AI generated over RMB 650 million in revenue, with an ARR (Annualized Revenue Run Rate) approaching $500 million (approximately RMB 3.6 billion), but its costs are in the tens of billions.

Huang Lichong, President of Universal Capital International, stated in a media interview that spinning off Kling essentially transforms the AI business from a "cost center sustained by the parent company" into an "independent, market-driven AI technology company with self-sustaining financing," using external capital to shoulder the high costs of video large models.

ByteDance, a competitor, also faces escalating computing power costs.

Its Doubao large model's daily Token usage has surpassed 120 trillion, with its B-side Volcano Engine MaaS capturing a 49.5% market share. However, Doubao only launched its first tiered subscription plans in May 2026 (RMB 68/month, RMB 200/month, RMB 500/month), meaning commercialization is still in its infancy.

Source: IDC's "Latest Analysis Report on the Chinese Enterprise-Level MaaS Market"

Industry insiders describe the situation as follows: "The more popular Doubao becomes, the more it's called upon, the greater the computing power consumption, and the deeper the losses. Free is a traffic strategy, but free is also killing cash flow."

Beyond cost pressures, the greater allure lies in valuation unlocking.

It's understood that Morgan Stanley previously valued Kling at just $6 billion, but with the spin-off plan advancing and market valuation logic for pure AI companies changing, the Pre-IPO round valuation of $18 billion represents a threefold increase for the same assets.

The core reason these AI businesses command high valuations post-spin-off is that, when consolidated in major companies' financial statements, they are seen as profit-draining cost centers. Once independent, they are revalued based on sector scarcity, revenue growth, and market potential, with price-to-sales ratios of several dozen times not uncommon.

Of course, major companies have deeper strategic considerations for spinning off AI assets.

One aspect is opening up capital exit pathways. ByteDance's overall valuation is around $550 billion, but its path to a full listing has long been blocked, leaving shareholders facing various forms of exit constraints outside public markets. Against this backdrop, ByteDance launched the "Doubao Independent Equity Incentive Plan," granting stock options linked to Doubao's performance to employees in the Seed department, establishing a valuation anchor and organizational foundation. While ByteDance currently states it has "no plans to spin off" its AI4S business, the implementation of the independent equity plan has laid the groundwork for Doubao's future independent capitalization.

The other aspect is achieving independent breakthroughs in business models. On June 8, Alibaba established the Token Foundry Business Unit, led personally by CEO Wu Yongming, integrating star products like Happy Horse and Happy Oyster with the Tongyi Large Model to accelerate AI commercialization. Spinning off AI businesses from major companies helps get rid of (break free from) the constraints of internal resource allocation and pricing systems within the group, enabling rapid growth under independent business models.

Can Technological Iteration Outpace the Growth in Computing Costs?

If spinning off is a rational choice after major companies have crunched the numbers, then for AI products like Kling, the prerequisite for completing independent financing and even an IPO is whether the unit economic model of this business can succeed.

In terms of absolute growth, Kling has delivered impressive results. Over two years since its launch, Kling's quarterly revenue surged from RMB 150 million in Q1 2025 to RMB 650 million in Q1 2026, a year-on-year increase of over 300%. Its ARR rose from $100 million in March last year to nearly $500 million in March this year, a fourfold increase in one year. The company has raised its full-year 2026 ARR guidance from $300 million to $500 million.

Source: Kuaishou's official Weibo account

In terms of revenue structure, approximately 70% of Kling's revenue comes from B-side APIs and enterprise services, with overseas markets contributing 70%. As of the end of Q1, Kling had over 60 million global users, serving more than 30,000 enterprise clients and developers. From an application scenario perspective, using AI to produce micro-short dramas can reduce costs to below one-third of traditional levels and shorten production cycles by over 60%. In the film and television production sector, the TV drama "Peaceful Years" was able to compress a two-month production task to less than two weeks after using Kling.

However, compared to the global AI video sector, Kling's revenue scale is still in its early stages.

Overseas video generation leader Runway is valued at approximately $5.3 billion, with an actual annualized ARR of around $150 million as of mid-2026 and a full-year target of $265 million. Kling's high valuation means capital markets are offering it a multiple of Runway's premium, reflecting expectations of scarcity as the "first global AI video stock," but also imposing higher validation requirements on its unit economic model.

Because on the other side of revenue growth is the rigid rise in computing costs. Video generation inference computing power demand is approximately 2,000 times that of traditional text large models, with the physical cost of generating a 1080P video ranging from $0.5 to $1.

In contrast, the lesson of Sora's shutdown is stark: Sora's cumulative total revenue from in-app purchases was only about $2.1 million (approximately RMB 14.51 million), but its daily computing power costs to maintain operations were as high as $15 million, meaning its cumulative earnings couldn't even cover a single day's computing expenses at its peak.

While Kuaishou's situation is better than Sora's, structural cost pressures remain significant.

Of the projected RMB 26 billion in capital expenditures for 2026, the vast majority will go toward Kling's computing power procurement and model iteration. Although Kling has achieved positive marginal profits at the inference level, with its ARR growing fourfold in a year, the parent company's capital expenditures have also doubled over the same period.

In response, He Jinghang, Director of Corporate Ratings for Fitch Ratings Asia Pacific, remains cautious: "Fitch does not currently expect Kling to make a very significant contribution to Kuaishou's profitability or cash flow in the coming years."

A deeper issue is that Kling achieved a fundamental reversal in its revenue structure in Q1 2026, with B-side APIs surging to around 60% and P-side professional user subscriptions dropping to about 40%. The core risk points in its unit economic model have shifted accordingly, from "declining C-side payment willingness" to "price competition pressure on B-side API calls" and "scalability in managing computing power costs."

The true meaning of a "commercialization inflection point" is not how fast ARR grows, but whether the unit economic model holds.

For Kling, this means whether the revenue generated from a single video generation can cover its computing power costs and achieve scalable profitability on that basis. The problem it faces is whether computing power costs, which rise with usage, can be outpaced by the unit cost reductions brought about by technological iteration.

This is a dilemma.

If video generation tools are priced far above user expectations (e.g., users expect a single price below $0.1, while actual generation costs are $0.5–$1), payment conversion will be constrained.

If pricing is too low, computing power costs will erode gross margins, with scale leading to deeper losses.

The way out depends on three variables: first, whether model efficiency optimizations can reduce single-generation costs to below $0.1; second, whether the ARPU (average revenue per user) of professional creators and enterprise clients can continue to rise to cover costs; and third, whether the price of computing hardware can continue to decline.

Leading securities firm research reports generally estimate that if inference chip costs fall by 40% annually and model compression technologies continue to iterate, video generation unit costs could drop by about 60% over the next 18–24 months.

However, all three paths currently carry significant uncertainty.

The lesson of OpenAI shutting down Sora has shown that video generation projects relying solely on technological breakthroughs without a commercial realization logic are unlikely to succeed.

OpenAI's notice regarding the suspension of Sora services

With Kuaishou's content ecosystem and commercialization team, Kling has natural advantages in customer acquisition, conversion, and retention, but whether this "short-video DNA transplantation" logic can persist after independent operations remains to be verified.

It's worth noting that if Kling ultimately succeeds in an independent listing, it will explore a viable path for independent capitalization of high-quality AI assets from major Chinese internet companies.

Coincidentally, competitors are making similar preparations: ByteDance's Doubao has completed institutional groundwork through the Independent Equity Incentive Plan, Alibaba has made organizational preparations by establishing the Token Foundry Business Unit, and Baidu's Kunlunxin has entered the tutoring (coaching) stage for the STAR Market...

However, whether this path can truly succeed depends not on how high the pre-money valuation is, but on whether, after each funding round and even post-listing, the capital markets' ongoing pricing logic aligns with the company's actual operating performance.

After all, as "Klings" embark on the path of independent financing, the market will no longer just ask about ARR growth but will constantly pose a more fundamental question: When will they achieve profitability?