【Tide Business Review/Text】

Liu Wei only remembered she still had an iQIYI membership when she received the deduction notification. "I must have signed up for it last year to watch some show. I haven't watched in ages; I'd better cancel it."

Just as she was about to cancel, her mother leaned over with her phone. "The show you mentioned wanting to rewatch is available on Hongguo. Check it out—it's free."

Liu Wei glanced over—no ads, no membership pop-ups, and automatic playback for the next episode with a swipe. "Wow, really. I wonder if they'll start charging later," she mused.

On the same day Liu Wei made this comment, late in the night on May 4, Li Liang, Vice President of ByteDance Group, responded on social media to rumors that "Hongguo will soon implement full paid subscriptions."

Li stated, "There will be no mandatory fees, nor will we alter the core model of free drama viewing for users." However, to enrich the platform's content ecosystem, Hongguo has begun introducing select long-form film and television (TV/film) content, with a small portion requiring VIP access as per copyright holder requirements.

A platform that started with free short dramas is now introducing long-form film and television content and establishing a membership system. This indicates that short drama platforms' encroachment on long-form video territory has evolved from competing for user attention to direct business model competition.

No wonder iQIYI is rushing headlong into AI.

At iQIYI's World Conference on April 20, founder Gong Yu announced that 117 celebrities had joined the Nado Pro AI Talent Pool, claiming, "In the future, 100% live-action productions may become intangible cultural heritage." The statement was immediately refuted by multiple celebrity studios.

iQIYI later clarified that "joining the talent pool" merely indicated celebrities' interest in AI film and television projects, with specific collaborations still requiring separate negotiations and licensing. The announcement contained "corporate copywriting errors."

As short drama platforms begin to "serve full courses" and legacy long-form platforms paint futures with AI, the boundaries of competition are blurring, leaving "veterans" like iQIYI struggling to pivot.

01 AI: The Common "Escape Route" for Long-Form Platforms?

iQIYI's urgent shift toward AI narratives reflects sustained pressures on the long-form video industry's business logic.

The long-form model is essentially a game of "expectations": platforms continuously produce content to expand their copyright libraries, attracting long-term user subscriptions. After content costs are amortized, membership fees become profit.

In practice, however, the model resembles "monthly movie tickets"—users often subscribe for one or two shows and leave once they finish.

In 2023, *The Knockout* drove iQIYI's membership revenue up 24% in a single quarter, only to see it decline after the show ended—a typical (textbook) example.

When long-form platforms fail to deliver hit content, financial pressures mount rapidly.

iQIYI's financials show 2025 revenue at RMB 27.29 billion, down 7% year-on-year—the second consecutive annual decline. Net loss attributable to parents reached RMB 206 million, a stark contrast to the RMB 764 million profit the previous year. Membership revenue fell from RMB 20.3 billion in 2023 to RMB 16.81 billion in 2025. As of May 2026, iQIYI's stock hovers around $1.2, with market cap down over 95% from its 2021 peak.

But revenue pressures are just the surface issue.

The deeper challenge for iQIYI and peers is the structural reallocation of user time.

Data from the *China Network Audio-Visual Development Research Report (2026)* shows that as of December 2025, daily per capita usage of micro-short dramas reached 129 minutes, up 28.4% year-on-year, surpassing long-form videos to rank second among audio-visual apps.

QuestMobile data similarly reveals that in 2025, short drama users averaged 118 minutes daily, exceeding long-form's 109 minutes for the first time. Long-form's share of total user time fell from 17.8% in 2023 to 11.3% in 2025, while short dramas rose from 2.7% to 10.8%—nearly equal.

This isn't just another "long-form is dying" refrain.

The critical point is that short dramas aren't just capturing fragment (fragmented) time—they're replacing the dedicated entertainment blocks once reserved for long-form content.

With user growth nearing its ceiling (QuestMobile notes just 19.04 million net monthly active user growth in February 2026, a 1.5% YoY increase), every minute short dramas gain means long-form loses.

These three pressures—user attrition impacting membership and ad revenue, financial strain forcing tighter content budgets, and reduced spending affecting quality—create a vicious cycle that further erodes user loyalty.

Trapped in this dilemma, long-form platforms are collectively betting on AI.

iQIYI initiated its Hong Kong listing process in late 2025 and submitted its application to the HKEX on March 30, 2026. With its U.S. stock languishing and financing windows narrowing, iQIYI needs a new narrative to reengage capital markets—and AI fits the bill.

02 Same AI, Different Strategies

iQIYI's aggressive AI push isn't unique in the industry.

In the first half of 2026, major long-form platforms simultaneously doubled down on AI, though their approaches reveal divergent understandings of "what problems AI should solve."

For example, Tencent Video emphasizes ecosystem building. At the April 15 Network Audio-Visual Conference, Tencent Vice President Sun Zhonghui noted that 5–20 person "creative teams" now leverage AI to accomplish workloads that once required hundreds, leading to exponential content growth and market stratification.

Tencent's core strategy is to bundle AI toolchains, IP licensing, and new revenue-sharing mechanisms for external creators, empowering small teams to become independent production units. The platform's role shifts from content organizer to tool provider and creator ecosystem operator.

Youku, meanwhile, focuses on integrating technology into production systems. Youku President Wu Qian defined AI at the same event as a "productivity system" rather than a standalone tool, emphasizing "art + technology" dual drivers. AI is embedded throughout content production, while a talent pipeline combining "writing, directing, acting" is built in parallel.

In Wu's view, "the AI era's scarcest resource isn't technological tools but super-creators who master both creativity and technology"—essentially arguing that Youku's AI strategy starts with talent development, binding technical and creative capabilities from the ground up.

Mango TV, conversely, prioritizes organizational reform. In February 2026, Mango Supermedia upgraded its AIGC Innovation Content Center to an independent top-tier division, on par with its Program Production Center, focusing on virtual events, AI live-action dramas, AI animated series, and other monetizable content.

When comparing the three platforms, similarities outweigh differences: all position AI as an enabler for creators. Tencent builds ecosystems, Youku develops talent systems, and Mango restructures organizations—but all aim to make AI a force multiplier for creators, not a replacement.

iQIYI's approach at the April 20 conference, however, took a more controversial path. Its "AI Talent Pool" initiative, in industry eyes, wasn't merely a technological upgrade but a reallocation of power—actors shifted from collaborative partners in the creative chain to data assets callable by the platform.

This insecurity quickly festered.

Traditional film and television institutions have yet to fully adapt to AI-era technological systems, and the industry is still learning to coexist with AI. iQIYI's AI Talent Pool, however, let technology outpace rules and consensus.

03 AI "Can't Save" Long-Form Videos

Currently, AI does boost efficiency in script generation, scene previsualization, post-production editing, and special effects rendering. For high-stakes, long-cycle premium long-form content, any risk-reducing, cost-cutting innovation is worth exploring—a logically sound industry approach.

But the question remains: Does efficiency guarantee user buy-in?

In late 2025, Amazon Prime Video introduced AI-generated English dubs for two popular anime series, only to face severe backlash. Viewers criticized the stiff, emotionless voices that often mismatched the visuals, "like a glitched text-to-speech tool." After public outcry, Amazon quietly removed the AI dubs. Voice actor Daman Mills called it "a massive insult to all performers."

iQIYI's "AI Talent Pool" controversy stems from the same issue: when platforms package actors' likenesses, voices, and performance styles as "callable data assets," they overlook a fundamental truth—audiences don't love actors for their data parameters but for the fleeting, human moments they deliver. Feng Yuanzheng put it bluntly: "AI actors' tears are painted on, but mine flow from within—they have warmth, taste."

As users drown in a deluge of daily content, the threshold for creating top-tier premium content only rises. AI may help produce more content, but the truly valuable pieces become even harder to make. Those human moments—a hesitant glance, a subtle smile—AI can imitate but never replicate.

This explains why Youku President Wu Qian emphasized, "Embracing AI isn't about replacement but enablement."

For long-form platforms, AI integration is inevitable. But as they chase AI efficiency, safeguarding content value remains paramount.

For audiences, whether a show is worth their time is the only currency that matters.

Take Liu Wei: she doesn't care about industry grand narratives. She only knows she didn't renew her membership for the show she wanted to finish. She'd already seen the ending on Xiaohongshu—a three-minute recap that was "quite entertaining."

"Saves time," she said.