Key Challenges for Sora 2: How to Monetize Innovation and Creativity?

Produced by | Business Show

On the inaugural day of the National Day celebrations, OpenAI revolutionized the video landscape with the official launch of Sora 2. Industry insiders are proclaiming, 'The GPT-3.5 moment for AI video has arrived.'

This time around, it's not merely a model upgrade; OpenAI has also introduced a standalone app to the forefront. Some speculate that this could evolve into the 'AI version of TikTok,' yet the complexities beneath the surface run much deeper.

Today, let's delve into the ChatGPT moment for AI video from a commercial and product logic standpoint, scrutinizing where its technology and product truly excel. Once the product is crafted, commercialization becomes imperative. So, where do its commercial challenges lie?

Firstly, let's examine its most striking technological breakthrough. This time, it's genuinely 'explosive,' marking a generational leap in both technology and product. The initial Sora was merely 'watchable,' but this iteration has evolved to 'performable,' nearing the effect of a 'world simulator.'

Previously, AI-generated videos often fabricated physical logic, such as a basketball shot defying angles and miraculously sinking into the basket.

Now, behold Sora 2. It can generate Olympic-level gymnastics moves, accurately simulate buoyancy and the physical laws governing object rigidity in 'stand-up paddle board flips,' and create scenes like 'a cat gripping an object tightly and executing a three-and-a-half-rotation jump.'

It adheres strictly to real mechanics and can even simulate 'plausible failures,' such as 'a basketball shot missing and bouncing off the rim.'

In Sora 2, if an athlete misses a shot, the basketball will realistically bounce off the backboard. Interestingly, the 'mistakes' made by Sora 2 often resemble those of an 'agent' implicitly simulating within it; while not flawless, it has made significant strides in adhering to physical laws compared to its predecessors.

Indeed, for any practical 'world simulator,' this is a crucial capability. You must simulate not only 'success' but also 'failure.'

More critically, synchronized audio and video have finally been achieved. When characters switch languages, their lip movements sync perfectly, and even ambient sounds like wind rustling and the clicking of a cat's paws on a keyboard are precisely matched to the visuals. Content that previously necessitated professional post-production can now be generated with a single click by inputting text prompts.

Even when transitioning between multiple camera angles, details like sunlight angles and water stains on the ground remain consistent, eliminating any discrepancies. In the officially released videos, character expressions, lighting, and physical rules are nearly impeccable, making it challenging to distinguish between reality and simulation.

Whether it's Olympic-level gymnastics, cyberpunk scenes, or dialogues with sound effects, everything can be generated with a single click, reducing the technical barrier to entry to its lowest point.

However, in the commercial realm, no matter how impressive the technology, it ultimately boils down to a pivotal question: How to commercialize it for consumer (C-end) users?

Undoubtedly, on the path to becoming a 'universal simulator' and an 'AI system capable of operating in the physical world,' people can derive immense enjoyment from the models built along the way. On the flip side, it's about how to monetize this experience.

From the currently available information, Sora's commercialization path has started to take shape, roughly divided into three major directions: C-end payments, business-to-business (B-end) services, and ecosystem monetization.

Firstly, C-end payments represent its most direct monetization starting point. The Sora App initially adopted an invitation-only system, primarily offering free usage. However, ChatGPT Pro subscribers can access the higher-quality 'Sora 2 Pro' model through sora.com and in-app channels.

Several months ago, the Sora team began internally testing the 'upload your own image' feature, which has been widely embraced.

In this application, you can create content, remix content generated by others, discover new videos in a customizable Sora feed, and even insert yourself or friends into video scenes using the 'Cameos' feature.

With the 'Cameos' feature, you only need to complete a brief audio and video recording within the app (used to verify your identity and capture your appearance and voice characteristics) to seamlessly integrate yourself into any Sora-generated scene.

However, it's particularly noteworthy that Sora previously employed a points-based payment system but has since abolished it. Instead, paying users can enjoy unlimited video generation services, and professional users can receive high-priority generation experiences, thereby reducing creative constraints and boosting user payment willingness.

This tiered strategy of 'basic free + premium paid' indeed lowers the barrier to entry for ordinary users while filtering out high-value users through professional features, continuing ChatGPT's payment conversion experience.

Next up are B-end industry services, which appear to offer a more imaginative growth trajectory. Sora 2's technological capabilities have demonstrated cross-domain adaptability, with clear application scenarios in multiple industries such as film and television, education, and real estate.

For instance, in the film and television industry, it can swiftly generate scene visual sketches or animation sequences based on scripts to aid in pre-production; in the education sector, it can create interactive science popularization and historical reenactment videos to make knowledge transmission more engaging; in the real estate industry, it can create virtual showrooms for potential buyers to experience unbuilt projects immersively.

If, in the future, OpenAI opens API interfaces, it can also provide customized solutions for enterprises, securing stable revenue from tool licensing.

However, this field will also encounter competitive pressures. Once vertical domains launch commercially viable products with favorable performance-to-price ratios based on existing data, if the data training costs for scenario-based models are sufficiently low, it will impact Sora's high-end market.

Finally, let's explore Sora's unique differentiating factor, which is the imaginative space for advertising and brand collaborations stemming from the social ecosystem. Indeed, leveraging the core gameplay of 'Cameo digital avatars' and 'Remix secondary creation,' Sora has significant potential to pioneer a new form of advertising.

Based on Cameo's invitation co-creation mechanism, brands can launch 'avatar co-creation' activities, inviting users to customize avatar scenes with brand elements. For example, a sports brand could enable users to generate videos of themselves competing alongside virtual athletes, achieving viral spread through the Remix feature.

Note that in this mode, users serve as both content consumers and disseminators, allowing brands to reach target groups at a lower cost.

At the same time, OpenAI emphasizes that the product prioritizes recommending content that inspires creativity rather than merely extending user engagement time. This 'non-addictive' positioning may also attract brands averse to traditional information feed advertising, opening up a unique advertising collaboration avenue.

However, no matter how promising the commercialization blueprint appears, it cannot evade real-world challenges.

Firstly, there's the 'minefield' of copyright. Sora 2 adopts a 'default allow usage' copyright rule, where copyright holders must 'actively opt out' to prevent their works from being used for AI generation. This rule will undoubtedly raise strong concerns in the film and television industry. OpenAI may still face numerous copyright disputes in the future.

Secondly, there's the fragile social foundation. We know that Douyin's (TikTok's Chinese counterpart) strength lies in its content ecosystem and algorithm accumulation with hundreds of millions of users. Sora, starting with 'AI + social,' can initially attract users with novel gameplay, but once the novelty wears off, users may uninstall it due to 'friends not playing anymore' or 'lack of new content.'

The core of social products is 'relationship chain accumulation.' Relying solely on technological gimmicks cannot retain long-term users. Previous examples of social products like BeReal 'catching fire quickly but cooling down even faster' also underscore this risk.

Furthermore, computational costs pose an unavoidable obstacle. High-fidelity video generation places enormous demands on computational power. Although the payment model can cover some costs, the computational investment for free users still requires continuous support. If payment conversions do not meet expectations, it could impose a heavy cost burden.

Of course, its opportunities are also evident. AI has genuinely lowered the creative barrier to its lowest point, enabling everyone to become a 'director.' If it can successfully navigate the path of 'creating content - co-creating - accumulating interest communities,' it might give rise to a new ecosystem.

From a commercial perspective, it could be OpenAI's first 'advertising-native' C-end product. The video format is inherently suitable for brand interaction. If 'avatar licensing + Remix dissemination' can succeed, brands can engage in scalable co-creation.

Currently, invitation codes have been released in the United States and Canada, with iOS available first, while Android and web versions are still in the pipeline. Everyone, please exercise patience and wait.

In the end, the impact of Sora 2's technological breakthroughs is commendable, and its commercialization path is clear. The move to transform video generation from a 'tool' into a 'social carrier' is also a thoughtful strategy.

However, whether it can overcome the pitfalls of copyright, ecosystem, and computational costs, leap out of the 'novelty trap,' and transform from a 'toy' into a 'sustainably profitable product' remains to be seen and tested by time and the market.

Nevertheless, the growth of new species is never premeditated; it's trialed. When short videos first emerged, no one expected TikTok to be the ultimate victor.

So at this stage, it's better to maintain observation and see what new paths it can forge.

First, maintain imagination and innovation, then continue exploring commercialization models. After all, the starting point of every technological innovation lies hidden in many 'unclear' possibilities.

'The End'