As a valued member of CO/AI, you're invited to join us at HumanX 2025, the premier AI conference shaping the future of technology.

Why Attend HumanX?

Don't miss this opportunity to be part of the AI revolution.

Register now with our special code HX25p_coai and save $250 on your general admission pass!

Opening shots

Google introduced Veo 2, an advanced AI video generation model that claims to outperform OpenAI's recently launched Sora. The timing is notable - OpenAI's Sora has faced mixed reviews since its release, with users reporting inconsistent quality and technical limitations, particularly around physics and complex motion.

Market positioning

OpenAI opted for a subscription based model - ChatGPT Plus ($20 monthly) includes 50 video generations at 480p resolution. Google has implemented a waitlist most likely to manage computational demands and ensure system stability as they scale and more will be known about how they plan to deploy their tool in the coming days.

Meta has meanwhile carved out its own path, focusing on open-source research and collaboration. Their recent release of nine major research projects, including video watermarking technology and virtual agent controls, suggests they're playing a longer game focused on infrastructure rather than immediate commercial applications.

Technical reality check

The limitations of current AI video generation have become increasingly apparent. Take the viral example of Sora's gymnast video, which featured physically impossible movements - a phenomenon experts have dubbed "jabberwockies." These errors stem from fundamental limitations in how these models process physics and anatomy.

At the same time, Google alongside its Veo 2 tool, has introduced Whisk, an image generation tool that emphasizes iterative refinement and visual exploration. This more focused approach might prove more immediately useful than attempting to solve all video generation challenges at once.

The path forward

The current state of AI video generation mirrors earlier stages of text and image AI - impressive capabilities mixed with notable limitations. The real question isn't which company will win this round, but rather how these tools will reshape creative workflows once they mature.

Some creative experimentation is already happening at the edges. Both OpenAI and other companies have launched AI Santa experiences, demonstrating potential applications in interactive entertainment. While these might seem gimmicky, they provide valuable insights into how AI video can enhance real-time interactions.

The next few months will likely bring rapid improvements in video generation quality and reliability. The question is whether the tech giants can overcome current limitations before smaller, more focused competitors find profitable niches in specific use cases. The race isn't just about technical capabilities - it's about finding the right balance of features, accessibility, and pricing that will drive widespread adoption.

Looking ahead, the true winners might not be those who perfect video generation first, but those who best integrate it into existing creative workflows. The battle for AI video is just beginning, and the front lines are being redrawn with each new release.

We publish daily research, playbooks, and deep industry data breakdowns. Learn More Here