Google Veo 3 AI Video Hits Prime-Time Ads

Google’s Veo 3 Debuts in Prime-Time Advertising

In a landmark moment for AI-driven marketing, Google’s new video generation model, Veo 3, powered a GTA-style commercial for betting platform Kalshi during the NBA Finals. The ad, conceived and built in just two days by creator PJ Ace, represents an unprecedented push of generative video into mainstream television.

Unveiled at Google I/O last month, Veo 3 joins a growing roster of AI tools transforming ad production. From ideation to final render, PJ Ace tapped Gemini and ChatGPT to draft the script, then asked Gemini to craft the precise prompt for Veo 3—effectively letting AI write for itself before generating the visuals.

How the Spot Came Together

According to Ace, the process involved 300–400 individual generations, stitched into rapid-fire scenes that showcase Kalshi’s brand amid chaotic, high-energy action. Each short clip captures a moment—whether a SWAT team standoff or a neon-lit chase—but continuity often breaks when tying them together.

Benefits and Limitations

• 95% cost reduction compared to traditional TV ads
• Rapid production: ad built in just two days
• Persistent scene continuity issues hamper narrative flow
• High volume of generations needed adds hidden complexity

Broader Implications for Advertising

The Kalshi spot underscores a shift in ad production economics. AI workflows can drastically cut payroll and agency fees, raising questions about the future of creative teams. As brands chase efficiency, they must also balance brand safety, quality standards, and creative direction.

The Road Ahead for AI-Driven Video

While Google’s longer-form tool, Flow, promises more coherent storytelling, Veo 3 highlights where generative video excels—and where it stumbles. Marketers and agencies will need robust testing, human-in-the-loop checks, and clear guidelines to ensure AI-enabled ads elevate rather than confuse their audience.