Google’s answer to OpenAI’s Sora has landed – here’s how to get on the waitlist
One of the many AI treats Google threw into the crowd at the Google I/O 2024 keynote was a new video tool called Veo – and the waitlist for the OpenAI Sora rival is now open to those who want early access.
From Google early Veo demos, the generative video tool certainly looks a lot like Sora, which is expected to release “later this year.” It promises to create 1080p resolution videos that can “last more than a minute” and in a variety of cinematic styles, from time-lapses to aerial drone footage. You can see an example further down this page.
Veo, the engine behind a broader tool from Google’s AI Test Kitchen called VideoFX, can also help you edit existing video clips. For example, you can give it an input video in addition to a command, and it will be able to generate additional scenery – Google’s example is the addition of kayaks to a coastal scene from the air.
But like Sora, Veo is also only open to a select number of early testers. You can now sign up to become one of those ‘Trusted Testers’ via the Google Labs form. Google says it will “review all submissions on a rolling basis” and some questions – including one asking you to link to relevant work – suggest it might initially only be available to digital artists or filmmakers.
Still, we don’t know exactly what the criteria is for being an early Veo tester, so it’s worth signing up if you’re keen to get started.
The AI video tipping point
Veo is certainly not the first generative video tool we’ve seen. As we noted when the Veo first launched, games like Synthesia, Colossyan, and Lumiere have been around for a while. OpenAI’s Sora has also reached the mainstream with his throwback music videos and strange TED Talk promos.
These tools are clearly reaching a tipping point, as even relatively conservative Adobe has shown how it plans to integrate generative AI video tools into its industry-standard Premiere Pro, again “later this year.”
But the significant computing power required to run Veo’s diffusion transformer models and maintain visual consistency across multiple frames is also a major bottleneck in a wider rollout, which explains why many are still in demo form.
Still, we’re now reaching a point where these tools are ready to partially jump into the wild, and being an early beta tester is a good way to get a feel for them before the inevitable monthly subscriptions are defined and rolled out.