"Cloud Run is effectively acting as our hosting platform," Higgs explains. "All our developers have to do is get their service into a Docker container and everything else will be taken care of. Deployment is done. Scaling is done. Logging is done. Even basic dashboarding is done for them."

This has a knock-on benefit for the customer. "The fact that we can spin up services so easily on Cloud Run means that we tend to spin things up as separate services, which lets us ship rapidly and get features to customers much faster," says Higgs. 

And because VEED's developers don't need to spend time managing infrastructure, they can focus time and energy on improving features and services. VEED has made significant savings on its cloud spend too, thanks to the autoscaling feature of Cloud Run, which automatically scales the resources it uses up and down as necessary, ensuring VEED only pays for what it needs.

This ability to scale resources as necessary also applies to VEED's use of Cloud GPUs, which the company uses to render the videos of its more than 12 million monthly users. 

With spiky demand for its render nodes, each powered by a GPU, VEED's compute-power costs would rise rapidly if not managed effectively. Using Cloud Load Balancing, VEED is able to scale its usage to ensure it isn't wasting resources.

With Google Cloud, VEED can also keep scaling up its resources as it grows, allowing it to offer users a stable and reliable video editing platform, no matter how many customers are using it. At the same time, Google Cloud helps VEED to keep latency to a minimum. As Higgs explains, "The geographically distributed nature of Google Cloud means we can run things across multiple regions to keep latency down. No matter where they are, customers not only get their videos faster, but experience better quality editing too when there is low latency."

VEED has grown rapidly, doubling its annual recurring revenue every year since it began monetizing the business. The amount of data VEED processes has grown too, from 0.23 petabytes of ingested data in 2021 to 4.05 petabytes in 2024. Like its compute power expenditure, VEED's storage costs could easily rise if left unchecked, so VEED created a tiered storage system with Cloud Storage. This ensures that its systems have fast access to the data they need, while data that is not needed so often, such as users' old videos, can be stored in a more cost-efficient way. By utilizing the storage tiers, Veed saved over 50% on their storage costs.

Another AI-powered feature that VEED is currently developing is its lip-sync tool. This uses machine learning to enable users dubbing their videos into another language to synchronize the speaker's lips to the audio track. Using models in Model Garden on Vertex AI, VEED is training its models to automatically regenerate the speaker's lip movements so that they match the voice. For VEED, using Vertex AI is another way to reduce its development time and keep its team lean by relying on the existing optimized AI infrastructure of Google Cloud. 

"We have no desire to build an entire AI team from scratch. That would get very expensive," explains Higgs. "But with Vertex AI providing a lot of the infrastructure, such as data storage, analysis, and everything else, we can take the path with the fewest operational overheads and let our staff focus on what they do best."

As VEED grows, Higgs makes use of weekly check-ins with the Google Cloud team to stay abreast of emerging technologies, while experimenting with Vertex AI models, as the company develops more AI features to make video editing even simpler. 

"Without Google Cloud, we wouldn't have grown to where we are today," says Higgs. "It let us grow faster with fewer staff, and instead of having to hire a sizable platform team, I could spend that money on features for our customers. All of which gives us a clear advantage going forward."