Migrated to Google Cloud in six months with negligible downtime
Consolidated platforms on Google Cloud, cutting transfer costs
Built cost savings with Axion instances and BigQuery analytics
Ad tech leader Verve migrated Dataseat to Google Cloud to unify infrastructure and cut costly cross-cloud transfers. The move boosts scale, helping process 1.5M ad auctions per second and deliver privacy-first campaigns.
Ad tech leader Verve migrated Dataseat to Google Cloud to unify infrastructure and cut costly cross-cloud transfers. The move boosts scale, helping process 1.5M ad auctions per second and deliver privacy-first campaigns.
Real-time demands don't wait for cloud sprawl
Dataseat — the demand-side platform part of the Verve Group since acquisition in 2022 — evaluates up to 1.5 million ad auctions per second, optimizing bids in real time without tracking user behavior. Their privacy-first approach demands more than just speed. It requires infrastructure that's capable of making precise, split-second decisions at massive scale. As Verve expanded through multiple acquisitions, its stack grew across several clouds. For Dataseat's engineers, that shift meant tackling duplicate systems, massive dataset transfers, and rising costs — challenges that ultimately led to migrating to Google Cloud as part of the broader unification effort.
"Training our model requires a high volume of data, which becomes costly and complex to manage across different platforms," Paul Hayton, Dataseat CTO. With a split infrastructure, friction crept into every layer. Model training lagged, reporting pipelines became brittle, and even small changes took extra coordination. As the platform scaled, the burden grew heavier. Consolidating systems became a priority.
Training our model requires a high volume of data, which becomes costly and complex to manage across different platforms.
Paul Hayton
Co-founder and CTO, Dataseat
As other parts of the business began operating on Google Cloud, migrating Dataseat there became a practical next step. The team started scoping what a full migration would entail — and how to roll it out while maintaining ongoing campaigns. With 24/7 bidding and thousands of ad placements per second, downtime wasn't an option.
Behind the migration was a team that thrives on a challenge. "They love learning new stuff," said Dr. Hayton. "From our point of view, it was brilliant — new platform, new world, let's go."
Strategic migration makes room for continuous innovation
The Dataseat team approached migration the same way they approach ad auctions: methodically, and with no margin for error. Rather than lift and shift, they stood up a parallel infrastructure on Google Kubernetes Engine (GKE) using Terraform, connected both environments through VPN, and gradually rerouted traffic to Google Cloud. The full cutover happened in just six months with only 90 minutes of planned downtime. "The fact that we got it done in six months is a testament both to my team and Google's investment in our goals," said Dr. Hayton.
With production now fully running on Google Cloud, the team has turned its focus to performance tuning and simplification. They're currently testing Axion-backed C4A instances side by side with their existing infrastructure, measuring bidder response times and compute efficiency to identify opportunities for reducing machine count and infrastructure costs. With their previous solution, Dataseat had reached the scaling limits of x86-based servers for in-memory key-value stores; with C4A, CPU utilization is much lower, so the same load can be handled with fewer servers. On the analytics side, they've started replacing legacy reporting pipelines with BigQuery, improving reliability and offloading operational overhead from their in-house systems.
Next, Dataseat plans to explore Google Cloud's machine learning ecosystem, including MLOps tooling and Gemini for AI-assisted experimentation. "We're already benchmarking Axion and exploring Gemini. The opportunity is to move faster, experiment more, and keep building a platform that runs at scale," said Dr. Hayton. These capabilities will help accelerate development cycles, deepen contextual optimization, and support the next phase of scale — without sacrificing the platform's privacy-first foundation.
That early adopter mindset has positioned Verve as close collaborators with Google on emerging technologies — from Axion benchmarking to scaling contextual AI across their platform. "Verve is at the pulse of the time. They're one of the first to try anything new from Google," said Sinan German, Google Cloud account manager for Verve and Dataseat. The result? A more efficient, cost-effective platform that scales easily, runs lean, and stays true to Dataseat's privacy-first promise.
The fact that we got it done in six months is a testament both to my team and the Google support throughout. With our previous cloud provider, we'd file tickets and wait. With Google, we had a Slack channel, quick responses, and a partner invested in our goals.
Paul Hayton
Co-founder and CTO, Dataseat
Dataseat, part of Verve, is a privacy-first ad tech company helping global brands reach audiences through contextual ads, real-time bidding, and data insights — without user tracking.
Industry: Media and Entertainment
Location: United Kingdom
Products: BigQuery, Google Kubernetes Engine
