How a gaming company cut cloud costs 30% on AWS

The challenge

Snail Digital Technology (Snail Games) runs online games for players around the world. As the player base grew, their infrastructure had to do two things at once: stay fast for players wherever they are, and keep costs under control as traffic grew.

On their previous Google Cloud setup, the team faced:

• Rising cloud costs as player numbers grew.
• Latency for players far from their main region, which hurts gameplay — especially in North America and Europe.
• Slow releases, making it hard to ship new features and events quickly.
• No easy path to AI features they wanted to build, such as player support and chat moderation.

What Bytenamics did

Bytenamics planned and ran the migration from Google Cloud to AWS, then helped the team put AI to work on the new platform.

1. Assess and plan

We mapped the existing workloads, player traffic and costs, and designed a target architecture on AWS sized for global play. The plan covered more than 40 services and their databases.

2. Build a global foundation

We set up a secure multi-account AWS environment and placed game services closer to players, so traffic travels a shorter distance. Game services now run in US East, Europe (Frankfurt) and Asia Pacific (Singapore), with Amazon CloudFront and AWS Global Accelerator in front.

3. Migrate without disrupting players

Workloads moved in planned waves, each with testing and a rollback plan, to keep downtime for players to a minimum. The full move took 14 weeks, with no unplanned downtime.

4. Optimize cost

After the move we right-sized resources and applied AWS pricing options to bring monthly spend down — mainly Savings Plans, AWS Graviton instances and Spot Instances for non-critical workloads.

5. Innovate with AI

With the platform on AWS, the team started building AI features using Amazon Bedrock: a player support assistant, in-game content generation and chat moderation.

1. Week 3: Assessment completeArchitecture and migration plan signed off.
2. Week 6: Foundation liveAWS accounts, networking and security in place.
3. Week 14: Migration completeAll game services running on AWS.
4. Week 18: First AI feature launchedPlayer support assistant on Amazon Bedrock.

The results

• 30% lower cloud spend compared with their previous Google Cloud costs.
• Lower latency for players worldwide — 45% lower on average in North America and Europe.
• Faster time to market for new features and in-game events — releases went from monthly to every two weeks.
• AI-ready platform, with 3 AI features now in production or in development.

What's next

Next, the team plans to use AI for game analytics and personalized player offers on AWS.

Planning a similar move? See our migration and generative AI services.