Tencent Cloud Software / SaaS Global Translated from Chinese

Hyperparameter Technology (超参数科技)

Hyperparameter Technology needed huge elastic compute for game-AI training, low-latency global deployment for live gameplay, and a better way to coordinate fast-changing resources across many countries and workload types.

Read the original story Explore all use cases

AI Bot peak online concurrency reached nearly 1 million per day.

Tencent Cloud identified nodes with latency within 30 ms.

The company serves users in more than 50 countries and regions.

Reported impact

Business outcome

Tencent Cloud helped Hyperparameter Technology scale AI Bot services to global reach, lower training cost, and keep live interactions responsive even as bot populations and user coverage grew rapidly.

Solution

Hyperparameter Technology runs its AI Bot SaaS business on Tencent Cloud so it can train, deploy, and scale game AI globally without carrying fixed infrastructure overhead. The company uses a cloud-native architecture compatible with native Kubernetes, which means teams can load container images without managing cluster nodes while still maximizing cloud-server compute capacity. Tencent's large resource pool and spot instances reduce the cost of deep-learning training for increasingly diverse and independent AI Bots, and the platform can scale as the company moves from simple bot behavior to large populations of interacting agents. For live gameplay, Tencent's global footprint and customized BGP networking place compute near users, find nodes within 30 ms, and optimize routing during overseas congestion peaks. The result is a low-latency AI operations platform that supports very high peak concurrency across dozens of countries.

Data flow

Training jobs and runtime AI Bot services are packaged into Kubernetes-compatible containers and scheduled on elastic cloud servers and spot instances; global network routing then places gameplay close to users and keeps latency low during live interactions.

Solution architecture

5 components · 3 layers

Compute
- Kubernetes-Compatible Containers Package AI Bot services so they can be deployed without managing cluster nodes directly.
- Spot Instances Supply lower-cost compute capacity for deep-learning training workloads.
- Cloud Server Delivers the core compute power behind AI Bot training and runtime scaling.
Serving
- Global BGP Network Routes gameplay traffic through low-latency paths and optimizes performance during overseas congestion.
Orchestration
- Cloud-native Architecture Provides the base design that lets AI Bot services scale elastically across Tencent Cloud.

Architecture clues

Routing policy is tuned for overseas congestion peaks to preserve real-time gameplay quality.
Tencent spot instances are used to reduce deep-learning cost.
The platform is compatible with native Kubernetes and removes the need to manage cluster nodes directly.

Evidence from the source

Hyperparameter's AI Bot roadmap requires huge elastic compute for self-training and for large numbers of interacting AI agents. (original: 这个过程中需要大量的训练数据积累，甚至要在没有玩家对战信息的情况下自训练，要追赶人类行为逻辑就需要海量弹性的计算资源。)

Tencent Cloud found node resources with latency within 30 ms and optimized routing specifically for overseas congestion peaks so AI Bots could compete with players in real time. (original: 腾讯云的全球布局以及优质的 BGP 网络：为超参数覆盖游戏客户时，找到延时在30ms以内的节点资源，同时对于海外高峰期网络拥堵问题，针对性进行路由策略优化，保障 AI Bot 与玩家对战的实时智能。)

Tencent Cloud helped Hyperparameter Technology reach nearly one million peak daily AI Bot concurrency and serve hundreds of millions of users in more than 50 countries. (original: 腾讯云始终陪伴超参数一起成长，实现超参数AI Bot每天峰值在线数近百万，服务全球50余个国家的数亿用户。)

Using a cloud-native architecture, Tencent Cloud's global footprint, and customized BGP networking, the company can use native Kubernetes compatibility and high-cost-performance spot instances. (original: 通过云原生架构、腾讯云的全球布局以及优质的 BGP 网络定制解决方案：1.使用云原生架构：一方面兼容原生K8S...借助腾讯云海量资源池，利用竞价实例的闲时高性价比。)

Transparency Evidence & methodology 100/100 reference value · 79% confidence · 1 source

Review how this reference was scored and trace each normalized claim back to its documented source.

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value100/100

Problem clarity20/20
Measures whether the customer's operational problem is specific enough to compare.
Implementation depth25/25
Rewards named components, data flow, integrations, and architecture detail.
Outcome specificity20/20
Measures whether the reported benefit is concrete rather than a generic claim.
Quantified impact25/25
2 metrics describe measurable impact.
Deployment scale10/10
Rewards evidence of a live or scaled deployment and meaningful integration scope.

Evidence confidence79%

Source authority20/25
The main evidence is an official vendor-published customer story.
Claim traceability21/25
10 of 12 key fields link to a named source.
Direct evidence20/20
4 direct source quotes retained.
Date provenance0/10
The publication date has no documented provenance.
Source transparency8/10
1 source reference document publisher, URL, authority, retrieval, and supported fields.
Implementation specificity10/10
Rewards source-backed components, data flow, quantified impact, and a specific outcome.

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

Tencent Cloud
Hyperparameter Technology (超参数科技) customer story

First-party
https://cloud.tencent.com/customer/chaocanshu

Source type
Vendor-published customer story
Supports
- Customer
- Business problem
- Solution
- Products
- Solution components
- Data flow
- Integrations
- technical area
- use case category
- Business outcome
- outcome category
- Metrics
- Architecture
- evidence quotes

Date evidence

First archived Feb 17, 2023 · Internet Archive · Independent First Wayback Machine capture; confirms the page existed by this date, not when it was published.
Open date source

Not separately sourced Industry, Region, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.

Similar use cases

See all like this

Tencent Cloud Archived · Mar 11, 2026 Aye Solutions Software / SaaS Southeast Asia Aye needed to prove that face-based payments could resist photos, videos, deepfakes, masks, and replay attacks while still delivering a smooth checkout experience that merchants and end users would trust. Aye has worked with Tencent Cloud for about 2 years. Fraud Detection AyeFace · Identity Verification · Liveness Detection

Tencent Cloud 3D AI Studio Software / SaaS Global 3D AI Studio needed a production-grade foundation for highly elastic generative 3D workloads so demand spikes and long generation times would not undermine reliability, user trust, or its planned expansion from B2C SaaS… 3D AI Studio reduced 3D generation time from a seven-minute bottleneck to a three-to-five-minute window. Tencent Cloud

Tencent Cloud CGTrader Software / SaaS Global CGTrader needed to remove manual bottlenecks from 3D asset production, help buyers source high-quality models at scale, and lower the technical barriers that made production-ready 3D creation slow and dependent on… Tencent HY 3D Global