Hello, this is CUBIG—enabling enterprise data to be truly usable for AI.
Following our previous announcement of being selected as a Finalist, we are back with even more exciting news thanks to your support. CUBIG has achieved 2nd place (Runner-up) in the global telecom innovation competition ‘T Challenge 2026,’ marking a historic first for a Korean company. 🎉
Out of 12 finalists from around the world, CUBIG was ranked 2nd, following Stanford University. This achievement reaffirms that our approach to ‘AI-Ready Data Infrastructure’ is highly recognized and valid on the global telecommunications stage.
What is T Challenge 2026?
T Challenge 2026 is a global open innovation program jointly hosted by Deutsche Telekom and T-Mobile US. Rather than being a simple idea contest, it is a platform designed to identify and validate technologies that can be practically applied to real-world telecommunications environments.
Under the theme ‘Building the AI-native Telco,’ the program aimed to discover solutions that enable next-generation networks to operate and optimize themselves autonomously based on AI.
Numerous teams from across the globe applied, and only 12 were selected as finalists to undergo a rigorous three-month development and validation process. Among these elite teams,
CUBIG secured the 2nd place position, right after Stanford University.
CUBIG’s Proposal: An AI operating layer based on ‘LLM Capsule’
Autonomous networks are structures that read, judge, and act on vast amounts of operational data in real time. Therefore, data must be organized and connected in a form that AI can utilize, and an operating system must be in place so that LLMs and agents can use it immediately.
In particular, because telecommunications companies handle data containing personal and sensitive information, it is often difficult to input it into AI immediately due to regulatory and security issues. This point has repeatedly become a bottleneck in the process of AI adoption for global telecommunications companies.
CUBIG’s LLM Capsule solves exactly this bottleneck.
✅ Supports both a plug-in method integrated into existing work tools and an independent web environment
✅ Real-time restructuring of a telecommunication company’s operational data—such as network configuration information, log data, and failure handling records—into a form that AI can utilize immediately
✅ Designed to convert data into a state connectable to AI and maintain that state consistently across various environments
✅ Sensitive information is converted according to policy standards rather than being connected to AI as-is; during this process, the visual structure and context of the data—such as document format, table structure, and log flow—are preserved
✅ Results are later restored back to fit the business context
The Next Stage of AI: Decided by the Operational Layer, Not the Model
This runner-up finish at T Challenge 2026 is not a final destination for CUBIG, but a starting point for our next phase.
The next stage of Enterprise AI will not be decided by ‘who has the largest model,’ but by ‘who can make AI-ready data to use’.
CUBIG is the company building that critical operational layer.
▪ Telecom: AI-Ready operational infrastructure for Autonomous Networks.
▪ Finance: Regulatory-friendly data operations that connect to AI while protecting sensitive information.
▪ Healthcare: Converting clinical and operational data into AI-usable states.
▪ Public Sector: Data governance infrastructure that satisfies both regulation and utility.
The demand for AI-Ready Data is growing fastest in highly regulated industries.
CUBIG will continue to lead this demand by developing technologies that protect data safely while transforming it into a truly actionable state.
📎Read more about CUBIG’s T-Challenge
📰Read More
Curious about our T Challenge-proven LLM Capsule?
🔗Request a Demo