Tech Mahindra announced its Zero Gravity Telco Architecture in Pune on 9 September 2026, presenting it as a framework for communication service providers (CSPs) moving towards AI-native and increasingly autonomous operations. The announcement also introduces the Zero Gravity Index, which the company describes as a diagnostic for assessing an operator’s readiness and sequencing its transformation.

This is not the launch of a network product, hardware platform or deployed autonomous network. Tech Mahindra’s announcement describes a strategic architecture and maturity assessment, but does not provide a customer deployment, performance benchmark, implementation timeline, price or detailed technical specification.

Contents

What Tech Mahindra announced

A CSP is an organisation that provides communications services such as mobile, fixed-line and broadband connectivity. These operators typically run large technology estates assembled over many years, with network, billing, customer-service and operational systems that may use different data definitions and embedded business rules.

Tech Mahindra calls the accumulated effect of these systems “Legacy Gravity”. In the company’s description, decades-old platforms, embedded rules and inconsistent definitions can make it harder to introduce new initiatives or coordinate data and decisions across an operator.

The Zero Gravity Telco Architecture is intended to address that problem by:

  • simplifying fragmented technology estates;
  • moving important business rules and definitions into a shared, governed layer; and
  • creating foundations that AI agents can use for operational tasks.

The company combines this proposed target architecture with the Zero Gravity Index. Tech Mahindra says the index assesses an operator’s maturity and recommends a sequence for moving from legacy environments towards adaptive and intelligent operations.

The release does not explain the index’s scoring system, criteria or validation method. It also does not establish whether the architecture is delivered as software, a reference architecture, a consulting methodology or a combination of these.

How the proposed architecture is meant to work

The central idea is to make operational context more consistent before adding AI systems to core telecom processes.

AI systems used for operational decisions need access not only to raw data, but also to definitions, permissions and business rules. For example, an automated system must understand what a service, customer state or network event means, which actions are allowed and who is authorised to approve them. If those meanings are scattered across incompatible systems, each AI project may need to interpret the same information separately.

A governed data or business-rules layer generally provides shared definitions, named ownership, access controls and rules for how information can be used. Tech Mahindra’s stated approach is to externalise relevant business context from legacy systems into such a shared layer, then allow AI agents and other operational tools to use it.

That could make multiple AI initiatives more consistent than isolated pilots built around separate interpretations of an operator’s systems. It could also make it easier to audit or update common rules if they are managed centrally.

However, the announcement does not disclose the architecture’s layers, interfaces, data models, agent controls, security design or interoperability mechanisms. The technical description therefore establishes the intended direction, not the implementation details.

“Autonomous operations” also covers a range of possible arrangements. It can mean automated monitoring, recommendations, workflow execution or systems that act with limited human intervention. The release does not define the level of autonomy targeted by Zero Gravity or specify where human approval would remain necessary.

Why the foundation matters for telecom AI

Telecom operators are seeking to use AI across increasingly complex environments, but adding an AI model does not by itself resolve fragmented data or conflicting operational rules. A system may generate technically plausible recommendations while lacking the context needed to apply them safely or consistently.

The proposed focus on shared definitions and governance addresses this less visible part of AI adoption. In principle, an operator that first organises its rules and data may have a more reusable foundation for service management, network operations and other automation tasks.

That does not mean the approach will produce the same result at every CSP. The effort required to simplify legacy systems, document rules and establish ownership depends on each operator’s architecture and operating processes. Tech Mahindra’s release provides no quantitative evidence that the framework reduces cost, improves reliability or accelerates deployment.

The announcement places Zero Gravity in the telecom industry’s broader movement from connectivity-led operations towards intelligence-led operations. Tech Mahindra also links the initiative to principles associated with TM Forum’s AI-Native Open Digital Architecture. The supplied evidence does not establish formal adoption, certification or technical conformance.

What to watch next

The most useful next evidence would be publication of the Zero Gravity Index’s criteria, scoring method and recommended transformation stages. Technical documentation describing the proposed layers, data models, interfaces and governance controls would also clarify whether the framework is a deployable architecture or primarily a strategic methodology.

Operator deployments would provide a stronger test of the proposal. In particular, reported changes in automation coverage, service reliability, operating cost, incident response or time required to launch new services would help distinguish an implemented capability from a conceptual target state.

The handling of human oversight will also matter. Telecom operations can affect service availability and customers at scale, so any AI-agent system will need clear permissions, audit trails, approval paths, rollback mechanisms and failure containment. None of those mechanisms is specified in the current announcement.

Sources