Bengaluru, India | Mar 24: As India races to cement its position as a global AI powerhouse, San Francisco based Operant AI today announced the launch of its AI Infrastructure Ecosystem Partnership Program to embed real-time AI defense directly into the inference infrastructure powering the world’s most demanding enterprise AI workloads.
The explosive growth of AI agents, MCP-connected systems, and large-scale inference deployments has fundamentally changed the threat landscape. Enterprises are running increasingly autonomous AI systems with access to sensitive data, critical tools, and external connections often without the security controls required to govern them. Operant AI’s platform addresses this directly, providing the only inline, real-time security layer purpose-built to discover, detect, and defend AI, Agents, and MCP as they run inference for customers and enterprises at scale.
India’s AI Infrastructure Surge Creates Urgent Security Imperative
India’s AI infrastructure expansion is accelerating rapidly across both government and private sector initiatives. Under the IndiaAI Mission, the government has already deployed 38,000 GPUs, with an additional 20,000 GPUs which was recently announced at the India AI Impact Summit 2026 last month. Across public and private deployments, India now has more than 80,000 GPUs in operation with mega investments reshaping the AI landscape.
Data-centre operators including ST Telemedia Global Data Centres, CtrlS, Sify Technologies, Nxtra by Airtel, and Princeton Digital Group are expanding capacity for AI workloads, while GPU cloud providers such as Yotta Data Services, NxtGen Datacenter & Cloud Technologies, and Neysa Networks are building specialized AI compute platforms.
India’s total data-centre capacity is estimated to grow from roughly 2 GW today to 8–9 GW by 2030, driven by large investments from hyperscalers such as Google, AWS, Reliance, and other global infrastructure providers.
However, this infrastructure expansion is outpacing the development of security frameworks needed to protect production AI systems. According to the FICCI–EY Risk Survey 2026, released in February, 61% of Indian business leaders now identify cyber-attacks and data breaches as posing “significant financial and reputational risks” to their organisations, elevating cybersecurity from an IT concern to a board-level priority.
As the race to build AI infrastructure accelerates, security is emerging as a critical differentiator. With AI systems becoming increasingly agentic and interconnected through emerging standards such as the Model Context Protocol (MCP), the potential attack surface is expanding. Prompt injection attacks, tool-level vulnerabilities, and runtime manipulation are among the growing risks facing production AI deployments.
Infrastructure providers joining Operant AI’s GPU Ecosystem Program gain the ability to offer customers not just raw compute power, but verifiably secure inference environments where models and agents can run safely in production. As enterprises increasingly scrutinize the security posture of their AI infrastructure, the ability to demonstrate robust, real-time protection at the inference layer is becoming essential to winning deals and retaining customers.
GPU-Accelerated Security for High-Performance AI Systems
The new GPU-accelerated versions of AI Gatekeeper and MCP Gateway provide blazing-fast security scanning and validation for models, agents, tools, and the complete inference-layer AI toolchain. By leveraging GPU acceleration, these solutions ensure that security measures keep pace with the high-throughput demands of production AI systems without introducing latency bottlenecks. Organizations can now protect their AI infrastructure with enterprise-grade security that operates at the speed their models require.
Real-Time Security at the Inference Layer: Key Capabilities
Operant AI secures models and agents directly at the inference layer—the precise moment when inputs are processed and outputs are generated. This runtime security approach monitors every interaction with the model in real-time, analyzing requests and responses as they flow through the system to identify and neutralize threats before they can compromise model behavior.
Key capabilities include:
-
Prompt Injection Detection and Blocking: Real-time analysis at inference time prevents malicious prompts from altering model behaviour, extracting training data, or bypassing safety constraints
-
Zero-Day Attack Protection: Behavioural monitoring detects anomalous patterns in how models are queried or how agents interact with tools, blocking novel exploits before they succeed
-
MCP Security for Agents: Runtime validation of agent actions and tool calls ensures autonomous systems operate within safe parameters, preventing unauthorised resource access or unintended behaviours
-
Intelligent Rate Limiting: Inference-time controls reduce unnecessary model invocations and prevent costly token consumption from automated attacks or inefficient agent behaviours, optimising operational costs while maintaining security
“We are at an inflection point where the scale, speed, and autonomy of AI systems have outpaced the security controls designed to govern them.The path to responsible AI isn’t just about building better models—it’s about securing them at the moment they matter most,” said Vrajesh Bhavsar, Co-founder and CEO at Operant AI.“As Indian enterprises deploy AI models and agents across financial services, healthcare, and public sector environments, the inference layer is where security must be enforced. Our GPU-accelerated AI Gatekeeper and MCP Gateway, combined with the GPU Ecosystem Program, deliver the speed and protection that India’s production AI systems demand.”
Bhavsar brings more than two decades of hardware AI and cybersecurity expertise to Operant AI. A kernel engineer by training, he built core security technologies for iOS and macOS at Apple — including Data Protection and the Secure Enclave that underpin the iPhone’s security architecture — before founding the Machine Learning business unit at ARM, where he brought on-device AI and security to billions of global devices and built an ML ecosystem program with more than 50 partners
“As autonomous agents become more sophisticated and models take on increasingly critical roles, securing the inference layer is no longer optional. It is the foundation on which safe, agentic systems must be built. We’re transforming models and agents from vulnerable systems into trustworthy, production-ready intelligence that organizations can deploy with confidence. Our goal is to ensure that the AI momentum is not only powerful, but also secure” he added.
Real-Time Defense for the World’s Fastest AI Inference Platforms
The newest generation of AI inference platforms — purpose-built silicon delivering thousands of tokens per second — has unlocked a category of AI applications that simply weren’t possible before: real-time agentic workflows, sub-second reasoning, and complex multi-step tasks running continuously at enterprise scale. Customers across healthcare, financial services, government, and enterprise SaaS are betting their most ambitious AI initiatives on this kind of infrastructure, precisely because speed at this level changes what AI can do.
“As AI moves to always‑on agents, the bar for infrastructure gets higher: it has to be performant and open by design. Partnering with Operant AI lets Tenstorrent customers pair our high‑throughput, Tensix‑based inference platforms with real‑time visibility of agents at the inference runtime layer, enabling them to scale their AI initiatives with confidence,” said Aniket Saha, VP of Product Strategy, Tenstorrent.
Together, Tenstorrent and Operant AI are bringing high‑throughput, real‑time inference and real‑time AI monitoring into a single, integrated infrastructure stack. By sitting directly in the inference stack on top of Tenstorrent hardware, Operant AI gives joint customers real‑time visibility and enforcement over what their models are doing, so they can take full advantage of Tenstorrent’s performance with the added benefit of enterprise‑grade compliance and traceability.
The AI Infrastructure Ecosystem Partnership Program is open to AI infrastructure providers, inference platforms, and MCP-compatible application vendors. Partners gain access to Operant AI’s runtime defense capabilities — including AI Gatekeeper, MCP Gateway, and Agent Protector — enabling them to deliver integrated, real-time security directly to their customers. Operant AI is the only vendor featured across six of Gartner’s critical AI security reports, including the AI TRiSM Market Guide, API Protection Market Guide, MCP Gateways Innovation Insight, and How to Secure Custom-Built AI Agents.
