Sarvam Bets on Google Cloud India's Former Chief to Drive Enterprise Growth

Indian AI company Sarvam has appointed Sashikumar Sreedharan as president and chief business officer. Sreedharan, most recently managing director of Google Cloud India, will also join Sarvam’s board.

Sreedharan will oversee marketing, finance, forward-deployed engineering, service delivery and commercial operations. Forward-deployed engineers work directly with customers to adapt and deploy technology, giving his role responsibility for much of the path from enterprise sale to production use.

A cloud operator gets a wide remit

Sreedharan brings more than three decades of experience in enterprise technology. He took charge of Google Cloud India in July 2025 after serving as chief operating officer for Google Cloud Asia Pacific from September 2023. Before joining Google, he led Microsoft India’s enterprise business as managing director for more than five years.

Those roles gave him experience selling cloud infrastructure and software to Indian companies, public-sector organizations and regulated industries. Sarvam now needs similar capabilities for long procurement cycles, security reviews, systems integration and ongoing support.

Capital and customers raise the operating load

Sarvam says more than 100 enterprises and government institutions are building with its platform. The company also signed an HQ IDS agreement covering AI adoption across the Indian Army, Navy and Air Force. HQ IDS, short for Headquarters Integrated Defence Staff, coordinates work among the armed services.

In June 2026, Sarvam raised $234 million in the first close of a planned $300 million Series B round. The financing valued the company at $1.5 billion after the investment. That funding supports expansion, while the growing customer base creates demand for repeatable onboarding, implementation and service processes.

Sarvam’s enterprise pitch centers on customer control. The company says organizations can keep their data, models and AI agents within their chosen environment. This approach targets buyers with strict requirements for data residency, security, customization and operational control.

The stack now spans models to deployment

Sarvam describes its offering as a full-stack AI platform, meaning it supplies models, developer interfaces, applications and deployment tools. Its current catalog includes:

• Language and voice tools: speech-to-text, text-to-speech, translation APIs and voice agents tuned for Indian languages.
• Document systems: digitization services and enterprise document AI.
• Software agents: work agents and coding agents designed to complete multistep tasks.
• Inference and edge deployment: Token Factory for running model requests and Sarvam Edge for processing on devices.
• Custom model services: training and fine-tuning for organizations that need specialized models.

Platform growth can outpace model adoption

Sarvam also hosts third-party models through Token Factory. It charges roughly ₹29 per million input tokens for Sarvam-105B and about ₹20 per million for DeepSeek V4 Flash. Tokens are the units into which a model divides text for processing.

The lower price for the Chinese-developed DeepSeek model creates a commercial question. Customers can increase their use of Sarvam’s infrastructure without adopting Sarvam’s own flagship model. Model-level usage, margins and pricing will show how the company balances its domestic model program with its broader inference business.

Sarvam has often been compared with France’s Mistral as a potential national foundation-model provider. Its third-party catalog gives it an additional role as a distribution and deployment layer for models developed elsewhere. Each business has different economics: proprietary models depend on direct adoption, while an inference platform can earn revenue across a wider catalog.

Sreedharan connected the appointment to Sarvam’s domestic model ambitions in the announcement: “India should be a maker of the intelligence the world runs on. For me, joining Sarvam is a commitment to that opportunity.”

No technical migration has been announced

Sarvam disclosed no immediate changes to its APIs, model endpoints, SDKs, prices or service terms alongside the appointment. The announcement therefore creates no migration requirement for existing users. Its operational effects will emerge through several areas:

1. Integration support. Responsibility for forward-deployed engineering could expand hands-on assistance with customer deployments. Sarvam has not published staffing levels or eligibility criteria.
2. Pricing and model lifecycle. Changes would appear in token prices, rate limits, context windows, versioning policies and model deprecation schedules.
3. Service levels. Formal commitments covering uptime, latency, support response and incident handling would strengthen the platform’s production offering.
4. Deployment controls. Contracts and technical documentation need to specify data location, retention, model isolation and the environments supported by each product.
5. Public-sector requirements. Work with HQ IDS could produce additional security and compliance features, although Sarvam has not announced a related product roadmap.

By placing marketing, finance, implementation and service delivery under one executive, Sarvam has concentrated responsibility for converting its model and API catalog into repeatable enterprise deployments. Customer renewals, deployment times, service commitments and usage across first-party and third-party models will provide the clearest measures of the hire’s impact.