Bench Mark Infotech outlines conditional ₹100 crore GPU expansion plan
Key highlights
- Bench Mark Infotech proposed about ₹100 crore of phased GPU computing investment on November 13, 2026.
- The proposal has no disclosed operational GPU fleet, secured financing or executed third-party marketplace contracts.
- Bench Mark Infotech has over 19 years of technology-infrastructure experience, including artificial intelligence labs and data-centre systems.
- The company plans two models: an owned GPU cloud and a third-party capacity platform.
powered by IRIS

Bench Mark Infotech Services Limited proposed on November 13, 2026 a phased investment of about ₹100 crore in graphics processing unit (GPU) computing for GPU-as-a-Service and cloud infrastructure, subject to approvals, funding arrangements, commercial evaluation and market conditions. The plan would add customer computing-capacity services to its existing IT infrastructure integration business.
Know Bench Mark Infotech Services inside out
Results, valuation, peers and who's buying, in one view.
What has Bench Mark Infotech proposed?
Bench Mark Infotech has proposed a phased expansion into GPU-as-a-Service, or GPUaaS, and cloud infrastructure, supported by an investment of about ₹100 crore in GPU computing in the near future. The November 13, 2026 announcement describes a contemplated strategic direction rather than completed capital expenditure, and makes the investment conditional on approvals, funding arrangements, commercial evaluation and market conditions.
The proposed business would change the delivery model for part of Bench Mark Infotech's IT infrastructure activity. Its current work is built largely on engineering, procurement, installation and systems integration, while the proposed GPU business would involve owning, managing and offering computing capacity for ongoing customer use. The company said it would continue serving existing IT infrastructure and systems-integration customers alongside the new service.
The announcement does not state a deployment date, GPU fleet size, technology supplier, customer commitment or procurement order for the ₹100 crore proposal. Bench Mark Infotech said the timing, asset mix and final investment amount may change as demand, procurement costs, infrastructure readiness, customer requirements and risk-management considerations develop.
How would Bench Mark Infotech offer GPU capacity?
Bench Mark Infotech plans two connected routes for GPU capacity: a company-owned GPU cloud and a third-party GPU rental platform. Under the owned-cloud model, the company would buy and deploy GPU-based computing infrastructure for enterprises, artificial intelligence developers, research organisations and other users through pay-per-use access, dedicated capacity or contracts.
The proposed owned cloud is intended for artificial intelligence model training, fine-tuning, inference and other compute-heavy workloads. Bench Mark Infotech would own and operate the relevant computing assets, with rental income identified as the intended income mechanism. The filing does not disclose expected utilisation, prices, contract lengths or revenue from this service.
The second route is a proposed platform for spare GPU-server capacity owned by independent operators and compatible infrastructure providers. Subject to onboarding criteria and commercial terms with each provider, Bench Mark Infotech intends to help users find capacity, set up access, make payments and receive support. The marketplace model could expand available capacity without the company owning every server listed.
The two models differ in both capacity ownership and proposed income source. The owned cloud would use Bench Mark Infotech's GPU assets and target rental income, while the platform would connect users with third-party infrastructure and target platform and infrastructure-management income. Neither model has disclosed executed marketplace contracts, an operational GPU fleet or assured future revenue.
Got a follow-up? Ask IRIS
Why is Bench Mark Infotech entering GPU-as-a-Service?
Bench Mark Infotech says the proposed service builds on over 19 years of technology-infrastructure experience and past work designing, supplying, integrating and commissioning artificial intelligence laboratories, data-centre systems, storage and networking solutions. Management described the move from delivering infrastructure projects to operating computing capacity as a next step from that technical work.
Chairman and Managing Director Vineet Kumar Gupta said on November 13, 2026 that businesses, developers and researchers need GPU computing for artificial intelligence work. He said each rollout phase would be matched to customer demand and approvals in place, which states management's intended pacing rather than a disclosed demand forecast or operating target.
Bench Mark Infotech's existing offerings include local area network and wide area network infrastructure, wireless communication systems, structured cabling, multimedia and audio-visual systems, smart classrooms, access control, surveillance, cybersecurity, data storage, data-centre solutions and fibre-optic infrastructure. The proposed GPU services would add managed computing capacity to these integration and infrastructure offerings.
Bench Mark Infotech is headquartered in Kolkata and has a branch office in Patna. It says it is empanelled with BSNL as a national-level system integrator and with RailTel Corporation of India Limited as a business partner, but the filing does not say either relationship provides funding, GPU supply, customers or contractual support for the ₹100 crore proposal.
What conditions could affect the ₹100 crore rollout?
Bench Mark Infotech says the ₹100 crore rollout depends on internal and statutory approvals, funding, technology procurement, infrastructure availability, contractual arrangements, market demand and other risks. The company has not described the proposal as committed capital expenditure and has not identified a financing source, amount or terms for the planned phases.
Procurement costs and infrastructure readiness are specifically identified as factors in rollout decisions. Because the service would require deployment of GPU-based computing assets and, separately, onboarding of third-party capacity providers, timing may depend both on equipment procurement and commercial arrangements with marketplace participants.
The release also identifies government actions, local, political and economic developments, and technological risks as factors that could cause actual outcomes to differ materially from forward-looking statements. Bench Mark Infotech gives no assurance of future revenue, profit or returns from the GPU plan and says it has no obligation to publicly update forward-looking statements for subsequent events or circumstances.
No financial targets accompany the proposed expansion. Bench Mark Infotech has not disclosed an expected revenue contribution, margin, break-even point, capacity-utilisation threshold, customer count, launch schedule or split of the approximately ₹100 crore between the owned cloud and third-party platform, leaving the commercial scale of each phase unresolved.
Conclusion
Bench Mark Infotech's November 13, 2026 proposal would extend its model from IT infrastructure project delivery into computing capacity offered as a service. The approximately ₹100 crore plan combines an owned-cloud route intended to generate rental income with a third-party marketplace route intended to generate platform and infrastructure-management income without ownership of every server.
The company has not set dated milestones for the rollout. Subsequent disclosures on approvals, funding arrangements, GPU procurement, commissioned capacity, third-party provider agreements and the first commercial deployment would show how the demand-led, phased plan is progressing.


