L&T partnership: enterprise AI factories
Larsen and Toubro (L&T) is the Indian engineering, construction, and technology conglomerate. The L&T-Nvidia partnership focuses on building enterprise AI factories for L&T's industrial customers in manufacturing, energy, and infrastructure.
| L&T-Nvidia partnership | Detail |
|---|
| Use case | Enterprise AI factories for industrial customers (manufacturing, energy, infrastructure) |
| Deployment model | L&T builds and operates the AI factory on customer premises or in L&T-managed data centers |
| Customer size | Mid-market and enterprise industrial customers (Rs 1,000 Cr+ revenue) |
| AI workloads | Predictive maintenance, quality control, supply chain optimization, process automation |
| GPU stack | Nvidia Blackwell B200 + Spectrum-X networking + AI Enterprise software |
| Pricing model | Capex + managed services; L&T finances the AI factory for the customer |
L&T's competitive position is the industrial-customer relationship. L&T has decades of relationships with Indian industrial customers (Tata Steel, Reliance, ONGC, NTPC, Indian Oil) and the operational expertise to deploy AI factories on customer premises. The Nvidia partnership gives L&T access to the Nvidia AI stack and the credibility to win enterprise AI factory contracts.
E2E Networks partnership: sovereign AI compute
E2E Networks is an Indian Tier-3 cloud provider that has been a key player in India's sovereign cloud market. The E2E-Nvidia partnership focuses on providing sovereign AI compute to Indian enterprises and government, on a Tier-3 sovereign-cloud model where the data, the compute, and the model weights stay within Indian jurisdiction with no cross-border data transfer.
| E2E-Nvidia partnership | Detail |
|---|
| Use case | Sovereign AI compute for Indian enterprises and government workloads |
| Deployment model | GPU rental + managed services in India data centers only (Tier-3 sovereign cloud) |
| Customer focus | Indian government, BFSI, healthcare, telecom - regulated industries with data-residency requirements |
| GPU stack | Nvidia H100 / H200 / Blackwell cluster + AI Enterprise software + sovereign deployment |
| Pricing premium | 10-20% above AWS / Azure / GCP equivalent (justified by sovereign-data-residency value) |
| Competitive position | Leading Indian Tier-3 sovereign cloud; the right pick for data-residency-sensitive AI workloads |
The customer use cases are government AI workloads (defense, intelligence, citizen services), regulated-industry AI workloads (BFSI, healthcare, telecom), and Indian enterprise AI workloads that have cross-border data-transfer restrictions. E2E's competitive position is the Tier-3 Indian cloud pedigree and the sovereign-data-residency guarantee.
Sovereign AI infrastructure thesis
The sovereign AI infrastructure thesis is the strategic context for Nvidia's India push. The thesis: AI is the new general-purpose technology, and countries that do not build their own AI infrastructure will be dependent on foreign AI providers for the most strategic applications (defense, healthcare, education, economic policy). The response: each country builds its own AI factory layer, its own model layer, and its own application layer.
Nvidia's role in the strategy is the GPU infrastructure layer. The three partnerships (Yotta, L&T, E2E) are the channels through which Nvidia technology reaches the sovereign AI infrastructure market. The combined infrastructure investment across the three partnerships is expected to reach 5B to 10B US dollars over 2026 to 2028, with the GPU cluster deployments as the largest share.
What to watch in 2026-2027
Three signals settle the rest of 2026 and 2027 for India's AI infrastructure buildout. First, the IndiaAI Mission deployments: the first 10,000-plus GPU clusters are expected to be operational by Q2 2027, with the AI applications (agriculture, healthcare, education) running on top. The deployment cadence is the proof point for whether the move faster message has been received. Second, the IndiaLLM and BharatGPT models: the Indian-language LLMs trained on sovereign infrastructure are expected to be released in late 2026 or early 2027. Third, the enterprise adoption: 2026 enterprise AI adoption in India is expected to grow 40 to 60 percent YoY, driven by the AI factory infrastructure, the falling cost of inference, and the maturation of the Indian AI talent pool.
What to budget for
Three cost paths settle the AI infrastructure decision for Indian enterprises. Cloud GPU rental (entry): Rs 50 to 200 per GPU-hour for H100 or H200 rental from Yotta, E2E, or AWS, Azure, GCP India region. Right pick for early-stage AI projects, experiments, and small-scale inference. On-premises GPU cluster (mid-market): Rs 5 to 50 Cr for a 100 to 1,000 GPU cluster (capex) plus Rs 2 to 10 Cr per year for operations. Right pick for mid-market enterprises with sustained AI workloads. Enterprise AI factory (large): Rs 50 to 500 Cr for a 1,000 to 10,000 GPU cluster (capex) plus Rs 10 to 50 Cr per year for operations. Add Rs 5 to 50 Cr per year for the engineering team to build, deploy, and operate the AI workloads.
Resources and next steps
- Nvidia newsroom: nvidianews.nvidia.com
- IndiaAI Mission official: indiaai.gov.in
- Yotta Data Services: yotta.com
- IndiaAI Impact Summit 2026 recordings: indiaai.gov.in/summit-2026
- Nvidia India partner directory: nvidia.com/en-in/partners
Fazlur Rahman
