Netweb Technologies India Ltd Q4 FY2026 Concall Summary

Key Takeaways

• Landmark year, AI-led: Revenue reached ₹2,183.6 crore, up 90% YoY, with the AI segment surging 459.6% to contribute 43.4% of the total, a result of years of in-house R&D and disciplined execution on large national-scale orders.
• Guidance implies deceleration: Management guided 35–40% revenue growth for the next couple of years, a sharp slowdown from FY26's 90%, with EBITDA margins held at 13–14% despite a strategic order book exceeding annual revenue.
• Order book is the anchor: A disclosed order book of ₹2,100 crore, rising to ₹2,400 crore including L1 bids, provides near-term visibility, though the absence of recent large L1 wins drew analyst questions.
• Balance sheet stretched for growth: Net borrowings rose ~₹270 crore to fund working capital for strategic orders, while inventory days climbed to 86 as the company stockpiled critical components amid global AI supply pressure.
• What to watch next quarter: Execution on the strategic order book over the next three quarters, any new large deal wins to replenish the pipeline, and whether the base business sustains its 30–35% CAGR.

The AI Transformation

• AI segment dominance: AI systems revenue grew 459.6% YoY to become 43.4% of total operating revenue in FY26, driven by national-scale orders and in-house design and manufacturing capabilities for latest-generation AI systems.
• Segment mix reset: Management now expects AI to remain around 35% of the business going forward, with the balance shared equally between HPC and Private Cloud, a structural shift from the earlier 25–30% AI expectation.
• Private Cloud not shrinking: The segment fell to 24% of revenue only because AI grew much faster; in absolute terms, Private Cloud is growing with good momentum.
• Product readiness for inference: The company has added products that can handle inference exclusively or both training and inference, positioning for what management sees as an additional demand layer on top of training.
• AI demand runway: Management sees AI demand continuing for at least the next 18 to 24 months, with component pricing pressure persisting as global demand remains unabated.

Sanjay Lodha: "The AI demand is unabated. It's very difficult to fulfill the domestic demand itself. So really, frankly speaking, we are -- the focus on exports, we are not getting time to do it."

Order Book & Strategic Deals

• Order book exceeds annual revenue: Disclosed order book stands at ₹2,100 crore, rising to ₹2,400 crore including L1 bids, which is more than FY26's full-year revenue and provides near-term visibility.
• Strategic order execution timeline: Large strategic orders are expected to be completed phase-wise within the first three quarters of FY27, with the organic order book building out within 18 to 20 weeks.
• Pipeline conversion dynamics: Approximately 60% of the pipeline converts, with conversion taking around 18 months total; the pipeline is dynamic with new entries and exits.
• L1 win gap unaddressed: An analyst noted no meaningful large L1 wins in recent quarters; management responded with general commentary on robust order book and AI demand pipeline without providing specific reasons for the gap or naming new opportunities.

Financial Health & Margins

• Margin resilience explained: PAT margin held at 9.3% for FY26, which management attributes to disciplined execution despite high top-line growth; Q4 EBITDA margin was sequentially flat at ~12.5% because strategic orders spilled over, and adjusted for forex, margins were better than last quarter and within guidance.
• Working capital stretched: Net borrowings increased ~₹270 crore, largely short-term, to fund working capital for large strategic and regular orders; receivable days improved to 86 from 114, but inventory days rose to 86 due to raw material buildup for large orders and securing key inputs.
• Cash conversion cycle: Stood at 84 days as of March 2026, expected to operate in the 90–110 day range going forward, with some further increase in inventory days projected.
• Capex light: No significant capex expected in FY27, only routine capex of ₹20–25 crore, as the company is capability-driven rather than capacity-driven, with utilization at only 65–70%.
• Gross margin question deflected: An analyst asked why gross margins did not improve sequentially despite a higher share of base business revenue in Q4; management pointed to strategic order spillover and forex-adjusted margins without directly addressing the base business mix question.

Sanjay Lodha: "You have to understand the basics of the company, actually, the company doesn't work on capacity. We work on capability. If tomorrow, NVIDIA's new Vera Rubin is coming up as a new, we are already set up for the partnership for that."

Manufacturing & Infrastructure Build-out

• New facility commissioned: A 15,000 square foot state-of-the-art production facility was commissioned to manufacture systems not previously made in-house, including the complete range of liquid-cooled systems, designed to scale beyond 150 kilowatt per rack.
• India AI Mission momentum: Government investments are flowing through two routes: facilities via CSPs and on-premises infrastructure; the first tranche is already impacting the country, with enterprises and CSPs all building facilities and data center investment increasing from the corporate side.
• Government spending not budget-constrained: The government's current AI budget is expandable; the main criterion is speed of spending and infrastructure build-out, not the budget ceiling.
• Sovereign market focus: India is not a hyperscaler-driven market currently; it is primarily a sovereign and enterprise-driven market, and Netweb is not targeting hyperscalers at this point, though it can bid if they show interest.

Sanjay Lodha: "Government's current spending budget, they are even ready to expand that also if they're able to do it. So it's really -- the budget is not the main criteria there. The criteria is how fast they are able to spend the money and how fast they are able to get in the new infrastructure."

HPC & the Agentic AI Question

• HPC growth understated: HPC grew around 31% but appears low only because AI grew much more; it is growing well, driven by the National Supercomputing Mission 2.0 and adoption of HPC+AI across verticals like automobile, oil and gas.
• Agentic AI not an immediate threat: Management does not believe Agentic AI will significantly impact HPC sales in the current year in this geopolitical region; Intel-related developments will take more time to create impact on HPC market growth in India.
• Inference as additive demand: Inference demand is only going to go up and will generate additional demand on top of training; the company is ready with products for both.

Guidance & Commitments

Key Metrics