Netweb Technologies India Ltd Q4 FY2025 Concall Summary

Key Takeaways

• FY25 was a breakout year, driven by AI: Revenue rose 58.7% YoY for the full year, with the AI segment surging 112% and contributing 14.8% of the top line, as enterprise and government adoption of private LLMs accelerated.
• Growth commitment held firm at 35–40% CAGR: Management reiterated confidence in sustaining this pace for the next few years, anchored by a booming data center market and a design-in-advance pipeline that converts in 8–12 weeks.
• Margins are deliberately capped to capture share: Operating EBITDA margin is guided at 13–14%, a level management calls comfortable versus single-digit competition, with only 30–40 bps of operating leverage expected next year.
• Cash conversion is the price of 60% growth: The cash conversion cycle improved to 73 days, but management was explicit that it will not sacrifice growth to optimize this number further, framing the tension as a mathematical outcome, not a risk.
• What to watch next quarter: Whether the India AI mission RFP enters the pipeline and can pull the growth trajectory above the 35–40% baseline, and whether the AI revenue share tracks toward the 19–20% target for FY26.

AI: The Growth Engine

• Segment scale and trajectory: AI contributed 14.8% of FY25 revenue with 112% YoY growth, and management expects it to reach 19–20% of revenue by the end of FY26, implying continued outsized expansion within the overall 35–40% CAGR.
• DeepSeek as a demand catalyst: The emergence of DeepSeek has opened opportunities by convincing Indian entities that previously thought they could not build fundamental models to now attempt it, exponentially increasing demand for AI compute.
• Skylus.ai fills a gap no single competitor addresses: Launched in FY25, this perpetual-license GPU infrastructure product has no direct single-product rival; a combination of 2–3 competing products can only come near it, and adoption is increasingly going up.
• India AI mission is the unbooked upside: The sovereign AI mission is not yet included in the order funnel; management will reflect it only when an RFP is a few months from release, leaving a potential growth accelerant outside the committed 35–40% baseline.

Lodha: "DeepSeek has rather opened a lot of opportunities for us... people who are thinking that they cannot build fundamental models, they also started thinking of building their fundamental models. So, all that definitely increases the demand for AI compute, exponentially actually."

Margin & Pricing Philosophy

• EBITDA margin guided at 13–14% for FY26: Management sees this as a comfortable band, noting that competition operates on single-digit margins, and prefers to price optimally to capture a large market rather than charge higher margins early.
• Operating leverage is modest by design: Only 30–40 bps of Y-o-Y margin improvement is expected from operating leverage, even if growth continues at 35–40% or higher, because the company is reinvesting to stay price-competitive.
• Full-stack integration protects margins: The company designs and manufactures hardware completely in-house, owns its software stack, and runs its own supercomputing R&D, a structure that structurally supports better margins than assemblers or resellers.
• Software is bundled, not broken out: Software and hardware are sold together as a solution; software revenue does not appear separately on invoices, so margin attribution between the two is not disclosed.

Lodha: "Our competition normally is all on the single-digit margin, wherein basically we are around 14%. So basically, we are already at a comfortable position. And plus basically, since there is a huge market to be captured, we want to price our products optimally and just not just try to basically charge margins initially or something of that nature."

Cash Conversion & Balance Sheet

• Cash conversion cycle improved to 73 days from 88 days QoQ: The improvement was driven by a deliberate push for early payment from large customers, though management expects the cycle to stay in a 70–90 day range going forward.
• Growth is the structural drag on cash: Growing at close to 60% means a chunk of revenue is continuously reinvested into working capital; as growth smoothens to 35–40%, cash will start coming back, but management will not sacrifice growth to optimize the metric.
• Quality of receivables is a mitigant: Management described negligible bad debt and kosher aging, arguing the cash conversion cycle is not a risk but a mathematical consequence of the growth rate.
• Balance sheet remains unlevered: Net free cash rose to ₹162.13 crore from ₹73.72 crore in December 2024, and the company carries zero net debt, with ₹170 crore cash on hand.

Sancheti: "I don't believe to call it as a risk off because our quality of cash conversation cycle is extremely good, and I'm sticking my neck out to say that. But you tell me, if you were me, would you sacrifice growth for the kind of business that we are to get a better cash conversion cycle?"

Order Book, Pipeline & Conversion Dynamics

• Order book is a poor lens for this business: The execution cycle is 8–12 weeks, so the order book is typically consumed in the next quarter, making backlog a less meaningful indicator than for longer-cycle peers.
• Pipeline is heavily government-skewed by visibility, not by end-market: Private-segment orders convert very immediately and do not sit in a similar structure; what shows up in the pipeline is largely government business, where gestation is longer and RFPs are designed months in advance.
• Design-in work starts 6–8 months before an RFP: The company proactively works on solutions before an RFP is published, meaning pipeline can convert in 30 days, 6 months, or a year depending on project complexity, with supercomputer planning starting earliest.
• Order book conversion is effectively 100%: Even for L1-stage orders, conversion is treated as certain, reflecting the bespoke, pre-designed nature of the solutions.

Lodha: "Our company cannot be really judged in terms of order book because order -- basically, order execution cycle for us is somewhere around 8 weeks to 12 weeks."

Business Mix & Segment Trajectories

• Government-enterprise split is balanced at 50:50: Both segments are large-enterprise in character, with government payments sometimes arriving faster than private ones because all sales are on 90-day credit terms.
• HPC and private cloud are secular growers: HPC has expanded from research into oil and gas, consumer, automobile, and manufacturing; the data center market is booming and not expected to cool soon, supporting a sustained good mix for both segments.
• AI enterprise workstations will reach ~20% of revenue: The shift from 15% to 20% will not materially alter the mix of other segments, because the overall base is also increasing.
• Exports will stay at 5–7% of turnover: Domestic demand has been phenomenal, and growing export share is difficult when the company is already growing 40% overall; management expects exports to remain consistent and might rise 1%, but will not guide beyond that.
• Services are a pull-through, not a standalone business: As the installed base grows, the company picks support and services contracts selectively where it has domain expertise, such as oil and gas, but does not position itself as a software or services company.

Competitive Position & Partnerships

• NVIDIA relationship is a deep OEM partnership: Netweb is among very few partners globally that design and develop products for NVIDIA, though margin-sharing or co-development terms were declined as confidential.
• Skylus.ai is a moat-widening product: No single competing product exists in the market; the closest alternative is a combination of 2–3 different products, and the perpetual-license model brings ease of usage and competitive advantage as AI systems adoption increases.
• PLI 2.0 validates the Make-in-India thesis: The first claim of ₹5.94 crore under the IT Hardware PLI scheme was received, underscoring the company's positioning as a domestic manufacturer competing with MNCs.
• SAP S/4 HANA deployment supports scalable growth: The system went live to enhance control, real-time insight, streamlined workflows, and stronger governance, a foundation-laying move for the next phase of scale.

Vikram: "Overall, if you see as such a single product available in the market, which can compete with Skylus.ai, we don't find any such option available."

Guidance & Commitments

Key Metrics