E2E Networks Limited — Q1 FY25 earnings call

E2E Networks reported robust financial results for Q1 FY25, with revenue from operations surging 112% year-on-year to ₹41.7 crores. This growth was accompanied by a 168% increase in EBITDA to ₹27.4 crores, leading to a significant EBITDA margin expansion of 1403 basis points to 66.34%. Net profit also saw a substantial rise of 44% YoY to ₹10.1 crores, translating to a diluted EPS of ₹6.75, up 42% YoY. Quarter-on-quarter, revenue grew 40% from ₹29.6 crores in Q4 FY24, and EBITDA increased 79%.

The company emphasized its position as a leading Indian player in the cloud GPU market, with the majority of its workloads now in the AI/ML domain. Management highlighted the superior compute power of GPUs, capable of delivering 1000x the compute of a typical CPU server for massively parallel AI/ML tasks, despite being 10x more expensive and consuming 10x more power. This shift to high-density GPU compute is a 'broad trend' contributing to the reduction in data center costs as a percentage of revenue and is expected to sustain margin growth in the medium to long term.

E2E Networks confirmed that the deployment of 185 crores CAPEX from Q4 FY24 was completed by early April, with the overall inventory, including other assets, now utilized at over 90%. The company's Monthly Recurring Revenue (MRR) for June stood at ₹14.5 crores, reflecting 90-95% utilization of its 450 AI H100 units. To meet ongoing demand, E2E is already in the process of acquiring another 256 H100 units, which are expected to go live in the current quarter (Q2 FY25).

Management articulated a significant opportunity for India in the cloud GPU sector, noting that while India accounts for 20% of the world's data, it only has 2% of the world's compute capacity, necessitating a 10x increase in compute. This presents a chance for India to 'leapfrog directly into the GPU segment.' E2E is also seeing 'some level of traction outside India,' with a prediction that 25-30% of its AI diffused deployment could be for overseas customers 'over a year.'

E2E Networks differentiates itself through its proprietary cloud software platform, which has been developed over the last decade, with 4-5 years specifically on cloud GPU. This platform, which supports 25-30 different microservices and focuses on national language processing, computer vision, LLMs, and diffusion models, is seen as a 'product business' rather than a services business. The continuous evolution of this software, driven by customer feedback and industry trends, creates 'stickiness' and higher lifetime value for customers, enabling the company to scale its business 10x without a proportional increase in team size.

The company clarified that there was no formal announcement of ₹800 crores CAPEX for FY25, but rather an aspirational figure. E2E maintains a flexible CAPEX strategy, adapting investments based on current demand and pipeline, with a shortened hardware planning cycle of 6-8 weeks. Future CAPEX will be funded through a judicious mix of vendor financing, debt, internal accruals, and potentially some equity raise. Asset turns are expected to remain between 0.5 and 0.6.

E2E Networks is exploring three avenues for government business following its MeitY impanelment, including supporting MSME partners, evaluating direct workload matches (18-24 months outlook), and joint ventures with other partners. The company is also strategically shifting towards a 'higher ARPU customer' profile, moving away from a large number of small customers. While this led to a higher customer concentration of 45-50% from the top 10 clients in Q1 FY25, management noted the 'bursty' nature of GPU workloads means this should be viewed over a full year.