This is what I am also concerned and can add to depreciation.
A very valid concern if you are an Nvidia shareholder. The problem is you are extrapolating this problem to E2E which is nothing more than an ant in this field. People are betting that the ant can become a slightly bigger insect and not that it will become an elephant.
9 Likes
A well written thesis on E2E and what the market is likely discounting at the current share price.
11 Likes
2026 can be the best year for E2E. Reasons as follow
- Increase in AI adaption by enterprises
- Voice agents are getting very good traction and specially in first line customer support. This will increase the inference and traning demand by X times in India.
- Data sovereignty is the need. It is not just voluntery compliance. Keeping critical Data outside India or In India with US hyperscaler under US cloud Act is the big risk
- Global Demand, Many Global companies will come to India for AI ready Infrastructure. They need it and In India we will have enough to atleast start onboarding some global companies.
- E2E has right to win, It is not about who has maximum GPUs at arsenal. But now it is all about who can optimally run a AI workload and generate good throughput from GPU clusters and E2E know this, Thanks to amazing support they get directly from Nvidia Team.
9 Likes
What is also very important is that this not GPU rental but rather AI software services. I am sure there is more to come and this revenue should have pristine margins and contribute heavily to the bottom line.
1 Like
This is not Software order. They have given GPUs to Zoho. They did the POC for sometime.
As new GPU comes in the order size should increase
1 Like
HI Hitesh, from where did you source this Zoho information. If you could also help us track infor on e2e. Thanks.
Looking to get views from the community on E2E Networks.
E2E recently undertook a large capex cycle (~₹1,200 crore), which significantly increased depreciation (~₹200 crore annually). At that stage, revenue growth did not scale proportionately, raising concerns around near-term ROCE and capital efficiency, and contributing to valuation correction.
Some factors that may change the outlook:
- Early signs of order inflow from India AI Mission and related AI infrastructure demand.
- Potential improvement in cost competitiveness versus hyperscalers like AWS/GCP, aided by currency movements. GCP cost A100 80 GB at $3600 vs E2E costing is $1120 for same machine. (3x+ difference)
- Domestic AI workloads and data residency requirements that may structurally benefit local GPU infra players if utilization improves.
Questions for discussion:
- What GPU utilization levels are needed for acceptable ROCE on the current asset base?
- How strong is revenue visibility from government and enterprise AI workloads over the next 12–24 months?
- How should steady-state margins and earnings be viewed once the capex phase normalizes?
Would appreciate insights from members tracking E2E closely.
2 Likes
Hi Devender,
On your question: 1. How strong is revenue visibility from government and enterprise AI workloads over the next 12–24 months?
There are 2 answers:
a) Yotta CEO says that there is already demand of 1 lakh GPU’s - which India AI mission is aware of and he expects this to be implemented in next 18-24 months. India's Compute Capacity To Surge: Yotta CEO Predicts 1 Lakh GPUs Within 18 Months
b) Also, the e2e CEO mentioned that they are in advanced stages of Blackwell as they can see robust demand. The truth is that e2e already has almost its entire capacity booked out from the 2 AI order.
1 Like
I know about GNANI order of India AI mission. Which company has given the 2nd order or how big is the second order?
4 Likes
The other order is by Gan studio. you can check all e2e orders from India AI mission on this website: India Al Cloud Computing Portal
1 Like
they cannot grow like markets are expecting without equity dilution & L&T has to infuse more and looks like promoters eventually wants to give it away to L&T and as valuation has been done on this blog post (taking equity dilution into consideration ) i feel everything is priced to perfection today with next 10 year of high growth assumptions.
3 Likes
Request you to share your assumption of why it is priced to perfection with next 10 year of growth.
At approx 5000 GPUs, E2E can do ~600cr revenue in FY27:
1) Almost 50% of the above is coming from India AI orders.
2) TTM 12 months revenue of 155 cr is without India AI orders (25% of 600cr). This revenue should also grow (assumption)
3) Remaining 25% of the 600cr is what E2E has to generate as incremental revenue through sales effort.
At 25% PAT margin- it’s 150 cr PAT. This business has operating leverage and blackwell has higher pricing - ~1.5x times hopper series. So , margin expansion should happen.
Valuation: At 4000cr mkt cap - this translates to ~25 PE. Lets assume +/- 20% variance in estimates - so the FY27 P/E can be 20-30x.
Growth beyond FY27E: With current debt capacity of ~500cr , 360 cr of cash yet to be deployed from the L&T fund raise and internal Cash flow generation of atleast ~100cr - E2E should be able to buy between 1000- 2000 GPU’s. This implies another 20-40% growth in FY28E.
To me - E2E lookes like 2x/3x in next 2 years.
Humbly request and welcome criticism and doubts on my above thesis. Opposite views will help me fill gaps on my current thesis. :-)
7 Likes
yes that’s what i mean by priced to perfect execution - increasing revenue by 4-5x in 3 years (58% cagr), if that happens like you said its trading at 25-30x earnings.
2nd thing you kind of missed is even if they be able to execute this they have to depreciate all of it in 3 years something everyone is debating today that what is life of these GPUs, Given the roadmap of Nvidia , AMD and TPUs - in 3 years they all are talking about 10x performance improvement.
That’s where they have to get another larger order from somewhere Gov + others and replace everything to state of the art that will require lots of capital and equity dilution.
and lastly we all are hoping this today’s trend will continue - hoping there is no AI bubble and there will be no AI bubble 3 years from now.
you can check detailed valuation here - kind of gives good prospective my only problem is despite so much of growth you are only expected to make 15-20% cagr by considering there can be n numbers of things which could go wrong E2E Networks Limited - by Dhruva Pandey
4 Likes
I think you are not considering depreciation of approx 150-200 cr for FY27?