ESDS Software Solution IPO: I am bullish, but not blindly. This is that rare IPO where I actually like the business, the growth and the use of proceeds together - even though the valuation is not cheap. I went through the 25 August 2026 RHP, the FY24-26 financials and the live IPO data, and there are genuinely interesting things here - plus one or two I would investigate much harder before calling this a straightforward "AI/cloud growth" story.
The IPO in brief
ESDS is raising ₹720 crore, a 100% fresh issue - no OFS, so the money comes into the company. Price band ₹408-429; at the upper band, post-issue market cap is roughly ₹5,028 crore. And this is overwhelmingly a capex IPO: about ₹576 crore (80%) goes into cloud and data-centre infrastructure, the balance to general corporate purposes. It is not generic capex either - the RHP breaks it down to roughly ₹96 crore for 80 cloud-node servers, ₹170 crore for 20 high-spec GPU servers, and ~₹44 crore for storage. The issue opened 28 August, closed 1 September 2026 (allotment 2 September), and was subscribed an extraordinary ~135.9x.
What I like
Start with the numbers, because they are genuinely good.
| Metric (₹ cr) | FY24 | FY25 | FY26 |
|---|---|---|---|
| Revenue | 286.5 | 361.3 | 472.2 |
| YoY growth | - | 26.1% | 30.7% |
| EBITDA | 101.9 | 154.9 | 234.2 |
| EBITDA margin | 35.6% | 42.9% | 49.6% |
| PAT | 13.6 | 55.6 | 120.8 |
| PAT margin | 4.8% | 15.4% | 25.6% |
Over FY24-26 that is a ~28% revenue CAGR, ~52% EBITDA CAGR and ~198% PAT CAGR - earnings growing far faster than revenue because margins expanded dramatically. One honest caveat: FY24 PAT was only ₹13.6 crore, so the ~198% PAT CAGR is flattered by a very low base. What reassures me is that the margin expansion is not an other-income mirage - other income actually fell (₹15.3 cr FY25 → ₹8.4 cr FY26) while finance costs dropped sharply. This looks like real operating leverage.
And the balance sheet went the right way while all this happened. Borrowings fell ₹149 cr → ₹62.7 cr → ₹42.9 cr, taking D/E from 0.66x to 0.08x. FY26 finance cost was just ₹11.79 crore - about 5% of EBITDA, with interest coverage around 15x. Debt is simply not a concern here.
Three more things I like: the integrated offering (cloud + data centre + managed services + SaaS - one vendor, not five); proprietary technology in eNlight Cloud and its auto-scaling, which is more than just renting server capacity; and a sovereign-cloud positioning that matters a great deal to the Government and BFSI customers for whom data localisation is non-negotiable.
The business
ESDS is a full-stack, AI-enabled cloud and data-centre player - IaaS + Managed Services + SaaS - serving 2,501 customers across Enterprise, Government and BFSI, from 5 Tier-III data centres (>75,266 sq ft). The FY26 revenue mix is well spread: IaaS ~₹207 cr (44%), Managed Services ~₹195 cr (41%), SaaS ~₹70 cr (15%).
The market behind it is large: the RHP pegs India's cloud market at ~$8.3bn in CY23 rising to ~$24.2bn by CY28 (~24% CAGR), with data-centre colocation growing ~21%. So the TAM is not the problem. The real question is how much of that ESDS can capture profitably against AWS, Azure, Google Cloud and large domestic operators.
Valuation, and the peer problem
Now the part that keeps me honest. At ₹429, on a post-issue diluted EPS of ~₹10.31, ESDS is priced at ~41.6x FY26 earnings. That is not cheap. And if it lists around the current grey-market premium of ~₹250 (~58% over the band), you are effectively buying it near ~66x on day one.
The peer check does not help. ESDS names E2E Networks as its listed comparable in the RHP - but E2E is loss-making, so its P/E is meaningless. There is no genuinely useful, profitable listed peer to anchor against, which means at ~42x you are paying for future growth - particularly the AI/GPU expansion - not for a discount to comparables.
The risks that matter
I am bullish, not blind.
- Customer concentration. Top customer = 15.93% of FY26 revenue; top 10 ~45%. Meaningful.
- Government dependence. 27.4% of FY26 revenue - stability, but tender, renewal and payment-cycle risk.
- Competition. AWS/Azure/Google and large domestic operators have vastly deeper pockets, and they set the price.
- Capex execution. ₹576 crore into GPUs, servers and storage - hardware that dates quickly. The real question is utilisation and return on that ₹576 crore, not just whether revenue grows.
- Geopolitical. A Russian BFSI customer hit by sanctions saw its ESDS revenue collapse from ₹73 cr (FY25) to ₹13 cr (FY26) - a live example of an external shock flowing straight to the top line.
- Litigation / contingents. Contingent liabilities ~₹55 cr (mostly ₹52.6 cr performance bank guarantees, plus GST demands); ongoing tax and other proceedings; and a Supreme Court dispute with SUSE LLC in which ESDS's SLPs were dismissed in May 2026.
The two things I would investigate hardest
This is where I go from "nice IPO" to "do the forensic work first."
1. The $1.25bn Sharon AI deal - and who is the customer. ESDS has signed a five-year, ~$1.25bn AI-infrastructure agreement with Australian neocloud Sharon AI, involving ~8,000 NVIDIA B300 GPUs. It is easy to misread this as ₹11,000+ crore of ESDS revenue. It is not. ESDS is the customer here - committing to take AI capacity from Sharon AI - which makes it simultaneously a huge downstream revenue opportunity and a huge capacity commitment that ESDS must then monetise. That single distinction is, to me, the most important diligence question in the entire IPO.
2. The ₹1,188 crore customer advance. FY26 operating cash flow was an extraordinary ₹1,367.7 crore on just ₹472 crore of revenue, and cash jumped from ~₹61 cr to ~₹1,253 cr. The reason is ~₹1,187.6 crore of customer advances on the balance sheet at March 2026. An advance can be fantastic - if it is genuine, contracted business. But I want to know exactly who paid it, under what contract, and what obligation ESDS carries against it before reading that cash pile as a pure positive. This matters far more to me than the headline ₹120.8 crore PAT.
My take
Put it on a scorecard. Business, growth, margins, debt, market size and IPO utilisation are all genuinely strong. Valuation is expensive (~42x). The peer comparison is nearly useless because the named peer loses money. Customer concentration is worth watching. The AI/GPU opportunity is potentially transformational but capex-intensive. And the cash flow needs forensic understanding, because ₹1,188 crore of customer advances drove the FY26 surge.
The one sentence I would remember: ESDS is not a bad company being sold at a good price - it looks like a good, fast-growing company being sold at a price that already assumes a meaningful part of the AI/GPU growth story.
So I like this IPO, because the earnings growth, margins, falling debt and genuine capex give me a business I would be comfortable tracking well beyond listing day rather than flipping on. But the real question the valuation forces is simple: can ESDS grow into the price? And the honest answer runs through the ₹1,187 crore advance and the $1.25bn Sharon AI commitment - understand those two, and you understand whether 42x is expensive or fair.
Good business. Expensive valuation. Worth watching.
Educational only, not investment advice. DYOR.
Sources: ESDS Software Solution RHP (SEBI filing, 25 August 2026) and abridged prospectus - issue size, price band, use of proceeds, FY24-26 financials, revenue mix, customer concentration and contingent liabilities; Economic Times, IPO360, IPrOspecta, Sushil Finance, Paytm Money, InCred Money and other IPO trackers; the Sharon AI agreement per Sharon AI / SEC disclosures; GMP and ~135.9x subscription per Economic Times / IPO Watch (as of 2 September 2026). Figures are approximate and as of the IPO - verify live before acting.
Follow along: @tyrovirtuoso on X