---
title: "ESDS Software Solution — Deep Dive Analysis"
ticker: "ESDS"
date: "2026-09-15"
tags: ["deep-dive","fundamental","cloud","data-centre","ai","gpu","ipo"]
summary: "Up 309% from its ₹429 IPO in seven sessions on the AI/GPU story — but ESDS is the paying customer in the $1.25bn SharonAI deal. 25-point checklist score 16.5/25: avoid at CMP."
model: "deepseek-v4-flash"
watchlist: true
priceAtAnalysis: 1754
support: 1582
resistance: 1828
supports: [1438, 1308]
resistances: [1900]
verdict: "AVOID AT CMP"
---

# ESDS Software Solution — Deep Dive Analysis

**Report date:** 15 September 2026 | **CMP:** ₹1,754 (intraday, 15 Sep 2026 ~10:50 IST) | **Last close:** ₹1,740.40 (11 Sep 2026) | **Model:** deepseek-v4-flash (DeepSeek)

---

## 1. Quick Snapshot

| Metric | Value | Note |
|---|---|---|
| CMP | ₹1,754 (intraday, 15 Sep 2026) | Last close ₹1,740.40 (11 Sep) |
| IPO price / listing | ₹429 / ₹757 (4 Sep 2026) | Listing gain +76.5% |
| Return since IPO | ~+309% | In seven sessions |
| Market cap | ~₹20,550 Cr | 11.72 Cr shares post-issue |
| 52-week high / low | ₹1,827.40 / ₹746 | Both set within the last seven sessions |
| Stock P/E (FY26 post-issue EPS ₹10.31) | ~170x | On pre-issue EPS ₹12.03 → ~146x |
| Industry P/E / peer median | 25.8x / 26.8x | Premium is extreme |
| FY26 revenue | ₹472.21 Cr | +30.7% YoY |
| FY26 EBITDA / margin | ₹234.23 Cr / 49.6% | Margin up from 35.6% in FY24 |
| FY26 PAT | ₹120.82 Cr | +117% YoY |
| ROCE / ROE (FY26) | 30.2% / 24.9% | RHP RoNW 22.85% |
| Debt/Equity | 0.18 (borrowings ₹42.9 Cr) | Debt sharply reduced |
| Book value / NAV | ₹52.66 per share | Trading at ~33x book |
| Piotroski score | 8.00 | Strong |
| Debtor days | 128 | Elevated |
| Earnings yield | 0.93% | vs 10Y G-Sec ~6.5% |
| Promoter holding | 39.46% (Sep 2026) | 46.06% pre-IPO |
| Dividend | Nil | No payout history |

**Data sources:** Screener.in consolidated (logged-in, 27 sidebar metrics verified, 15 Sep 2026), NSE bhavcopy (delivery data), NSE/Yahoo price data, RHP-linked company disclosures, broker notes. Quarterly financials are **not available** — the company listed on 4 September 2026 and has not yet reported a quarter as a listed entity. This is flagged wherever it matters below.

---

## 2. Business Overview

ESDS Software Solution (incorporated 2005, Nashik) is an AI-enabled **cloud, managed services, data centre infrastructure and software** provider. It is one of only two Indian players offering the full stack of GPU-as-a-Service (GPUaaS), cloud, managed services, data centre infrastructure and software, and the larger of the two by FY26 revenue (Nexdigm report cited in the RHP).

**Revenue mix (FY26)**
- Infrastructure-as-a-Service (cloud + colocation): ₹207.2 Cr — 43.9% of revenue
- Managed services: ₹194.6 Cr — 41.2% (roughly doubled YoY on AI/GPU-related work)
- Software-as-a-Service: ₹70.4 Cr — 14.9%

**Operating assets**
- Five data centres: Nashik, Airoli (Navi Mumbai), Bengaluru, Mohali, Noida (opened Oct 2025); ~75,266 sq ft aggregate
- Two planned: Kolkata and Sahibabad/Ghaziabad
- Asset-light government model: operates 3 of STPI's 6 Tier-3 compliant data centres (Bengaluru, Mohali, Noida)
- Proprietary platforms eNlight and SWARAJ Cloud (India's first "Made in India" cloud, 2011); holds India's only patent for vertical auto-scaling cloud technology (filed in India and the US); Security-as-a-Service (SECaaS) framework

**Customer base (as of June 2026)**
- 2,516 customers in FY26 (1,465 in FY24; 1,714 in FY25)
- 104 government clients on its cloud, 115 banks and financial institutions across 1,045 branches, 113 organisations on its SAP-HANA cloud, six smart cities
- Revenue retention 94.92%
- Average revenue per customer **fell** to ₹18.8 lakh in FY26 from ₹21.1 lakh in FY25 — new customer additions are smaller accounts
- Largest customer was ~15.9% of FY26 revenue. The FY25 top customer (a Russian BFSI entity, 20.15% of revenue, ₹72.8 Cr) collapsed to 2.80% (₹13.2 Cr) in FY26 after sanctions; a new UAE customer contributed ₹75.24 Cr in FY26

**Moat assessment:** The genuine moat is regulatory, not technological — India-resident data on Indian servers under Indian jurisdiction, which matters to BFSI and government clients bound by RBI/localisation rules. That is a real, defensible niche against AWS/Azure/Google. The GPUaaS business, by contrast, is a reseller position with no evident structural moat.

---

## 3. Financial Performance

### Annual P&L (Screener.in consolidated, ₹ Cr)

| | Mar 2021 | Mar 2022 | Mar 2023 | Mar 2024 | Mar 2025 | Mar 2026 |
|---|---|---|---|---|---|---|
| Sales | 172 | 195 | 208 | 287 | 361 | 472 |
| Operating profit | 62 | 55 | 47 | 101 | 155 | 234 |
| OPM % | 36% | 28% | 23% | 35% | 43% | 50% |
| Other income | 2 | 3 | -0 | 6 | 15 | 8 |
| Interest | 18 | 18 | 26 | 32 | 25 | 12 |
| Depreciation | 37 | 43 | 49 | 53 | 62 | 63 |
| PBT | 9 | -2 | -28 | 22 | 83 | 167 |
| Net profit | 5 | -3 | -22 | 14 | 56 | 121 |
| EPS (₹) | 10.51 | -0.30 | -2.42 | 1.35 | 5.54 | 11.98 |

**Key takeaways**
- Revenue CAGR of ~22% over FY21-FY26, but the inflection is recent: +38% in FY25 and +30.7% in FY26 (₹472.21 Cr per RHP-linked presentation)
- EBITDA margin expanded 35.6% → 42.9% → 49.6% across FY24-FY26 — genuine operating leverage as data centres filled up
- PAT went from ₹13.6 Cr (FY24) to ₹55.6 Cr (FY25) to ₹120.8 Cr (FY26). Note the low base: FY24 PAT was tiny, so "8x profit growth in three years" overstates the steady-state run-rate
- Interest cost halved to ₹12 Cr as borrowings fell from ₹259 Cr (FY24) to ₹42.9 Cr (FY26) — the balance sheet genuinely deleveraged before the IPO

### Quarterly trend

**Not available.** Screener shows no quarterly columns for ESDS because the company listed on 4 September 2026 and has not filed quarterly results as a listed entity. Its first quarterly print (Q2 FY27, September quarter) is expected around November 2026. Any "quarterly growth" claim circulating in the market is derived from RHP annual/pre-IPO data, not from reported quarters.

---

## 4. Balance Sheet (Screener consolidated, ₹ Cr)

| | Mar 2023 | Mar 2024 | Mar 2025 | Mar 2026 |
|---|---|---|---|---|
| Equity capital | 9 | 9 | 10 | 10 |
| Reserves | 201 | 217 | 408 | 534 |
| Borrowings | 179 | 259 | 127 | 96 |
| Other liabilities | 111 | 63 | 112 | 1,298 |
| Total liabilities | 501 | 548 | 656 | 1,938 |
| Fixed assets | 208 | 322 | 353 | 388 |
| CWIP | 0 | 0 | 0 | 4 |
| Other assets | 293 | 225 | 299 | 1,541 |
| Total assets | 501 | 548 | 656 | 1,938 |

**What actually happened in FY26:** the balance sheet roughly tripled in size, driven by a **₹1,176.6 crore customer advance** received for a GPUaaS project (recognised in other current liabilities, ~₹1,196 Cr) and the corresponding cash (company cash rose from ₹61 Cr to ₹1,253 Cr, inside "other assets" of ₹1,541 Cr). This is money received before service delivery — it is a liability and an obligation to deliver, not earned income or shareholder value created.

Two further balance-sheet facts worth knowing:
- Borrowings fell to ₹96 Cr on the Screener consolidated view (₹42.9 Cr in the RHP presentation) — the company is genuinely lightly levered
- **Hypothecated current assets of ₹1,464.54 Cr, or 96.72% of total current assets** as of 31 March 2026 (RHP risk disclosure). Nearly all current assets are pledged as security, which constrains future borrowing flexibility even though reported debt is low

---

## 5. Cash Flow (Screener consolidated, ₹ Cr)

| | Mar 2021 | Mar 2022 | Mar 2023 | Mar 2024 | Mar 2025 | Mar 2026 |
|---|---|---|---|---|---|---|
| Cash from operating activity | 49 | 16 | 55 | 57 | 163 | 1,368 |
| Cash from investing activity | -57 | -69 | -54 | -1 | -112 | -130 |
| Cash from financing activity | 22 | 74 | -19 | -71 | 7 | -45 |
| Net cash flow | 14 | 21 | -18 | -15 | 58 | 1,193 |
| Free cash flow | -7 | -44 | 19 | 34 | 49 | 1,243 |
| CFO / OP | 71% | 47% | 131% | 49% | 108% | 594% |

**This is the single most important table in the report.** FY26 operating cash flow of ₹1,368 Cr against EBITDA of ₹234 Cr is **not** operating excellence — it is almost entirely the ₹1,176.6 Cr customer advance. Strip the advance out and FY26 CFO lands near ₹180-190 Cr, which is consistent with FY25's ₹163 Cr and healthy versus PAT. The advance will unwind as revenue: as ESDS delivers the GPU service, the liability converts to revenue, and the "amazing working capital" reverses (Screener's working capital days of -806 is a direct artefact). **Treat FY26 CFO/OP of 594% as non-recurring.**

Capex is rising as designed: the ₹720 Cr IPO is 100% fresh issue, with ₹576 Cr earmarked for cloud and data centre equipment in FY27-FY28.

---

## 6. Shareholding Pattern

| Holder | Sep 2026 (post-IPO) | Pre-IPO |
|---|---|---|
| Promoters | 39.46% | 46.06% |
| FIIs | 1.01% | — |
| DIIs | 8.50% | — |
| Public | 49.90% | — |
| Others | 1.10% | — |

Number of shareholders: 1,80,892.

Only one post-listing column exists (the company listed on 4 September 2026), so a trend cannot be established yet. Notes:
- Promoter stake fell to 39.46% purely from fresh-issue dilution — **there was no offer-for-sale**, i.e. no promoter or investor exit at the IPO
- DII holding of 8.50% reflects the ₹216 Cr anchor book (19 anchor investors at ₹429; 82% of it went to domestic funds) plus post-listing buying. Motilal Oswal Digital India Fund's NAV rose 5.19% on the listing move, confirming pre-existing MF exposure
- Well-known individual investors Ashish Kacholia and Mukul Agrawal have been reported as pre-IPO/early holders

**Watch:** pre-IPO shareholder lock-in expiries and the eventual release of supply are the biggest structural overhang. With a 100% fresh issue, no early investor got liquidity at listing — that liquidity comes later, into the market.

---

## 7. Order Book & Revenue Visibility

There is no conventional order book to report. What exists instead is one very large GPU commitment and one very large customer advance — and understanding the direction of both is essential:

**The $1.25 billion SharonAI agreement (signed 1 April 2026)**
- SharonAI Holdings (NASDAQ: SHAZ) provides GPU compute capacity; **ESDS and its subsidiaries are the CUSTOMER** paying for it — confirmed by SharonAI's SEC filings (8-K / 424B3), which define ESDS, SPOCHUB Solutions and ESDS Cloud FZ as "the Customer"
- ~8,208 NVIDIA B300 (Blackwell) GPUs and ~17.83 PB of storage, hosted in Australia (NEXTDC facility); revenue start September 2026, with a 16 September 2026 target delivery date
- Five-year term, ~$250 million per year (~₹2,100 Cr), **take-or-pay** (ESDS pays whether or not it uses the capacity); payments monthly in advance; ESDS cannot terminate for convenience in the first 36 months
- **$140 million of letters of credit / bank guarantees** required as security — roughly twice ESDS's pre-IPO total asset base
- ESDS's plan is to resell this capacity to Indian enterprises — "buy wholesale, sell retail". No binding India-side customer contracts have been disclosed

**The ₹1,176.6 crore advance**
- Received by SPOCHUB Solutions (subsidiary) from an overseas enterprise customer for a GPUaaS project
- Sits as a liability until the service is delivered; refundability has not been publicly clarified
- Management has guided GPU revenue to begin from **Q3 FY27**

**The maths to keep in front of you:** ESDS must sell ~₹2,600 Cr+ of GPU capacity annually (at a 25-30% markup) to cover a ₹2,100 Cr annual commitment. FY26 revenue was ₹472 Cr. That is a 5-6x step-up in annual revenue, from a business with no disclosed downstream order book at that scale.

Screener's peer and small-cap data show no order-book line for ESDS — unlike defence/EPC names, this company's "visibility" rests on this single arrangement.

---

## 8. Management Commentary

- CMD Piyush Somani has described the arrangement with a total potential contract value of $1.25 billion, with revenue generation from Q3 FY27, and in a Hindi interview framed it as a seven-year commitment
- Management has publicly characterised GPUs as "appreciating assets" and referenced a "$14 billion business funnel" without publishing the order-book information that would make it underwritable
- On 9 September 2026, following the post-listing surge, the exchange sought clarification on the price movement. The company replied that it had disclosed all material information, had **no undisclosed price-sensitive information** explaining the rally, and considered the movement market-driven
- IPO positioning: "one of only two players in India" offering the full stack, with the government/STPI partnership and data-localisation guarantee front and centre

**Assessment:** management communicates the AI opportunity aggressively and the contract's economics thinly. The gap between the narrative ("$1.25 billion contract", "$14 billion funnel") and the filing language (ESDS as the paying customer with a take-or-pay obligation) is the most serious disclosure-quality issue in this story, and any investor is entitled to discount the narrative until the first listed quarterly numbers appear.

---

## 9. Technical Analysis

**Critical context:** ESDS has traded for only **seven sessions** (listed 4 September 2026). Standard indicators — EMA/SMA 20/50/100/200, RSI(14) on daily, MACD, Bollinger Bands — **do not exist yet** (they need 15-200 sessions of history). Any source quoting them for ESDS is fabricating them. What follows is the price, volume and delivery evidence that does exist.

### Price and volume (NSE)

| Date | Close (₹) | Volume (shares) | Delivery % | Note |
|---|---|---|---|---|
| 4 Sep (listing) | 908.40 | 1,57,90,052 | 66.59% | Opened ₹757, closed +112% over issue |
| 7 Sep | 1,090.05 | 7,42,653 | 99.32% | Upper circuit — no supply |
| 8 Sep | 1,308.05 | 1,27,61,472 | 26.13% | Circuit again on heavy churn |
| 9 Sep | 1,438.85 | 3,12,887 | 99.42% | Upper circuit, minimal volume |
| 10 Sep | 1,582.20 | 1,56,76,632 | 17.79% | Huge delivery-volume day, low delivery |
| 11 Sep | 1,740.40 | 74,98,763 | 17.87% | Continued surge |
| 15 Sep (intraday) | ~1,754 | — | — | High ₹1,827.40 (all-time high), low ₹1,681 |

**Trend:** vertical. Up roughly 309% from issue price in seven sessions, hitting upper circuits in most sessions, with no consolidation of any kind.

**Volume/delivery reading:** the high-volume, low-delivery days (8, 10, 11 Sep: 17-27% delivery on 0.75-1.6 crore shares) are **churn, not accumulation** — intraday and momentum money exchanging hands, the classic post-IPO speculative tape. The ~99% delivery days are circuit days where sellers simply did not exist. Neither pattern is evidence of institutional accumulation.

**Bulk deal:** on 10 September 2026, a bulk deal of **6,34,148 shares at an average ₹1,482.04** was recorded on NSE. This is the only reported bulk/block deal in calendar 2026 and, being a small ~0.5% of equity, does not establish institutional backing either way.

**Key levels**
- Resistance: ₹1,827 (all-time high, 15 Sep intraday); above that, blue sky
- Support 1: ₹1,582-1,600 (10-11 Sep breakout zone)
- Support 2: ₹1,438 (9 Sep circuit close)
- Support 3: ₹1,308 (8 Sep close) — a break here would mark the first failure of the post-listing trend
- Structural reference: the listing-day range ₹746-908 defines where genuine two-way trade first occurred

**Signal balance:** no indicator-based bearish evidence (no indicators exist), but also no base, no higher-low structure, and no established trend to deploy against. This is a momentum tape, not a technical setup.

### 25-Point Swing Trading Checklist (all 25 points assessed)

| # | Point | Assessment | Verdict |
|---|---|---|---|
| 1 | Price position (within 25% of high / at ATH) | Trading at all-time high | ✓ |
| 2 | Clear uptrend | Rising every session, but parabolic with no base — unproven trend | ⚠️ |
| 3 | Volume characteristics | Huge volumes, but low delivery on the biggest days = churn | ⚠️ |
| 4 | Clean price action | Upper circuits, gap-ups, 15% intraday swings — not orderly | ✗ |
| 5 | EMA alignment (9>20>50>200) | Impossible to compute — only 7 sessions of history | N/A |
| 6 | Chart pattern / base formation | No base exists; straight vertical move from listing | ✗ |
| 7 | Base quality | No consolidation at all | ✗ |
| 8 | ADR suitable | Very high daily range — suits only aggressive momentum traders | ✓ |
| 9 | Primary uptrend (broad market) | Nifty ~23,342, market not in distress | ✓ |
| 10 | Follow one trading model | Discipline point — must be an explicit momentum/IPO model | ✓ |
| 11 | Sector strength | AI, data centre and cloud are the leading themes | ✓ |
| 12 | Tailwind | Data localisation, AI capex, GPU scarcity — strong tailwinds | ✓ |
| 13 | IPO consideration | Explicitly an IPO-base situation; lock-in supply pending | ✓ |
| 14 | Price action strength vs peers | Best-performing IPO of 2026; massively outperforming | ✓ |
| 15 | Sector & government boost | STPI partnership, IndiaAI push, BFSI localisation mandate | ✓ |
| 16 | Fundamental quality | Genuine growth and low debt, but earnings-quality questions (advance-driven CFO, SPOCHUB margin) | ⚠️ |
| 17 | Quarterly earnings growth (30-50%) | No quarterly data as a listed company | N/A |
| 18 | Delivery percentage | 17-27% on peak-volume days; ~99% only on no-supply circuit days | ⚠️ |
| 19 | News & institutional activity | Heavy news flow, one small bulk deal, anchor MF interest — no institutional accumulation evidence | ⚠️ |
| 20 | Risk management (entry/SL/target defined) | No structural support exists to define a stop; SL would be arbitrary | ⚠️ |
| 21 | Market environment | Supportive — index near 23,342, no systemic weakness | ✓ |
| 22 | Liquidity | Excellent — daily turnover ₹100 Cr+ | ✓ |
| 23 | Relative strength | Extreme outperformance vs index and sector | ✓ |
| 24 | Breakout confirmation | There is no base to break out from; entering here is chasing | ✗ |
| 25 | Trading psychology | Maximum FOMO risk; buying after +309% is the textbook late entry | ⚠️ |

**Score: 16.5 / 25 — below the 17-point threshold, with critical fails on items 6 (no base) and 24 (no breakout to confirm).** By this checklist's own rule, ESDS is a **pass, not an entry**, at the current price. It may become a valid setup later — if and when it forms a proper base after the first quarterly results.

---

## 10. Block Deals & Insider Activity

- **Bulk deal, 10 Sep 2026 (NSE):** 6,34,148 shares at an average ₹1,482.04 — the only reported deal in calendar 2026. Small relative to equity (~0.5%) and not identifiable as a strategic accumulation
- No block deals reported in calendar 2026
- No promoter selling reported; promoter stake change to date is pure IPO dilution
- Anchor book (₹216 Cr, 19 investors, 82% domestic funds) is subject to standard lock-in

*(Deal data per Trendlyne/NSE-reported bulk deal records; Trendlyne's page was behind a bot-verification wall at the time of writing, so this is corroborated via secondary reporting of the same NSE record.)*

---

## 11. Forward Valuation

### The starting point

- FY26 PAT ₹120.82 Cr; post-issue EPS ₹10.31 (11.72 Cr shares) → **~170x trailing earnings at ₹1,754**
- On pre-issue EPS of ₹12.03 → ~146x
- Market cap / FY26 revenue ≈ 43x
- Industry P/E 25.8x; peer median 26.8x (LTTS 26.8x, Tata Technologies 48.7x, Inventurus 40.3x, Netweb 107.6x)
- Screener PEG 1.62 (3-year profit CAGR basis) — flattered by the low FY23 base (a loss year)

### Analyst views on record

| Source | View | Target / Estimate |
|---|---|---|
| Choice Institutional Equities | Buy (initiating) | ₹1,550 — already exceeded. Projects revenue ₹472 Cr (FY26) → **₹4,581 Cr (FY28)**; valued 18x FY28 EV/EBITDA |
| Swastika Investmart | Partial profit-booking on listing-day gains | Stop-loss ₹650-680; fresh entry only on pullback to ₹600-650 |
| Bonanza | Wait | Wants the first post-listing results before acting on AI numbers |

### Credibility check on the aggressive case

Choice's ₹4,581 Cr FY28 revenue implies roughly a **9.7x increase in two years**. For that to happen, the GPU reseller business must scale from zero to multiple thousands of crores while the ₹2,100 Cr annual SharonAI commitment runs in parallel. Back-checking the implied economics: to earn ₹4,581 Cr of revenue at, say, a 25% gross spread on GPU resale, ESDS would need to sell ~₹6,000 Cr of capacity — beyond even the $1.25bn contract's nominal size within five years. The estimate is therefore best read as a bull-case ceiling, not a base case. Note also that the only target published (₹1,550) has already been passed by the market.

### Scenario table (FY28 horizon)

| Scenario | Revenue FY28 | Assumed PAT margin | PAT FY28 | EPS FY28 | Implied P/E at ₹1,754 |
|---|---|---|---|---|---|
| Bear — GPU ramp slips, advance partly refunded, core grows 20% | ~₹700 Cr | 20% | ~₹140 Cr | ~₹12 | ~146x |
| Conservative — core grows 25%, GPU contributes modestly | ~₹1,100 Cr | 20% | ~₹220 Cr | ~₹19 | ~93x |
| Moderate — core 30% + partial GPU ramp | ~₹1,900 Cr | 21% | ~₹400 Cr | ~₹34 | ~51x |
| Analyst bull (Choice) | ₹4,581 Cr | ~20-22% | ~₹950 Cr | ~₹81 | ~22x |

Even the moderate scenario leaves the stock at ~51x forward earnings — a premium to every peer except Netweb — and the bear case leaves it at ~146x. The bull case only works if the GPU business lands almost perfectly.

**Verdict on valuation:** the stock is priced for the bull case and then some. There is no margin of safety at ₹1,754 on any scenario except the analyst bull case, which itself depends on contracts that have not been disclosed.

---

## 12. Recent News & Catalysts

**Positive**
- FY26 results (RHP-linked): revenue +30.7%, EBITDA margin 49.6%, PAT ₹120.8 Cr — genuinely strong operating performance in the core business
- IPO subscribed ~136x (QIB 261x, NII 193x, retail 40x); listed 4 Sep at ₹757 and closed day one at ₹908.40 — best-performing mainboard IPO of 2026 so far
- ₹1,176.6 Cr advance received from an overseas GPUaaS customer — real cash, real counterparty commitment, though it is a liability until delivered
- Choice Institutional Equities initiated with Buy (TP ₹1,550), citing the full-stack position and AI contract
- $576 Cr of the ₹720 Cr IPO earmarked for cloud/data centre equipment in FY27-28 — pure growth capital
- Motilal Oswal Digital India Fund NAV +5.19% on the listing — evidence of existing MF interest

**Neutral / watch**
- Exchange clarification sought on 9 Sep; company confirmed no undisclosed price-sensitive information
- Sep 16, 2026 target delivery date for the 8,208-GPU cluster — first hard execution checkpoint
- First listed quarterly results due ~November 2026 (Q2 FY27) — the first audited look at post-IPO economics
- One bulk deal (10 Sep, 6.34 lakh shares at ₹1,482)

**Negative / risk**
- The "$1.25 billion contract" is a **purchase commitment** (ESDS is the customer), not a customer order — confirmed by SharonAI's SEC filings
- SPOCHUB Solutions (subsidiary) generated ₹84.95 Cr revenue and ₹53.96 Cr PAT in FY26 — a 63.5% net margin — contributing ~45% of consolidated PAT, from virtually nil revenue in FY25; the promoter personally acquired a 1% stake in SPOCHUB for ₹2,000 in May 2025
- FY26 operating cash flow is advance-driven and non-recurring
- 96.72% of current assets are hypothecated (RHP)
- Customer concentration: largest customer 15.9% of FY26 revenue; the FY25 top customer (Russian BSFI, 20.15%) collapsed to 2.8% post-sanctions and a new UAE customer (₹75.24 Cr) appeared — concentrated overseas exposure keeps shifting
- ARPU declining (₹21.1 lakh → ₹18.8 lakh) as smaller customers are added
- CRISIL historically classified the company "Issuer Not Cooperating"; upgraded to BBB+/Positive/A2 by Sep 2025 — historical, but part of the governance record
- RBI data localisation limits which clients can use Australian-hosted GPU capacity — a direct constraint on the resale thesis

---

## 13. Key Risks

| Risk | Severity | Why |
|---|---|---|
| GPU resale execution — no disclosed India-side customers against a take-or-pay ~₹2,100 Cr/yr commitment | **Critical** | The entire thesis depends on it; the commitment is payable regardless |
| Valuation — ~170x trailing earnings with no listed quarterly track record | **Critical** | Any disappointment in the first results or the Sep 16 delivery checkpoint has enormous downside room |
| Earnings quality — ~45% of PAT from a newly profitable subsidiary at a 63.5% net margin, plus promoter's ₹2,000 purchase of 1% of that subsidiary | **High** | Requires explanation before it can be underwritten |
| Advance unwind — ₹1,176.6 Cr liability converts to revenue only on delivery; refundability unclear | **High** | Failure to deliver turns cash today into an obligation tomorrow |
| ₹140M letters of credit / guarantees required | **High** | Roughly 2x the pre-IPO asset base; funding path depends on IPO proceeds and lenders |
| Post-listing technical structure — no base, parabolic, 17-27% delivery on peak volume | **High** | Churn-driven price with no support structure; drawdowns will be violent |
| Lock-in expiries (pre-IPO and anchor supply) | **High** | A 100% fresh issue means no early investor has exited yet |
| Customer concentration and shifting overseas accounts | **Medium** | 15.9% top customer; sanctions-driven replacement of one large account with another |
| RBI data localisation limiting GPU resale pool | **Medium** | Cuts the addressable market for Australian-hosted compute |
| Competition (hyperscalers, other Indian DC players) | **Medium** | Core cloud moat is regulatory; GPU resale has no moat |
| Liquidity/microstructure — circuit limits still binding | **Medium** | Entry and exit can be impossible at the extremes |

---

## 14. Screener Pros & Cons

**Pros (Screener)**
- Company has reduced debt
- Company has delivered good profit growth of 84.3% CAGR over the last 5 years

**Cons (Screener)**
- None flagged. Note this is a machine-generated checklist that keys off reported ratios — it does not capture the earnings-quality issues in this report (advance-driven cash flow, subsidiary margin anomaly, take-or-pay commitment), which is exactly why a checklist alone is not sufficient for a stock this unusual.

---

## 15. Bull Case vs Bear Case

**Bull case**
1. Real, profitable, 30%+ growth core business with sector-leading margins (49.6% EBITDA) and a genuine regulatory moat in Indian data residency
2. ₹1,176.6 Cr of cash received upfront from an overseas AI customer — evidence that someone with money believes in the GPU business
3. India's GPU scarcity plus 2,500 existing customers (104 government, 115 banks) gives a ready-made distribution channel for GPUaaS
4. Debt cut to ₹42.9 Cr and ₹576 Cr of IPO money funding capacity — a clean balance sheet going into the capex cycle
5. If the ramp works, revenue goes from ₹472 Cr to multiple thousands of crores; Choice's ₹4,581 Cr FY28 estimate is the top of that range

**Bear case**
1. ESDS is the **payer** on a $1.25bn take-or-pay GPU commitment with no disclosed downstream contracts — the risk sits entirely on its own book
2. ~45% of FY26 profit came from a subsidiary that was dormant a year ago and reported a 63.5% net margin — unexplained, and the promoter bought 1% of it personally for ₹2,000
3. FY26's spectacular cash flow is a customer advance, not earnings; strip it out and the business looks normal, not transformational
4. At ~170x trailing earnings after +309%, the first disappointing data point (first results, Sep 16 delivery, or a lock-in expiry) has no valuation cushion
5. 96.72% of current assets hypothecated, $140M of guarantees needed, and RBI localisation limiting who can buy Australian-hosted compute

---

## 16. Final Verdict

**Verdict: AVOID AT CMP — genuinely good core business, but the price and the AI story have run far ahead of what is disclosed.**

ESDS is not a bad company — it is a decent, profitable, fast-growing Indian cloud and data centre business with a defensible regulatory niche, low debt and real cash on the balance sheet. The problem is what is being bought at ₹1,754.

Three things must be separated, and the market is currently conflating them:

1. **The core business** — worth owning at a sensible multiple. FY26: ₹472 Cr revenue, ₹234 Cr EBITDA, ₹121 Cr PAT, ROCE 30%. Even assuming strong 25-30% growth, this core supports nothing close to ₹20,550 Cr of market cap.
2. **The GPU advance** — real cash (₹1,176.6 Cr), but a liability that must be earned back through delivery, and one that distorts every cash flow statistic for FY26.
3. **The SharonAI commitment** — a ~₹2,100 Cr/year **cost** obligation with no disclosed matching revenue. The market has been reading it as an order book. It is the opposite.

Add a 63.5%-margin dormant subsidiary contributing roughly 45% of group profit with the promoter personally holding 1% of it, and the correct posture is not "buy the AI story" — it is "wait for the first listed quarter and the delivery checkpoint."

**Technical verdict:** 16.5/25 on the 25-point checklist, with critical fails on base formation and breakout confirmation. There is no base to trade against and no structural level to place a stop.

**By investor profile**

| Profile | Recommendation |
|---|---|
| Short-term / momentum trader | Only for those already positioned; entering after +309% is chasing. If trading it, treat as pure momentum with a hard, arbitrary stop and small size — circuit limits can trap you |
| Swing trader | **Avoid.** No base, no confirmed breakout, 17-27% delivery on peak volume, no EMA structure to trade against. Re-assess if a 4-6 week base forms above ₹1,400 after the first results |
| Long-term growth investor | **Wait.** Revisit after Q2 FY27 results (~Nov 2026) and after the Sep 16 GPU delivery checkpoint. A pullback into ₹1,200-1,400 with the base intact and confirmed GPU revenue would be a more reasonable entry |
| Value investor | Not applicable — ~170x trailing earnings, ~33x book, 0.93% earnings yield |
| Risk-averse investor | **Avoid.** Earnings-quality questions, take-or-pay commitment, lock-in overhang, parabolic tape |

**Entry framework if, and only if, the story validates:**
- Base case: a proper 4-6 week consolidation after the first listed results, with delivery % sustained above 40% on up-days
- Levels: support ₹1,582 → ₹1,438 → ₹1,308; resistance = ₹1,827 all-time high, then blue sky
- Position sizing: sub-₹25,000 Cr small/mid-cap with this risk profile → cap at 2-3% of portfolio, not 5%
- Hard invalidation: a close below ₹1,300 (8 Sep level) breaks the entire post-listing structure

**What would change this view**
| Event | Impact |
|---|---|
| Q2 FY27 results show GPU revenue actually starting and clean (excluding the advance) | Bullish — first evidence the model works |
| Binding India-side GPU customer contracts disclosed | Bullish, materially |
| Sep 16, 2026 GPU cluster delivered and commissioned on schedule | Bullish — first hard execution proof |
| First results show SPOCHUB's 63.5% margin is not repeatable | Bearish — ~45% of profit at risk |
| Advance proves refundable / project stalls | Bearish — cash becomes a liability |
| Core business growth decelerates below 20% | Bearish — the valuation has no support |
| Forming a proper base with >40% delivery on advances | Constructive — would make it tradable |

---

## 17. Disclaimer & Sources

This analysis is for information and education only. It is not investment advice, and it is not a recommendation to buy, sell or hold. It was prepared by an AI agent (deepseek-v4-flash via Hermes Agent) from public sources; every material number should be independently verified against the company's RHP, exchange filings and audited accounts before acting. Newly listed, circuit-bound stocks are exceptionally volatile.

**Sources**
- Screener.in consolidated page for ESDS (logged-in session, 27 sidebar metrics verified) — annual P&L FY21-FY26, balance sheet, cash flow, ratios, shareholding, peer table
- NSE bhavcopy daily files (4-11 September 2026) — delivery quantities and percentages
- NSE/Yahoo Finance price data (4-15 September 2026)
- ESDS Software Solution RHP (24 August 2026) and abridged prospectus; company investor-relations financials FY24-FY26
- SharonAI Holdings SEC filings (8-K and 424B3) — SharonAI–ESDS GPU capacity agreement terms
- Economic Times, BusinessLine, IndiaInfoline, Moneycontrol, DataCenterDynamics reporting on the IPO and listing
- INDmoney rally analysis (revenue/EBITDA/PAT/borrowings/customer/ARPU series; FY26 advance disclosure)
- Credivant red-flag review (12 September 2026) — hypothecation 96.72%, SPOCHUB analysis, lock-in and delivery checkpoints
- Dhruva Substack forensic review of the RHP — SPOCHUB PAT contribution, ₹2,000 sub-1% promoter purchase, customer concentration
- UnlistedZone research note on the SharonAI agreement structure
- Broker commentary: Choice Institutional Equities (Buy, TP ₹1,550), Swastika Investmart, Bonanza

**Data limitations stated plainly:** no quarterly financials exist for ESDS as a listed entity; no moving-average, RSI or MACD data is available (seven sessions of history); Trendlyne's bulk-deal page was behind a bot wall, so the 10 September bulk deal is reported from the NSE record via secondary sources; the company's FY26 "other assets / other liabilities" composition is taken from Screener plus the RHP-linked disclosure of the ₹1,176.6 Cr advance.