Esds Software Solution Q1 FY27 earnings call
In brief
ESDS Q1 FY27 revenue INR 133.66 cr (+7.28% YoY), PAT INR 29.28 cr (+14%); Sharon AI launch delayed to November.
- Management's tone
- Confident
- What was said
- Leaned positive
- Guidance
- None given
- Analyst pushback
- Low
- Stock, next session
- −5.00% (Nifty 50 +0.34%)
- Q1 FY27 revenue INR 133.66 cr (+7.28% YoY), PAT INR 29.28 cr (+14% YoY) and EBITDA margin 42%.
- Sharon AI's 8,200 B300 GPUs at Sydney delayed ~1.5 months; now expected live by early-mid Q3 FY27 (Nov first week).
- Plans to deploy ~1,500 own GPUs by Jan-Feb 2027 (mid/end Q4 FY27); funded 30% advance, 30% debt, 30% IPO.
- Domestic order book INR 3,000 cr over 3 years with 30-40% CAGR; international GPU pipeline exceeds 50,000 GPUs.
- Revenue mix Q1 FY27: IaaS 51%, Managed Services 31.31%, SaaS 17.55%; mix expected to shift further toward IaaS.
An AI read of the company's transcript · the filing
The numbers
What moved the numbers, as management explained it
- QoQ revenue fell 20.20% because Q4 FY26 included one-time technical design services revenue from subsidiary SPOCHUB for Sharon AI factory design. (one-off)
- Tax expense fell to INR 71.90 mn from INR 89.96 mn YoY, lifting PAT growth (14%) well above revenue growth (7.28%).
- Total expenses rose 8.42% YoY (INR 987.04 mn vs INR 910.39 mn), slightly outpacing revenue growth.
- Seasonality: Q1 is historically a softer quarter; management expects stronger Q3 and Q4 as AI factories bill.
The numbers management led with
- FY27 capex target: Approximately Rs 1,500 crores (with additional Rs 1,500 crores also referenced)
- International GPU pipeline: More than 50,000 GPUs in funnel
- Sharon AI deployment size: 8,200 B300 GPUs at Sydney AI factory, 7-year lease
- Own GPU fleet target: ~1,500 GPUs to be live by Q4 FY27
The company's filed results for this quarter are not on file with us yet; these are management's own figures from the call.
Guidance
Guidance on this call
| What | For | What management said |
|---|---|---|
| PAT margin on AI factory deals | — | We are targeting PAT margins in the range of 15% to 20%. |
| Domestic business CAGR | — | Domestic, currently, we are seeing a CAGR of 30% to 40%. |
| FY27 capex (concluded + additional) | FY27 | FY27, we are targeting to conclude around INR1,500 crores of capex, and additional INR1,500 crores. |
| Domestic order book | — | Domestic order book around INR 3,000 crores with 3-year duration. |
| Own GPU deployment target | Q4 FY27 | Target to touch around 1,500 GPUs, deploying by January or February next year. |
The business
By business
Infrastructure-as-a-Service (IaaS)
Largest vertical at 51% of Q1 FY27 revenue; covers cloud IaaS, GPU IaaS and co-location; growth driven by AI factory ramp with 8,200 B300 GPUs at Sydney going live next quarter.
IaaS share 51% of Q1 FY27 revenue · Industry pricing $5-6 per GPU per hour
Outlook: Share expected to stay steady and rise further as Sharon AI Sydney deployment goes live and own 1,500-GPU capacity is commissioned by mid/end Q4 FY27.
Managed Services
31.31% of Q1 FY27 revenue; people-based services for customer cloud and AI factory operations; FY26 had a one-time design services spike from SPOCHUB subsidiary.
Managed Services share 31.31% of Q1 FY27 revenue · FY26 one-time technical design services revenue from SPOCHUB
Outlook: Mix expected to shift toward IaaS as Sharon AI moves from design phase to live operations; people count held near 1,000.
Software-as-a-Service (SaaS)
17.55% of Q1 FY27 revenue; anchored by Swaraj Cloud platform with new launches Swaraj Garuda (APM) and Swaraj Jatayoo (database activity monitoring).
SaaS share 17.55% of Q1 FY27 revenue
Outlook: Management called SaaS the highest-margin vertical; no separate FY27 target given.
Balance sheet, capex and funding
- Customer advance ~INR 1,177 cr from Sharon AI, amortised monthly over the contract term.; working_capital
- IPO proceeds INR 720 cr received post 4 Sept 2026 listing; ~1,90,000 shareholders on the register.; funding
- FY27 capex target ~INR 3,000 cr (INR 1,500 cr concluded + additional INR 1,500 cr).; capex_planned
- Capex funding mix: 30% customer advance, 30% debt/lending, 30% IPO proceeds for own 1,500-GPU India build.; funding
- AGM postponed via ROC approval to handle post-listing shareholder logistics.; other
The industry, as management sees it
Management sees sustained supply-demand imbalance in GPUs for the next 3-5 years, with NVIDIA unable to meet demand and prices of GPUs/CPUs/RAM/storage rising structurally. International (especially US/Europe) AI labs willing to pay a premium and offer 12-18 months of advance; India demand still nascent. Older GPU SKUs (A100, H100) continue to see secondary demand from gaming, e-commerce and video-rendering customers after AI lab use.
Risks management named
- GPU supply constraint with delivery delays of 4-9 months
- Customer concentration risk on Sharon AI; lease expires after 7 years and renewal not guaranteed
- Quarterly seasonality — Q1 typically weak, Q3/Q4 historically stronger
- Quarter-on-quarter revenue volatility from lumpy design/managed-services work in subsidiaries
Q&A
13 analyst exchanges dominated by the Sharon AI deal mechanics — deployment timing, lease duration, rate-locking, GPU lifecycle and customer concentration. Management was confident and forthcoming on strategic/operational detail but uniformly declined to give revenue or PAT guidance. No adversarial pushback; tone was celebratory given the recent September 4, 2026 IPO listing. The few direct answers on order book size, customer advance quantum and capex targets carried the call.
Not answered directly
- FY27 and FY28 full-year revenue and profitability guidance
- Total order book quantum internationally beyond a 'much larger' qualitative descriptor
- Customer advance amortisation schedule and quarterly burn-down
- Total international GPU pipeline value in dollar terms
Asked for a number, answered without one
- FY27 and FY28 full-year revenue and profitability guidance: Management declined, saying the company has not planned to disclose forward-looking numbers and will report actuals each quarter.
- Per-GPU revenue / size of FY27-FY28 contribution from Sharon AI: Management pointed to industry pricing of $5-6 per GPU per hour and said 8,200 GPUs will all kick in at the same time; no deal-specific FY revenue given.
- Margin split for upcoming AI factory deals vs. Sharon AI pricing: Management gave only a 15-20% PAT margin target band, not deal-specific margin disclosure.
Every question, with its answer
1. Segmental revenue mix
Swechha Jain, ANS Wealth
Question. Could you explain the three verticals — IaaS, SaaS, Managed Services — in simple terms, the cross-sell dynamic and segment-wise margins? Also, where does own-GPU revenue get classified and what capacity do you plan to build?
Answer, Piyush Somani, Chairman and Managing Director. Explained IaaS (cloud/GPU/co-location infra), SaaS (Swaraj Cloud and dozens of products), and Managed Services (people managing customer cloud/AI factories). Stated company prefers to bundle all three so customer is 'fully dependent'. 'Our profitability is high on Software-as-a-Service. But now, with growth of AI factories, with growth of GPU business, we are seeing more numbers coming from Infrastructure-as-a-Service and better margins also coming from Infrastructure-as-a-Service.' Saurabh Daga added segment split: IaaS 51%, Managed Services 31.31%, SaaS 17.55%. GPU revenue booked under IaaS; currently 'majorly leasing from partners', no significant GPUs went live in Q1. Sized future build as 'sizeable' with AI factory ramps from Q3/Q4.
Follow-up. Inquiry on own GPU capacity plans and rationale for postponing the AGM (asked for pre-call presentations going forward).
Answer. Prasad Deokar attributed AGM postponement to spike in retail shareholders to ~1,90,000 post-IPO (listing 4 Sept 2026) and need to align timelines; ROC extension obtained.
2. Sharon AI deployment timeline
Aaryan Mehta, Shravas Capital
Question. GreenSquare was leased by Sharon AI on 21 Sept for 8,200 chips vs your 16 Sept target. What's the delay reason, when does capacity go live, and revenue begin? Also timeline on 600-800 own GPUs and further scale, and margin profile of these deals.
Answer, Piyush Somani, Chairman and Managing Director. Confirmed a 1-1.5 month delay; revenue from Sydney AI factory starts Q3 onwards. GreenSquare was building the data centre for 8 months; once billing starts, ~1 month more for rack installation/testing. Targeted 'early Q1 or mid of Q3'. Own GPU fleet targeted at ~1,500 by Jan-Feb (mid Q4). Working on multiple new deals; will inform regulators as closed. Margin profile: 'We are targeting PAT margins in the range of 15% to 20%. So, that's our target PAT margin. We will not work for anything lesser than that.'
Follow-up. Pipeline size and quantum of future deals relative to the 8,200 GPU Sharon contract.
Answer. Pipeline of multiple deals; specifics will be shared on closure via regulatory intimation.
3. India vs International GPU strategy
Anmol Garg, DAM Capital
Question. Will India focus shift to CPU or GPU additions? Sustained GPU demand from Indian clients or just pilot projects? Any domestic order book to highlight and FY27/FY28 guidance?
Answer, Piyush Somani, Chairman and Managing Director. Indian demand 'taking some more time to pick up' vs global. Plans to initially offer India-built capacity to international customers, then domestic as India demand picks. Indian GPU revenue 'has already started, but the numbers are not significant'. Domestic order book ~Rs 3,000 crore over 3-year duration; 'International is much larger than that'. Explicitly declined full-year guidance: 'we are not supposed to be giving any kind of guidance for full year.'
Follow-up. Order book duration and full-year guidance.
Answer. 3-year duration; declined numerical guidance.
Not answered directly.
4. Subsidiary revenue and customer advance
Divy Agrawal, Ficom Family Office
Question. SPOCHUB subsidiary booked Rs 85 cr revenue and Rs 54 cr PAT in FY26 vs zero in FY25 — nature of revenue, one-time? Status of Rs 1,100-1,200 cr advance and execution timeline. Also Managed Services up 157% YoY to ~Rs 76 cr — drivers?
Answer, Saurabh Daga, Accounts and Finance. FY26 SPOCHUB revenue was one-time technical design services for Sharon AI GPU cluster architecture, delivered via subsidiary employees. Advance pertains to same Sharon AI order, will go live Q3 FY27 and amortise monthly. FY26 Managed Services surge explained by same technical design work classified under Managed Services; future shift will be towards IaaS revenue as GPU factories go live.
5. Per-GPU pricing and capex structure
Sudhir Bheda, Bheda Family Office
Question. Sharon AI deal in plain terms — how many GPUs, per-GPU revenue, FY27/FY28 contribution, order book abroad. Pricing for new orders — same ~$5/hr?
Answer, Piyush Somani, Chairman and Managing Director. All 8,200 B300 GPUs in Sydney AI factory will kick in together post mid-Q3 deployment. Industry pricing 'upwards of $5 per GPU per hour, $5, $6 per hour — we are also going to bill our customer in that range only.' For India build, capital structure is 30-30-30 (advance from customer, lending, IPO proceeds). Holding off signing new customers at current capacity because GPU prices expected to rise into January; will lease out at higher rate.
Follow-up. Outlook on GPU pricing direction.
Answer. Asserted prices are 'continuously going up' for next 3-5 years due to demand-supply mismatch.
6. International GPU pipeline
Yash, Mavira
Question. Size of international pipeline (in GPUs), capacity to serve it, confirmation of 1,500 GPU deployment by Q4 FY27, and shape of domestic order book.
Answer, Piyush Somani, Chairman and Managing Director. Disclosed 'strong funnel of more than 50,000 GPUs right now' across geographies. Strategy is to block capacity before a 25% price rise and lease out after. Confirmed 1,500 GPU deployment by Q4 FY27. Domestic currently seeing 30-40% CAGR.
7. Capacity timeline and forward outlook
Kunal Bajaj, Choice Institutional Equities
Question. Was the QoQ revenue decline entirely Sharon AI technical billing or more structural? Can Sharon go live November first week? Nature of European banking client workload. IPO-funded capacity operational timeline; billing from Q4 or Q1 FY28? FY27/FY28 outlook?
Answer, Piyush Somani, Chairman and Managing Director. Expects Sharon live by November first week. Declined to comment on the end-customer's workload nature ('none of our purview'). IPO-funded capacity live by Q4 FY27; billing starts Q1 FY28. Declined forward-looking revenue/profitability numbers.
Not answered directly.
8. Geopolitical risk and GPU lifecycle
Zaksh Manekshana, Ambit PMS
Question. Client mix for international pipeline — US/Europe/China? Geopolitical risk of Chinese exposure. Will GPUs commoditise given B200/B300/Rubin launches? Demand-led or capacity-led approach to GPU procurement? International infrastructure plans?
Answer, Piyush Somani, Chairman and Managing Director. All international client onboarding requires NVIDIA approval. Asserts GPU is an 'appreciating asset' — H100 and older SKUs still in demand, CUDA framework keeps older GPUs relevant. Cites H100 spot price 'around double' of the B300 lease rate Sharon locked. Defended view that 8-year useful life is normal across AI lab → gaming → e-commerce migration. Initially conservative (customer first, then capacity); now building capacity ahead of RFS via auctions. Will build internationally in geographies offering green/clean power via local partners, plus India.
9. Sharon contract structure and GPU obsolescence
Vedant Sarda, Nirmal Bang Securities
Question. On Sharon AI — is revenue locked at a pre-decided rate while you hold GPU price upside? Won't GPUs become outdated after 5-6 years given latest tech?
Answer, Piyush Somani, Chairman and Managing Director. Confirmed Sharon revenue is locked at pre-decided rate ('Correct, Vedant'). On lifecycle: A100/H100 still command $3.5/hr — 'the price at which we locked the B300, that same price is being demanded by H100.' Pointed to 8-year useful life across customer segments and the CUDA-driven performance durability.
10. Revenue seasonality
Ravi Mehta, OneUp
Question. Disconnect between 7-8% YoY Q1 revenue growth and 30-40% order book growth — seasonality or otherwise?
Answer, Piyush Somani, Chairman and Managing Director. Acknowledged seasonality: 'Q4 is always a stronger quarter for us.' Expects seasonality to fade over next 2-3 years as AI factories deliver even monthly billing. Saurabh Daga added Q1 growth of 8% is appropriate given current market conditions for domestic customers and that YoY expansion will accelerate from Q3.
11. Sharon partnership economics and headcount
Sandeep Agarwal, Naredi Investments
Question. Why can't Sharon AI lease directly to end users — what is our value-add? Renewal risk post-5 years? Impact of rising spot rates on new deal pricing and margins? Any PLI/capex subsidy eligibility? Current employee strength and incremental hires?
Answer, Piyush Somani, Chairman and Managing Director. Sharon contract locked for 7 years (not 5); rates already discovered with them. Sharon has since closed 10+ AI factories and is in strong cash position. Rates will rise with supply-demand gap; ESDS manages OEM/ODM/NVIDIA/leasing-partner ecosystem. No PLI/subsidy eligible. Current headcount 993; ~100 additions for Sharon deal offset by churn — will not exceed 1,000. Model uses partner-staffing to keep people cost low.
12. Lease renewal and supply chain
Nupur Kogta, NIPL
Question. Risk of non-renewal after Sharon's 7-year lease; any similar risk on new deals? GPU availability as a constraint for India build?
Answer, Piyush Somani, Chairman and Managing Director. On 7-year expiry: 'we don't know what will happen' but will renegotiate at prevailing market rates; describes '1 plus 1 equals to 11' partnership. Confirmed GPU availability is already a constraint with delivery delays of 4-9 months, mitigated via partner relationships.
13. Core business and capex deployment
Aaryan Doshi, Divitiae Investments
Question. RHP says business is not seasonal — was the strong Q4 only because of ~Rs 75 cr from overseas customer? Core business growth in Q1? Deployment of Rs 1,177 cr customer advance + Rs 720 cr IPO into capex so far; total FY27 capex plan?
Answer, Piyush Somani, Chairman and Managing Director. Confirmed core business also grew YoY in FY26; the Q4 spike included one-time managed-services revenue for AI factory design/architecture. FY27 capex targeted at ~Rs 1,500 crore (additional Rs 1,500 crore was also referenced as a separate bucket) — funded by customer advance, IPO proceeds and partial debt. Saurabh Daga/Piyush could not give the exact deployed-to-date number on the call.
Partly answered.
What was said
Topic by topic, in the order it was spoken
Q1 FY27 Performance Highlights · Piyush Somani (CMD)
- Revenue from operations Rs 133.66 crores, +7.28% YoY
- PAT Rs 29.28 crores, +14% YoY
- EBITDA Rs 559 million, 42% margin
- Positioned the quarter as 'steady, disciplined growth' driven by customer demand and execution across all three segments
- Framed results as sustaining momentum following the September 4, 2026 IPO listing
Swaraj Sovereign Cloud and Product Launches · Piyush Somani (CMD)
- Swaraj Cloud branded as AI autonomous sovereign cloud platform integrating compute, storage, networking, databases, AI/GPU, cybersecurity and managed services
- Launched Swaraj Garuda — application performance monitoring and observability product
- Launched Swaraj Jatayoo — database activity monitoring product
- Stated goal of converting product investments into sustainable long-term value for customers and shareholders
Detailed P&L Walkthrough · Nadukuru Ramaiah (CFO)
- Total income Rs 1,351.71 mn vs Rs 1,257.21 mn YoY; total expenses Rs 987.04 mn vs Rs 910.39 mn
- PBT Rs 364.67 mn vs Rs 346.82 mn; tax expense Rs 71.90 mn vs Rs 89.96 mn
- Revenue +7.28% YoY but -20.20% QoQ — attributed to one-time design services revenue in subsidiary during Q4 FY26
- CFO characterised core business as growing steadily per MD commentary
In their words
We are targeting PAT margins in the range of 15% to 20%. So, that's our target PAT margin. We will not work for anything lesser than that.
It's not like an iPhone which becomes slow after 4 years or 5 years. It continues to become faster and faster because of CUDA.
There is desperation in international customers to get the GPU services. So, we are happy to provide services to them. They are good paymasters, giving 12 to 18 months of advances also.
To check next time
What management committed to on this call, or the dates they gave.
- Sharon AI Sydney 8,200-GPU deployment going live; first AI factory revenue recognition.; we are expecting that to go live by November first week
- Own 1,500-GPU India build progressing toward Jan-Feb 2027 (mid/end Q4 FY27) commissioning.; deploying those by January or February next year
- FY27 capex run-rate: remaining INR ~1,500 cr to be concluded.; we will be doing additional INR1,500 crores of capex in FY27
- Mix shift from Managed Services to IaaS as design revenue normalises.; the shift will be towards IaaS revenue
- New international AI factory contracts under the 50,000-GPU funnel.; strong funnel of more than 50,000 GPUs right now
- Customer advance of INR ~1,177 cr amortisation profile.; advance will get amortized over a period of contract on monthly basis
Transcript
We have not transcribed this call's recording. Read the company's transcript (PDF).
The stock after the call
| After the call | Close | Stock | Nifty 50 |
|---|---|---|---|
| Next session Fri 25 Sept 2026 | ₹1,760.50 | −5.00% | +0.34% |
| 5 sessions Thu 1 Oct 2026 | ₹1,434.10 | −22.61% | −2.78% |
From the close of Thu 24 Sept 2026, ₹1,853.15: the last close before the call, which began at 11:00 IST. Adjusted daily closes; the move includes everything else that happened in those sessions.