Recently I have been doing the research on the cloud computing space in the GPU era and I have found some astonishing facts. Some of them are as follows:
-
Global Cloud Computing Market: The global cloud computing market is projected to reach $2.5 trillion USD by 2032, up from $500 billion USD in 2022.
-
Top Cloud Service Providers: The three leading cloud service providers—AWS, Azure, and Google Cloud—are expected to hold a 36% market share in the $2.5 trillion cloud computing market (approximately $900 billion USD). This growth is driven by the ongoing AI revolution.
-
Indian Hyperscalers: Currently, there are no hyperscale cloud providers originating from India.
-
Yotta’s Initiatives: Yotta, a player in this space, has ordered 16,000 NVIDIA H100 chips, with a total acquisition and implementation cost of around $1 billion USD. The shipment was scheduled for receipt by June 2024. Yotta has launched its own cloud computing platform called Shakti Cloud, which has already sold out its capacity, with more companies on the waiting list.
-
Yotta’s Cloud Capacity: Yotta’s cloud capacity is sufficient to train two ChatGPT-4 models within three months.
-
Global NVIDIA Chip Shortage: Despite TCMC doubling its capacity in 2023-24, there remains a global shortage of NVIDIA chips. TCMC plans to double capacity again in 2024-25.
-
Challenges for Indian Data Centers: Indian data centers, even the latest ones, are not adequately equipped to handle the power and cooling requirements of next-generation GPUs like the H100 and Blackwell. Reengineering is essential to host cloud computing in Indian data centers.
-
Indian Public Cloud Services Market: In 2023, the revenue from Indian public cloud services (including IaaS, PaaS, and SaaS solutions) reached $8.3 billion USD, according to the International Data Corporation (IDC). The overall Indian public cloud services market is expected to grow to $24.2 billion USD by 2028, with a CAGR of 23.8% for 2023-28.
E2E Networks Concall Analysis
-
Capex and GPU Acquisition: In FY 2023-24, E2E Networks invested approximately ₹145 crore in fixed assets, primarily acquiring 450 NVIDIA H100 GPUs. The cost per GPU was approximately ₹32 lakhs.
-
Monthly Recurring Revenue (MRR): Following the recent capex, the company expects an MRR of ₹14-16 crore. For FY 2023-24, the MRR stood at ₹7.83 crore.
-
Incremental Revenue and Profit:
-
The increase in MRR amounts to ₹8.17 crore, resulting in an annual incremental revenue of around ₹98 crore from the current year’s capex.
-
The incremental PAT (Profit After Tax) is approximately ₹12.48 crore.
-
Assumptions: EBITDA margin at 51%, depreciation based on a 6-year straight-line method, interest cost at 9% for outstanding short-term and long-term borrowing, and a tax rate of 25%.
Note: There is a scope of improvement in EBITDA margin.
- Guidance for FY 2024-25:
- The management aims for an ideal capex of ₹800 crore in FY 2024-25.
- Assuming consistent parameters, the company could acquire 2500 H100 chips with this capex.
- Incremental revenue from the ₹800 crore capex, following the same MRR Increase/capex ratio, would be ₹540 crore.
- Incremental PAT would be approximately ₹79.52 crore.
- Assumptions: Source of funding – 50% debt: 50 % equity, EBITDA margin at 51%, depreciation based on a 6-year straight-line method, interest cost at 9% for outstanding short-term and long-term borrowing, and a tax rate of 25%.
Note: There is a scope of improvement in EBITDA margin.
- Total Peak Revenue:
- After the ₹800 crore capex, the total peak revenue is projected to reach ₹732 crore.
- The peak PAT is estimated at ₹111.34 crore.
Challenges/ Questions for which I am looking the answers for:
-
Sourcing NVIDIA H100 Chips: Will the company be able to source NVIDIA H100 chips?
-
Fundraising for Capex: Will the company be able to raise funds for the required capex?
-
Data Center Space: Will the company be able to obtain the required space from data centers for running the GPUs?
-
Selling Cloud Computing Space: Will the company be able to sell cloud computing space to customers?
-
Technical Readiness and Staffing: Is the company well-equipped technically and does it have the required number of employees to cater to such a large number of customers?
-
Product Strength and Demand Creation: Is the company’s product strong enough to create this kind of demand?
-
Competitive Strengths: What is the company’s right to win? What are its competitive strengths?
-
Software Stack Importance: What is the importance of the company’s software stack in the GPU era?
Feel free to share your findings, and I’ll be here to continue the discussion!