From the article that @satishwe posted, it looks like they are planning to locate some of their servers in India - at least enough to serve Indian users of ChatGPT. These are likely to be inference servers only.
It remains to be seen whether they will outsource this to someone (from the list of companies mentioned) or operate it themselves. If they are planning to operate it themselves, it could be that they just buy a company outright or buy some assets.
No chance ChatGPT would want to operate Data Centers. Their main focus will be buildin and refining their models. It will definitely go downstream to these AI cloud companies
OpenAI will 100% eventually operate all their data centers as it is their biggest expense. They have already committed $500B to build out data centers (Stargate) and buying stakes in small nuclear power companies.
When you are spending billions on training and inference costs every year, the savings are huge when use your own infrastructure instead of hyperscalers. It’s only that OpenAI is growing so fast that they can’t build out their own infrastructure fast enough and need to use hyperscalers to meet some of the demand.
if that was the case why is Open AI using Oracle OpenAI and Oracle reportedly ink historic cloud computing deal | TechCrunch
That’s because they aren’t able to build up capacity themselves fast enough to keep up with growth. But eventually, their objective is clearly to have their own infrastructure. It will not happen overnight - probably will take a decade or something like that.
For example: see their flagship Texas data center project - OpenAI's gargantuan data center is even bigger than Elon Musk's xAI Colossus — world's largest 300 MW AI data center could reach record 1 gigawatt scale by next year, threatens grid stability | Tom's Hardware
Approved fundraising
https://nsearchives.nseindia.com/corporate/E2E_29092025110051_OutcomeofBoardmeeting.pdf
Authorized share capital doubling from 25Cr to 50Cr. 1000Cr fund raise thro various ways.
Doubling of authorized capital.
Is this like diluting the existing capital ? Can we expect valuation to come down by 50%?
No, it depends on what price they will be raising. Somewhere around 2600-2800 should be the price based on 90 days weighted average volume method.
Objects of fund raising and how the funds will be used to drive the EPS in future,
Usually whenever a company does a QIP the stock price falls
Raising at 2800 would not be optimal. Expecially when they did the LnT at similar/higher valuations. Given many of their GPUs are booked now, they should be able to handle debt and wait for a higher price for fresh equity when necessary
would they expanding their capacity with the cash they have and loan together making almost 2500cr?
It is just an enabling provision.
L&T has a habit of getting an enabling provision passed every year. The provisions allows money to be raised anytime over the next 1 year. It does not mean they will 100% raise money
After taking over 21% of E2E, L&T is implementing some of their practices.
Nothing wrong in the company creating a war chest for future growth
Almost 100% guaranteed. The revently have increased their borrowing limit to 1000 crore.
E2E is contract bound to publish quarterly results by 21st next month based on the L&T contract. It’s 29th today, Q2 should have been published a week ago. anyone knows what’s up?
are we going to see the proposed capex on these products as well?
That is why you never want to deploy all at once. The industry is changing so quickly, but older products will still have their use case
interesting to see the company’s plans and comments about the cash deployment and future growth it foresees.
Podcast of the founders on the Blume Ventures Podcast
