Adding beds is the easy part. Making each bed work is the hard part.
India's hospital groups are building fast. In the most recent quarter they grew revenue by 21 per cent and added 15 per cent more working beds. Their profit margin did not move. It sat where it has sat for a while, at about 23 per cent. That gap between growing fast and earning more is what this edition is about.
- Revenue up 21 per cent, beds up 15 per cent, margins flat at 23 per cent. And ICRA expects another 14,500 beds to open across this year and next, costing around ₹30,000 crore.
- A new bed loses money before it makes any. It takes years to get from opening day to earning the group's average margin, and every group adding beds is carrying a lot of beds that are still on that journey.
- The popular fix is to fill the beds. I tested that and it does not hold up — the link between how full a hospital is and how profitable it is points one way in FY25 and the opposite way in FY26.
- What does show up clearly is shorter stays and higher prices. Cutting the average stay from 4.1 to 3.9 days treats about 5 per cent more patients using the same beds.
- If you sell services to hospitals, the flat margin is the addressable problem, and the two levers that move it without capital are stay length and case mix. Section seven.
- KKR is paying $1.4 billion for Medicover India. It looks like a turnaround bet. It is mostly a debt story, and 19 of its 25 hospitals already make money.
Growing revenue while margins stay put is not a failure. It is what happens when you keep adding capacity: the new beds drag the average down while the old ones hold it up. But it does mean the interesting question is no longer how many beds a group owns. It is how fast each one starts paying for itself.
02 — The reason
A new bed loses money before it makes any
This is the part that gets skipped when a hospital group announces an expansion. A bed does not switch on and start earning. It goes through stages, and most of them are loss-making.
Stages one to three are the expensive ones. You are paying the doctors, the nurses, the electricity and the loan, and the patients have not arrived in enough numbers yet. A hospital group in the middle of a building programme is carrying a lot of beds sitting in that zone.
Which explains the flat margin without needing anything to have gone wrong. Scale arrives in months. Profitability arrives in years.
You can see it in the most recent numbers. Manipal grew revenue 38 per cent, but its revenue per bed fell once the newly bought Sahyadri hospitals joined the average: ₹77,200 without them, ₹71,500 with them. Nothing got worse. Cheaper beds simply entered the calculation.
03 — The popular answer
So just fill the beds?
That is the standard prescription, and it is where most investment cases start. Indian hospitals run at roughly 63 per cent occupancy, so a third of the beds are empty on an average day. Fill them and the profit follows, or so the argument goes.
I tested it. I took the listed hospital groups and checked whether the fuller ones were also the more profitable ones.
Same companies. Opposite answer, one year apart. When the result flips depending on which year you pick, you have not found anything. There are only six companies with comparable numbers, and hospital groups report margins on slightly different definitions, so the test is not capable of settling it either way.
What that does and does not mean. Nobody sensible thinks empty beds are free. Occupancy clearly matters to a hospital's economics. What the numbers will not support is the simpler claim that occupancy is the lever — that raising it a few points reliably widens the margin. If that were true across the sector, it would be visible in the sector's own reporting, and it is not.
04 — What does work
Two things that show up clearly
One. Getting patients home sooner
A bed that is occupied for four days earns four days of revenue. A bed that treats the same patient in three and a half days is free sooner for the next one. Shortening the average stay from 4.1 days to 3.9 means roughly 5 per cent more patients through the same beds, with no construction and no new hiring.
And here is the awkward part. Doing this well makes your occupancy look worse. Beds sit empty for longer between patients. So a hospital group that is genuinely improving can report softer occupancy, and a group that is simply keeping people in longer can report better occupancy. Occupancy on its own cannot tell those two apart.
Two. Charging more per day
The spread between hospital groups on revenue per occupied bed per day is close to two to one.
View the data behind this
| Hospital group | Per bed per day |
|---|---|
| Max | Rs 77,800 |
| Manipal | Rs 70,780 |
| Fortis | Rs 68,770 |
| Medanta | Rs 66,500 |
| Apollo | Rs 63,310 |
| KIMS | Rs 44,640 |
Source: CRISIL peer comparison, as reproduced in listed-company disclosure. Manipal is H1 FY26. CRISIL notes that bed and revenue definitions vary between companies, so treat these as close comparisons rather than exact ones.
Max earns ₹77,800 a day per occupied bed. KIMS earns ₹44,640. That is not mainly about how well the bed is run. It is about which city it sits in and which patients can pay. Max's beds are concentrated in Delhi and Mumbai, where incomes and insurance cover are highest.
Which is worth sitting with, because it cuts against the deal logic. The price a hospital can charge comes largely with the building and the city. You buy it. You cannot create much of it afterwards. So the lever that shows the clearest link to returns is also the one an acquirer has least control over.
05 — The live test
The $1.4 billion deal, and what it actually is
On 6 August 2026, KKR agreed to buy Medicover's Indian hospitals for about $1.4 billion. Twenty-five hospitals across Telangana, Andhra Pradesh, Maharashtra and Karnataka.
Reported that way, it sounds like a bet on fixing an underperforming chain. Medicover's own executive told Reuters something more ordinary.
Look at the last card. The company said its debt was rising and becoming hard to sustain, and that is what sent it to private equity. This was a balance sheet problem before it was a growth story.
And 19 of the 25 hospitals already make money. So the buyer is not rescuing a failing business. The job is clearing the debt and getting the six immature hospitals through the stages in section two.
The utilisation gap is real, though. Getting from roughly 2,700 occupied beds to the 4,000 management wants means about 1,300 more patients in beds every day, close to half again what the group carries now, without building anything new.
Which makes this a clean test of the whole argument. If filling beds is the lever, this is where we will see it. KKR already owns Max Healthcare, the group with the highest revenue per bed in the country, so if any buyer has earned the benefit of the doubt on turning beds into money, it is this one.
06 — So what
Five numbers nobody publishes
Every hospital group in India collects all five of these. Almost none of them publishes any.
Occupancy hospital by hospital, not group average
An average of 60 per cent might be a set of hospitals at 85 and a set at 35. Those need completely different fixes, and the average hides which one you own.
Average length of stay, by department
Without it you cannot tell whether falling occupancy means the hospital is emptier or faster. Those are opposite conclusions from the same number.
Revenue per bed excluding acquisitions
Otherwise a falling figure is unreadable. Buying cheaper hospitals and running worse hospitals look identical in the group number.
Beds licensed, beds open, beds occupied
All three, because the gaps between them are large. A licensed bed is a property statistic. An occupied bed is the business.
Who is paying, with scheme patients shown separately
Ayushman Bharat patients pay less per day than private ones. A group can fill more beds and earn less per bed in the same quarter, and be doing something either very good or very bad. The group numbers will not tell you which.
This is the pattern I keep running into. Bed counts appear in every announcement because they are easy to publish and flattering. The numbers that would show whether those beds are working are collected by everyone and disclosed by nobody.
Which means that in three years, when someone asks whether $1.4 billion bought better hospitals or simply more of them, nobody outside the deal room will be able to answer.
07 — For services and technology leaders
Where the commercial opportunity actually sits
If you sell to hospital groups, the flat margin in section one is your addressable problem. Every finding above translates into something a provider can be paid to fix.
The framing that lands with a hospital CFO. Do not open with digital transformation. Open with the number in section one: revenue up 21 per cent, margin flat. Then offer the two things that move margin without capital, which are stay length and case mix, and the one thing that makes both measurable, which is a common data definition across the estate.
And be honest about what technology cannot fix. A bed in a city where patients cannot pay ₹70,000 a day will never earn what a Delhi bed earns. Price comes with the postcode. Throughput and mix are the levers that are genuinely available to an operator, which is precisely why they are where the services money is.
08 — Takeaways
Key takeaways
09 — Sources
Sources and evidence status
What is officially reported, what I calculated from it, and what is an estimate.
Before citing externally: the occupancy test uses six companies over two years and cannot establish cause either way. CRISIL cautions that peer definitions vary, so cross-company comparisons are approximate. Medicover's figures come from company statements and press reporting rather than audited accounts.