Satish Grampurohit

Last year, my net worth hit an all-time high. By another measure, I’d lost half my wealth.

A lot of you have asked me, “Where do you invest your own money? What does your portfolio actually look like?” So, I thought I’d finally answer it.

Over the last 20 years, I’ve made some of my best investment decisions, and some of my biggest mistakes.

In the last week of March 2020, with the world locked down and markets in freefall, I split half of my deployable portfolio into the safest asset I knew. The other half went into my highest-conviction investment.

If you’d like to know what those two investments were and how last 20 years shaped my portfolio, and where every rupee of my wealth is invested today, read the full blog below.

https://lnkd.in/gujsf9_5

Candid Health just raised $120M Series D and tripled its valuation in 18 months — while 42% of health systems are deploying AI but only 4% have achieved scaled, measurable outcomes. That gap is the whole story right now.

Here is what happened in healthcare AI this week:

→ Candid Health closed a $120M Series D led by Sixth Street Growth.
Valuation tripled since February 2025. 190% ARR growth YoY. 200+ provider orgs on platform.
Their model: treat RCM as a data engineering problem first — then layer in AI agents.
Result? Clients like Talkiatry cut manual billing work 40%. Nourish hit a 96.7% touchless claim rate.

→ HHS joined the White House’s $5B Genesis Mission on July 22.
NIH-led challenges targeting chronic disease root causes, pediatric cancer, and AI-accelerated drug discovery.

→ Luma Health’s Aditya Bansod put a name to the next frontier: “Headless AI.”
Not replacing EHRs. Not building another app. Connecting siloed systems so humans stop being the coordination layer.

→ Qventus data shows 94% of health system CIOs say AI delays create competitive disadvantage — yet 80% still cannot quantify ROI.

The funding is real. The urgency is real.
The implementation gap is realer.

Which of these three — RCM automation, federal AI research investment, or operationalizing AI past the pilot — has the most direct impact on your organization’s bottom line in 2026?

#HealthcareAI #Revenuecycle #DigitalHealth #HealthTech #AIAdoption

Grateful to moneycontrol.com for featuring the vision behind JAN AI 🙏

For me, JanAI has never been just another initiative. It is a people-first movement.
A movement that believes AI should not remain the privilege of a few, but become a force that unlocks the aspirations and potential of millions—especially across rural and underserved India.

This vision can only become reality through the coming together of Samaaj, Sarkar, and Bazaar—civil society, government, and markets—working together to unlock human potential at population scale, with AI as a powerful enabler.

We’re still at the beginning of this journey, and the opportunities for collaboration are endless.

If you have an idea, a partnership opportunity, or simply want to contribute to this movement, leave a comment below or reach out. We’d love to connect.

Come, join the movement! Together, let’s transform rural aspirations into AI-powered opportunities

Head Held High Foundation

Read the story here:
https://lnkd.in/gWxZUJUa

I did an interview with Tracxn recently, where I talked about how I stumbled into Investment Banking, what the early years actually looked like, how Dexter Capital Advisors has grown over the years, and what deal-making actually looks like once you are past the pitch.

Watch the full interview here: https://lnkd.in/g6XP9GEk

Devendra Agrawal, CFA Anuradha Aggrawal Rohit Pratap Singh Aiswariya G Shobita Dhar

#InvestmentBanking #PrivateEquity #VentureCapital #Leadership #Careers #DexterCapital

An excellent idea from Infosys, especially relevant for middle-aged gents like me – a company-wide fitness league where you have a different challenge every week. This week’s ask is the number of steps and here is my entry.
I hope you are doing yours as well Viral Thakkar and Alok Kumar !
#infyfitnessleague #hale #iwol

(And before anyone comments on the two wristwatches with 3 dials in total, the Titan watch is a Brand Equity Quiz prize from 2001. Those were the days! )

Most organisations make the build vs. buy vs. co-build decision at the point of budget approval and treat it as settled.

It rarely is.

The decision gets made on the wrong criteria. Available talent. Timeline pressure. A vendor relationship that already exists. None of these are the right starting point.

The right starting point is one question: what does this capability need to do that is unique to us?

If the answer is nothing, the case for building is weak. Standard customer service workflows, document extraction, basic fraud flagging. These problems have been solved. Buy the commodity, deploy it fast, and move on.

If the answer is something significant, building earns its place. A credit decisioning model trained on fifteen years of proprietary customer data. A risk scoring engine built around underwriting logic that reflects a market segment nobody else serves. Here the data is the moat. Owning the model means owning the advantage. The organisations that have built proprietary AI capability in credit risk and customer decisioning have created something a competitor cannot replicate by buying the same vendor.

The mistake most organisations make is applying build logic to commodity problems and buy logic to proprietary ones. They build the chatbot because the engineering team is available. They buy the credit decisioning engine because procurement is faster. The result is internal investment where it does not differentiate and external dependency where it does.

Co-build earns its place when the capability is genuinely proprietary but the infrastructure to deploy it is not. The organisation owns the data, the domain logic, and the decisioning framework. The partner owns the model architecture, the deployment pipeline, and the engineering depth to move at speed. Neither is substitutable. That is what separates a deliberate co-build from an outsourced build with unclear ownership.

The governance question most co-build arrangements fail to answer at contracting is straightforward: when the model produces a wrong outcome, who owns the response? If that question has no clear answer before deployment, the arrangement will answer it under pressure.

All three options are legitimate. The decision criteria are what most organisations have not made explicit.

Build where data and logic are proprietary and advantage compounds with ownership. Buy where the capability is commoditised and speed matters more than differentiation. Co-build where proprietary logic needs infrastructure that would take years to build alone.

Until that line is drawn explicitly, the sourcing decision will keep being made on availability and relationship. Not on strategy.

𝘚𝘩𝘢𝘳𝘦𝘥 𝘧𝘰𝘳 𝘪𝘯𝘧𝘰𝘳𝘮𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘱𝘶𝘳𝘱𝘰𝘴𝘦𝘴 𝘰𝘯𝘭𝘺 𝘢𝘯𝘥 𝘯𝘰𝘵 𝘳𝘦𝘱𝘳𝘦𝘴𝘦𝘯𝘵𝘢𝘵𝘪𝘷𝘦 𝘰𝘧 𝘢𝘯𝘺 𝘰𝘳𝘨𝘢𝘯𝘪𝘴𝘢𝘵𝘪𝘰𝘯 𝘐 𝘢𝘮 𝘰𝘳 𝘩𝘢𝘷𝘦 𝘣𝘦𝘦𝘯 𝘢𝘴𝘴𝘰𝘤𝘪𝘢𝘵𝘦𝘥 𝘸𝘪𝘵𝘩.

20 years, a reflection-

They did NOT go in a blink of an eye. There was a lot of toil, fights, ups and downs, loss of face, along with laughter, dance, smiles, and sorrows.

It was NOT about clients.
It was always about Teams – with them, about them, and towards them.

It was about Taking Risks and Putting Bets.
Some worked and some didn’t, but it was about learning from failures to push harder for something that worked.

It was Always about Mentors.
I got the space and nurturing when I needed it the most, as well as being allowed to fail if needed.

It was Never about a unit or a group.
It was about an organization, an ecosystem that comes together as a force.

It was Never about the work.
It was always about the culture and ways of working. As Bill Walsh said about the game: “The score will take care of itself.” I also believe that numbers will take care of themselves as long as you have set the culture right.

Finally, it is and it will never be about me.
It is and it will be about the leaders I worked with and learnt from, and the leaders I groomed. I am the sum total of all 😊

Infosys Consulting