1. Power is the product
Every serious conversation about compute ends at a substation. Capacity that can energize on a credible date is worth more than capacity that is merely announced. I treat the interconnection queue as the real roadmap.
Vikhyat Prakash
Intelligence is becoming an industry.
Industries are built out of concrete, copper and power.
I build the physical layer of AI: data centers, energy and the capital structures that carry them. This is my working view of how that gets done, what I have learned doing it, and where I am headed next.
Vikhyat (विख्यात) is Sanskrit for renowned. The name is a standard to earn by building, not a claim.
- Focus
- AI infrastructure and mega projects
- Method
- Systems integration across land, power, compute and capital
- Long arc
- Regenerative healthcare and immersive technology
The thesis
The constraint on AI is no longer ideas. It is delivery.
2. Speed is a design choice
Greenfield campuses take years. Existing, powered shells can move in months. The winning portfolio mixes both, and is honest about which is which.
3. Integration beats invention
Most mega projects fail between the parts, not inside them. Land, utility, equipment, financing and customers each work alone. The value is in the seams.
4. Capital follows certainty
Investors do not fund ambition. They fund a sequence of de-risked steps with clear gates. A project is a financing story that happens to involve buildings.
The build
How I think about a mega project, layer by layer.
Each layer has its own clock, its own risk and its own experts. A project moves at the speed of its slowest layer, so I work them in parallel and keep one map of all of them.
- 01Months to years
Land and site
Location is power, fiber, water, zoning and community, in that order of how often they kill a deal.
- 02Longest clock
Energy
Grid, on site generation and storage as one problem. Reliability and carbon intensity are both requirements, not trade offs.
- 03Supply bound
Facility and equipment
Cooling, electrical gear and long lead items. Procurement is strategy because the supply chain sets the schedule.
- 04Revenue gate
Compute and customers
Demand that signs before steel goes up. Contracts shape the design, not the other way round.
- 05Enables all
Capital and structure
Funds, special purpose vehicles and project finance sized to each stage, with clean governance between partners.
- 06Compounds
Operations and data
The system of record that ties it together. If you cannot see every site on one screen, you are not running a portfolio.
Lessons from the work
What building in AI and systems integration has taught me.
Name the real bottleneck.
It is almost never where the team is looking. Ask what would stop this project if everything else went perfectly.
Sell the sequence, not the vision.
A staged plan with exit points raises more money, and better money, than a bigger story.
Partners are architecture.
Who owns what, who gets paid first and who decides are load bearing. Draw them like you draw the one line diagram.
Instrument everything early.
A single source of truth for sites, vendors and milestones is cheaper in week one than in month eighteen.
Use AI on the work itself.
Research, diligence, site screening and reporting are where agents earn their keep today. The physical build stays physical.
Be precise about what is known.
Separate confirmed, estimated and assumed. Trust is the scarcest asset in a market this fast.
The industry
Where I think this goes next.
Now
A scramble for powered capacity. Retrofit and fast to energize sites command a premium.
Next
Power and compute are planned together. Campuses are designed around generation, not attached to it.
Then
Infrastructure becomes a financial asset class with standard instruments, data and benchmarks.
After
Compute is a utility. The question shifts from where it lives to what it is for.
Beyond infrastructure
Why any of this matters: people.
Compute is a means. The ends I care about are longer, healthier lives and better ways to learn, heal and connect.
Regenerative healthcare
Medicine is moving from managing disease to restoring function. That shift needs enormous computation for biology, imaging and modeling, and the infrastructure I build is how it gets paid for.
AR and VR
Immersive tools change how surgeons train, patients recover and teams see complex systems. I follow and build here because spatial interfaces are the natural window into large, physical projects.
Other projects
Agents for diligence and operations, data products for markets that lack visibility, and tools that make hard industries legible. Notes appear here as they mature.
Writing
Essays on what is being built.
01 The modern Turing test: can an agent make the sale?Why the Machine Payments Protocol points to the future of payments, with three case studies. 02 The past, present and future of the digital twin dreamThe magic of effortless AR is one step closer. The history of headsets and digital twins, why AI was the missing piece, and where it goes next.Contact
If you are building something physical and ambitious, write to me.
Send a note on LinkedIn. Partners, operators, researchers and founders are all welcome.
Connect on LinkedIn