Loading...
Loading...
The science of why institutions fail —
and the architecture to fix them.
01 — The pattern
A community foundation raises $2M, distributes it, starts from zero next year. Every year. For decades.
Capital consumed on contact
A hospital needs a $500K scanner that Medicare would pay for. Can’t buy it. Five-person committee for anything over $1K.
Authority bottleneck
A nonprofit loses its founding director. Within 18 months, programs collapse. New director starts over.
Memory walks out the door
A government agency adopts AI for case processing. No governance framework. 67% of orgs are doing this right now.
Delegation without architecture
These are not bad organisations. They are good organisations that fail structurally — and nobody can explain why.
02 — The missing science
What exists today
The questions IRSA investigates
Medicine has diagnostics. Architecture has load calculations. IRSA is building the equivalent for institutions — a formal design science that prevents structural failure.
03 — What changes
Before
Capital consumed in a single pass
After
Capital recycles indefinitely
→ $100K → $567K cumulative impact at 85% recycling
Before
Knowledge leaves when the director leaves
After
Memory is structural, encoded in the institution
→ Leadership changes without capability loss
Before
Governance is a policy doc a new board can override
After
Governance is constitutional — enforceable, persistent
→ Purpose survives political and economic cycles
Before
Every blended finance deal structured from scratch
After
Standardised functions compose into any instrument
→ From bespoke to replicable — like banking did for lending
Before
AI adopted with zero governance
After
AI governed as delegation, not technology
→ Decision surface stays bounded as capability scales
04 — Already built
Nobody paid for any of this. The ask is partly retrospective — pay for value that’s already been created.
79
Working papers
50
Interactive tools
3
Production systems
$0
External funding
Recognition & traction
Note: Deployment money landed before operating runway. That’s backwards — but it’s also validation. Someone already bet on the output. Now we need to fund the engine that produces it.
05 — The engine
IRSA isn’t a think tank that writes reports. It’s an applied research institute that investigates structural problems, develops the theory, and then builds the infrastructure to fix them. Research in, ventures out.
Research questions → applied infrastructure
“How can philanthropic capital recycle instead of being consumed?”
→ Produced Elevate — now in pilot with $150K deployment capital
“How does governance survive leadership change and AI delegation?”
→ Produced Constellation — in production, multi-org
“How can blended finance structures be standardised?”
→ Produced UACC framework — practitioners applying to live deals
“What’s next?”
Fund the institute and find out. The pipeline is the point.
Google started as search, not Alphabet. You’re not funding one product. You’re funding the engine that keeps producing them.
What a team could do
People
Research
Products
Presence
06 — The ask
Two ways to support this work, depending on what resonates.
Basket 1
Fellowship
$200K / year
Buy the founder’s time. A fellowship at Ormond College (University of Melbourne) covering salary and the freedom to do this work full-time. Currently unfunded — everything built so far was done while looking for the next paycheque.
“I want to support the thought leadership and the person behind it.”
Basket 2
Research Institute
$500K–$1M / year
Fund the charitable engine. Research fellows, field studies, applied infrastructure, books, events, institutional partnerships. The institute that generates ventures like Elevate and Constellation — and whatever comes next.
“I want to support the mission and see what this institute produces.”
79
Papers, unfunded
3
Systems built
$0
Funding to date
This is what one person did with nothing.
Imagine what an institute could do.
Roshan Ghadamian
roshan@irsa.institute
irsa.institute · elevate.gift · constellation.institute
The theory exists. The infrastructure exists. The institute needs to exist too.