The system that processes your taxes was written before the moon landing. The system that handles Medicare for 67 million people is from the 1970s. 95% of every ATM transaction you've ever made was processed by a language called COBOL — invented in 1959, taught at zero universities, maintained by engineers who are retiring at 10% a year. $3 trillion moves through this infrastructure every single day.
The federal government spends $83 billion a year — 79% of its entire IT budget — just keeping these systems running. Add state and local government, healthcare, and banking and the number climbs into the hundreds of billions. All of it spent maintaining technology that cannot support real-time processing, cannot connect to modern systems, cannot be secured against modern threats, and cannot deliver any of the operational gains that AI promises. Within a decade, the engineers who understand this infrastructure will not be scarce. They will be gone. And there is no plan to replace them — or the systems they maintain.
Modernization is not optional. No company operating today is capable of full-scale institutional modernization across government, healthcare, and banking. Luminary was built to be that company — and we are aggressively moving to lead it.
If you used an ATM this morning, COBOL processed it. If you filed your taxes, COBOL handled it — the IRS Individual Master File has been running on COBOL since the 1960s and has never been replaced. If you're on Medicare, your claims run through a COBOL system from the 1970s. If you collect Social Security, your benefits are calculated by 60 million lines of COBOL managing over a trillion dollars a year. If you flew anywhere recently, your reservation was processed on a COBOL mainframe. If your bank sent you a statement, the core ledger behind it is almost certainly COBOL.
This isn't a few legacy systems lingering in a few agencies. This is the operating layer of American life. And it's written in a language that was invented in 1959, that no university teaches, maintained by developers who are retiring at 10% a year, running on hardware that is no longer manufactured. Nobody is paying attention to this. It doesn't make the news until something breaks — like when unemployment systems in six states crashed during COVID because they couldn't handle the load and the governors had to publicly beg retired COBOL programmers to come back.
| Sector | What Runs on COBOL | Scale |
|---|---|---|
| Your Tax Return | IRS Individual Master File — every American taxpayer processed on code from the 1960s. Still COBOL and assembly. | Every U.S. taxpayer |
| Your Social Security | SSA retirement, disability, survivors benefits — 60 million lines of COBOL implementing thousands of laws | $1T+/yr in payments |
| Your Doctor Visit | Medicare claims processing, Medicaid eligibility, hospital billing back-ends, insurer policy admin — all COBOL-era | 67M Medicare patients |
| Your Bank Account | Core deposits, lending, 95% of ATM transactions, 80% of in-person transactions, card rails, payment clearing | $3T daily transactions |
| Your Unemployment | State UI benefits processing — crashed in NJ, CT, KS, CA, RI, IA during COVID. Only 22 of 50 states had modernized by 2020 | 28+ states still legacy |
| Your Flight | Airline reservation, ticketing, loyalty, scheduling — outages regularly make headlines when the mainframes fail | Global reservation systems |
| Your Insurance | P&C and life insurer core policy and claims systems, state workers' comp — decades old, coupled to actuarial models | Every major carrier |
| Your Employer | Fortune 500 payroll, HR, inventory, supply chain on IBM Z mainframes — 70%+ still running COBOL workloads | 70%+ Fortune 500 |
There is no master list of COBOL dependency by institution. No centralized registry. No government audit that covers the full scope. The data has to be assembled from GAO reports, state audits, vendor disclosures, and pandemic-era emergency appeals. Which means the real number is almost certainly larger than what's above — because nobody has finished counting.
The average COBOL developer is in their mid-50s. About 10% of the remaining workforce retires every year. 60% of organizations that still depend on COBOL say they cannot find replacements. Nobody is training new ones. And here's the part that should keep people up at night: in most of these systems, there is no documentation. The code is the documentation. The business rules — how taxes get calculated, how benefits get distributed, how claims get adjudicated — live inside variable names and conditional statements written by engineers who retired a decade ago. When they leave, that knowledge leaves with them. 42% of critical business logic is at risk the moment a key developer walks out the door.
That's the first cliff. The second is AI. Every governor, every bank CEO, every hospital CIO wants the benefits of AI. But you can't run machine learning on a mainframe. You can't plug a 1970s COBOL system into a modern API. You can't do real-time analytics on a batch-processing architecture designed before the internet existed. So before any of these institutions can use AI, they have to rebuild the foundation. The modernization wasn't just overdue. It's now the prerequisite for every other initiative on every other roadmap.
This isn't about preference. These aren't systems that could be better — they are architecturally incapable of doing what every institution now needs them to do. They were designed for batch processing in a world before the internet. They store data in formats — EBCDIC, packed decimals, VSAM flat files — that nothing modern can read natively. They can't be patched quickly when a security vulnerability is found. They can't scale horizontally when demand spikes. They can't expose data to other systems without layers of middleware bolted on after the fact. They are, by design, closed, rigid, and isolated. And every operational goal that matters in 2026 — real-time processing, interoperability, AI, cybersecurity, citizen-facing digital services — requires the opposite.
These two columns are not a comparison. They're a wall. You cannot get from the left to the right by patching, wrapping, or optimizing. The architecture doesn't bend that way. The only path is migration — pulling business logic out of decades-old procedural code, redesigning it for modern architecture, and rebuilding the foundation. That's what modernization actually means. And it's what every institution on the COBOL map is eventually going to have to do.
The numbers in this presentation are large. Billions in spending. Trillions in transactions. But behind every legacy system is a person waiting — for a tax refund, for a claim to be approved, for a permit to come through, for an account to work. These aren't abstract inefficiencies. They're real failures that hurt real people every day. And they're all caused by the same thing: infrastructure that was built before the internet and has never been replaced.
This is what modernization is actually about. Not upgrading servers. Not migrating to cloud. Fixing the systems that touch every American's life — their healthcare, their taxes, their bank account, their interactions with local government — and making them work the way they should have been working for the last twenty years.
Every state in America runs its own systems for taxation, licensing, benefits, and regulatory compliance. Most of it is decades old. Most of it can't connect to modern platforms. Almost none of it has been replaced. And there is no coordinated national effort to fix it. The modernization landscape is disjointed — a patchwork of isolated agency upgrades, incomplete migrations, and state-level initiatives that don't talk to each other. Most states haven't started. Most don't fully understand the scope of what needs to happen. The problem is misunderstood, uncovered, and massively underestimated. But it has to get done. Every one of these systems will eventually fail or be forced into replacement — and we have an opportunity to lead that process before the emergency arrives.
The model we're building in Indiana is designed to travel. A tax-system migration that works here works in Ohio. An MMIS modernization that passes CMS certification in Indiana is reusable in 49 other states. The platform components, the toolchains, the reference architectures — they're templates, not custom builds. Three states and two healthcare systems and this is a completely different company. Indiana is where we prove the playbook. What happens after that is arithmetic.
On April 28, 2026, Governor Mike Braun announced IN AI — a statewide initiative to make Indiana "the most AI-ready state in the nation." That's significant. It means the political will is there, the focus is declared, and the conversation about modernization is happening at the highest level. Indiana's agencies need the infrastructure rebuilt to deliver on that vision — and Luminary is the Indiana-based AI infrastructure company that's qualified to do it. The focus is aligned. The funding path exists. And we're already in motion.
It is that the country has built decades of operations around them. Government, healthcare, banking, education, and public services do not run on software alone. They run on laws, data, workflows, approvals, reporting requirements, employee habits, vendor contracts, and thousands of small exceptions that accumulated over time. The code is only one layer of the machine.
Modernization fails when that machine is treated like a software replacement project. Luminary's work starts earlier. We study how the institution actually operates. We map the systems, the data, the rules, the dependencies, and the human workarounds. We separate what is outdated from what is essential. Then we design the path from the old operating environment to a modern one.
That is the part most people underestimate. It is also the part that determines whether modernization creates real efficiency or just another expensive layer on top of the last one. Luminary is built for that layer: the space between legacy infrastructure and modern institutional operations.
Indiana is where we start. Two major reformation builds are in motion. The contracts are being pursued aggressively. Here's the gameplan.
We've already developed proposals for two entry points: a $100 million economic infrastructure reformation through the IEDC, and a unified Indiana education operating system through the Department of Education. We prove these, and everything else unlocks — state agencies, municipalities, banking institutions, healthcare systems. The detailed roadmap lives in our formal proposals. The principle is simple: land the flagship builds, deliver, and let the results open every door that follows.
Indiana is our home court. It's where we prove the model — then we take it to every state capital in the country.
Indiana spends 60–80% of its IT budget maintaining systems that are decades old. That's the pattern everywhere — agencies spending the majority of their technology dollars just keeping old systems alive, with almost nothing left for improvement. Modernization doesn't add cost. It redirects it. Pennsylvania's modest modernization program generated $37 million in realized savings in its first phase. Massachusetts committed $1.23 billion over five years because the math showed the cost of doing nothing was higher. Indiana has $52.2 billion in total annual revenue and a $22–23 billion general fund. The fiscal capacity is there. The precedent is set. And our $100 million IEDC proposal represents a fraction of what a serious five-year program would look like.
These are real IT budgets from real states — and in every case, the majority of the money goes to maintaining legacy systems, not replacing them. They're all looking for a modernization partner. That's what Luminary is.
| State | IT Budget | What They're Spending On |
|---|---|---|
| California | $13B | $198M for Employment Development modernization alone. Still running legacy systems across most agencies. Largest state IT budget in the country. |
| Texas | $5.4B | $1.3B to Dept. of Information Resources. $106.6M for Medicaid MMIS modernization. $55.7M for eligibility system redesign. Spending hundreds of millions and still early. |
| New York | $5.4B | $391M for centralized tech services. Building "One ID" for residents across all agencies. Whole-of-state cybersecurity push. |
| Massachusetts | $1.23B committed | FutureTech Act — signed into law. Five-year bond for IT modernization, cybersecurity, AI enablement + $400M federal match. The benchmark. |
| Pennsylvania | $37M saved | Cloud migration and procurement optimization generated $37M in realized savings — modest program, fast payback, new appropriations following. |
| Indiana — IEDC | $100M proposal | Luminary's starting reformation engagement. Opening wedge into a $1.0–1.5B five-year modernization envelope. |
| Indiana — DOE | Separate build | Parallel reformation for a $9B+ annual education system. Independent scope and timeline. |
The pattern across every state that's started: modernization spending generates measurable ROI within the first phase. The cost of maintaining legacy systems compounds every year. The cost of replacing them decreases as AI tooling improves. The longer you wait, the worse the economics get.
Every row you saw at the top of this presentation — taxes, benefits, banking, healthcare, unemployment, insurance, airlines, enterprise — represents institutions that are going to spend millions or billions replacing infrastructure that is actively failing. Not because someone decided it's a strategic priority. Because the developers are gone, the hardware is end-of-life, and none of it can connect to anything built in the last 20 years. The question isn't whether the money gets spent. It's who gets the contracts.
Healthcare is the clearest example of what's coming — and the largest single market in the modernization wave.
Medicare processes claims for 67 million Americans on COBOL from the 1970s. CMS just started converting its "Pricer" applications from COBOL to Java — in 2026. Every state runs its own legacy Medicaid system, and those systems depend on SSA's COBOL databases for identity and eligibility. The healthcare IT market is $354 billion today, headed to $981 billion by 2032. 60–70% of that goes to replacing legacy systems. CMS alone spends $6 billion a year on state Medicaid infrastructure and covers 90% of modernization costs.
All figures below are derived from peer-state benchmarks, McKinsey banking modernization data, Federal Reserve payments briefings, and our own Indiana cost analysis.
Every number below is derived from the contract sizes, peer-state benchmarks, and sector data already presented in this document. The savings estimates use documented figures — GAO maintenance costs, McKinsey banking ROI data, CMS spending levels, and Pennsylvania/Massachusetts realized results. The revenue scenarios model what happens as Luminary moves from Indiana to national scale.
These are not Luminary's revenues. These are the savings that accrue to the institutions we modernize — the ROI that justifies the contracts. Modern platforms reduce total cost of ownership by 20–40% and free budget that was trapped in maintenance.
| Sector | Current Maintenance Burden | Est. Savings Post-Modernization | Source / Basis |
|---|---|---|---|
| Federal IT | $100B/year across agencies (GAO) | $20–40B/year | 20–40% TCO reduction documented across modernization programs |
| State & Local IT | $96–128B/year (60–80% of $160B) | $19–51B/year | 20–40% reduction. PA realized $37M from modest program alone |
| Indiana Specifically | ~$600M–1.2B/year (est. 60–80% of state IT) | $120–480M/year | Proportional to state revenue. 0.4–0.6% annual reallocation sustains program |
| Healthcare / Medicaid | $212–248B/year (60–70% of $354B market) | $42–99B/year | 20–40% reduction. CMS $6B/year in directory fragmentation alone |
| Banking (U.S.) | $20–30B+/year est. (Canada's Big Five alone: $8B CAD) | $6–12B/year | McKinsey: 38% cost reduction post-modernization, 62% faster time-to-market |
| Per Community Bank | True cost 3.4× perceived (European case: €6.8M vs €2M perceived) | 38% reduction | Documented European mid-size bank case study — 18 months to realize |
Total estimated annual savings across all U.S. sectors if fully modernized: $87–202 billion per year. This is the value that justifies the spending. Luminary's contracts are a fraction of the savings they generate for the institutions we serve.
Three scenarios based on aggressive contract acquisition at each level of government, department, and sector. Contract sizes use the ranges established in this document. Each scenario assumes Indiana as proof of concept in Year 1, then scaling.
| Conservative | Moderate | Aggressive | |
|---|---|---|---|
| Year 1 | $45–55M | $55–70M | $65–85M |
| IEDC Phase 1 ($30–35M) + Dept. of Education Phase 1 ($10–15M) + municipal pilots ($2–5M) + first banking engagement ($3–5M) + Chamber + university network | |||
| Year 2 | $65–85M | $85–120M | $120–160M |
| Indiana continuing + 1–2 new state entries ($15–25M each) + Medicaid MMIS pilot ($5–10M) + 2–4 banking ($8–15M) + 3–5 municipal ($5–10M) + Schweikert/federal entry | |||
| Year 3 | $90–130M | $150–220M | $220–310M |
| 3–5 states active + 2–4 Medicaid systems ($20–40M) + 4–8 banking ($15–30M) + 5–10 municipal ($10–20M) + federal contracts expanding + managed services from completed Y1 projects | |||
| Year 4 | $140–190M | $250–340M | $380–480M |
| 5–8 states + 4–8 Medicaid + 6–12 banking + 8–15 municipal + federal expanded + platform licensing + managed services portfolio growing | |||
| Year 5 | $190–260M | $380–480M | $550–700M |
| 10+ states + national Medicaid presence + 8–15 banking + 15–25 municipal + federal portfolio + platform licensing + recurring managed services from all completed engagements | |||
The contract sizes driving these numbers are already documented in this presentation — $50–100M per state flagship, $15–40M per agency, $20–80M per Medicaid system, $10–75M per banking institution, $5–15M per municipality. The revenue grows because the model is repeatable: what works in Indiana works in Ohio, and the toolchains, reference architectures, and platform components we build in Year 1 reduce delivery cost in every year that follows. The managed services revenue from completed engagements compounds — every finished project becomes a recurring revenue stream.
Luminary is a holding company. Each platform below is a domain-specific AI operating system — not a tool, not a dashboard. They share infrastructure, they share intelligence, and each one opens doors the others walk through. Legal Insights gets us into the Chamber of Commerce and the Schweikert Medicaid engagement. Vestiq proves we deploy AI with real money. Blue Haven proves our systems run in production. Studios is how we deliver faster than anyone expects.
$3 trillion in daily transactions on code nobody teaches anymore. $100 billion a year in maintenance for systems that can't talk to anything modern. 50 states. 50 Medicaid systems. 50 unemployment platforms. All running on the same dying stack. And one company with a plan to lead the modernization of America's critical infrastructure.