Leverage Points
Beyond Tanks and Taxes: What’s Missing from the Czech 2025 Election Programs
Czech original version here.
In the run-up to the Czech elections, politicians argue over whether retirement should come a year earlier or later, how many tanks to buy, and who should get subsidies. But the world is moving faster: drone wars, AI, robots, new medicines, expensive electricity. If we don’t take this seriously, the country will end up a museum piece.
Here are eight leverage points—changes that could shift the whole system, and decide our security, health, and future.
1. Defense
The debate stops at “raise defense spending or not.” What we don’t hear is where the money should actually go. Meanwhile, Poland is shooting down drones with F-16s and missiles costing tens of millions each—while ten drones still slipped through.
The lesson from Ukraine is crystal clear: most new spending should go to producing drones, weapons to take drones down, AI to coordinate them, systems to defend against them, and radio systems to jam them. Plus: planning where drone wars will head in their second, third, and fifth iterations.
Want to buy 20 tanks for a few billion, like always? They’ll be destroyed by 20 drones worth one million. The drone ecosystem links naturally to startups, innovation, the Czech arms industry, data, and AI. We’re well positioned—what’s needed isn’t pocket change for “innovation,” but channeling new defense spending into this future.
2. Energy
a) Building giant nuclear blocks in Dukovany and Temelín will take 15–20 years and hundreds of billions. Don’t do it. The fastest path to cheap nuclear capacity is small modular reactors (SMRs)—faster to build, modular, and spread across multiple sites. Czechia already has know-how and infrastructure. ÚJV Řež is working with Rolls-Royce, GE, and NuScale. This is where to bet, accelerate, and support.
b) Why is electricity so expensive here when we have cheap nuclear and coal? Because EU power prices are set by the most expensive marginal source—often gas—and that price then applies to everyone. Low-cost producers pocket windfall profits, which the state partly claws back with special taxes. In recent years, that’s funneled 50–100 billion CZK annually into the state budget (windfall tax, VAT, ČEZ dividends, etc.). But it’s money people and firms pay on top—effectively a hidden gas tax. For an average household, that’s 5–8 thousand CZK a year. This must be fixed at the source: the flawed, bureaucratic EU pricing system.
3. Healthcare
a) How to cut costs and improve health? Make GLP-1 drugs free. These medicines don’t just reduce weight; randomized trials show they cut heart attacks and strokes by 20% (SELECT), slow kidney failure in diabetics (FLOW), lower inflammation and atherosclerosis, and protect the liver. They’re not “luxury slimming pills,” but a way to slash the system’s most expensive complications: heart attacks, strokes, dialysis.
Cardiometabolic disease—heart, vessels, diabetes, obesity—consumes a huge share of Czech healthcare spending. Heart disease alone takes about a fifth of the budget. Obesity and diabetes add billions more. Giving GLP-1 free isn’t an extra cost; it’s shifting money from late-stage complications into prevention and treatment, with payback in just a few years. Plus, it shifts the population trajectory toward longer, healthier lives.
b) CZ Biobank. I’ve been deep into AI for healthcare and biotech. Here’s what’s obvious: in ten years, we could live a decade longer with lower healthcare spending. The bottleneck isn’t AI models, it’s data.
Look at the UK Biobank: half a million volunteers, anonymized health data, thousands of studies, faster therapies. Or Finland’s FinnGen, which connects genetics with nationwide health records—GP visits, hospitalizations, prescriptions, deaths—making real-world evidence possible without costly trials.
We need the same for Czechia. Why? Because risks and treatment efficacy differ. We have different genetics, habits, care pathways, languages, and record formats. Models trained on Brits or Finns will systematically fail here. Open, anonymized, secure Czech health data would become fuel for science, medicine, startups, and prevention: better predictions for heart attacks and cancer, faster clinical trials, cheaper prevention. This is a system-shifting lever.
4. Pensions
While the government forces through a retirement-age hike of one month per year (a year over 12 years), AI and robots are about to smash this model to pieces—long before 20 years.
Will AI do our work? Take our jobs or pay for our pensions? Open questions, but they’ll define the future of retirement. Meanwhile, robots will transform care for the elderly. Universal home robots are on the way—able to “take care of the household” or “take care of grandma,” not just wash dishes. They’ll learn from video, copy how you make coffee or repair a roof, and share knowledge globally. Within months, they’ll match human skills. Expected within five years.
That means elder care will soon be easier and cheaper. What should the Czech government do? Help this industry grow here: robot hardware and software. Think about what happens to our “car assembly” industry in a few years (hint: robots). That industrial base is actually a good foundation for a strong robotics and defense sector.
5. Education
AI in schools. In the US, Alpha School is experimenting with “2-hour learning.” Kids learn just two hours a day with AI tutors. Teaching is fully personalized—AI instantly spots where a child struggles and addresses it directly. No waiting for the whole class, no boredom for the fast learners.
The rest of the day? Workshops on drone building (yes), game programming, sports, resilience training. Results? Students rank in the top 1% of US schools in English, math, reading, and science.
This could be the single most important thing we do for Czechia over the next 20 years. And it’s doable fast:
Send a scouting delegation to Alpha Schools.
Launch pilots with volunteer schools here.
Measure outcomes.
Scale if it works.
Start with the two AI hours—it’s the easy part. If results are even half as strong as in the US, it will be the biggest leap since Maria Theresa reformed Czech education.
6. State Governance
A relic of the 1990s: “delegated competence.” To save money, the state pushed hundreds of functions—permits, IDs, passports, licenses, registries—onto municipalities. Clerks sit in town halls but act in the state’s name. The result? Blurred accountability, inconsistent interpretations, bureaucracy. Citizens don’t know whether to complain to the town or the ministry. Cases bounce endlessly between offices.
The fix: pull state agendas back under the state, and digitize them. IDs, passports, licenses, benefits, permits. Leave municipalities their own scope: schools, transport, parking, local development. That way, people know clearly who’s responsible.
This isn’t just about saving billions through centralization and digitization. It’s about clarity and flexibility: one state = one methodology, one digital portal, faster timelines, transparent outcomes. Citizens and businesses will finally know where they stand.
7. Digital Justice and Open Data
Digital court files are already in one program—good. But they should also be open data. Then you can build analytics for the entire judiciary. Suddenly you’d see who judges badly or with bias. Who keeps giving suspended sentences to rapists. Which judges give custody to fathers 90% of the time—or to mothers. Who routinely gives probation where others give six years in prison (and vice versa).
Transparency would change everything.
8. AI in Government and Parliament
How many ministers use ChatGPT daily? I suspect very few. I’d want AI to be one of the “ministers.” No portfolio—just present everywhere. At every meeting it could draft options, pull data, compare proposals, flag blind spots—in minutes.
The easiest way? Don’t build billion-dollar portals. Just appoint a human “body” for AI: attends meetings, feeds queries, delivers top-tier output. Cost: $200 a month for ChatGPT Pro. Once others see the quality gap, adoption will spread.
And parliament? Recently MP Marek Benda bragged about inventing “riders” (sneaking hidden clauses into laws). Then MPs were shocked and vowed to fix it. But what if every bill was reviewed by AI before voting? Check for riders, unintended effects, conflicts, simplifications. Again, Pro version suffices. The AI’s output becomes a standardized checklist, a mandatory appendix to the bill. Pilot it for three months, then decide.
The debates we’re hearing—retirement tweaks, tank purchases, scattered subsidies—are the politics of the past. The world has already moved on. The real levers are here. If we pull them, we can secure our future. If we don’t, it will be very hard to catch up.