
Design. Detonate. Survive. As the Chief Engineer of the Kurchatov Institute, you must design atomic devices to satisfy the demands of the State. A hardcore physics-based puzzle game simulating hydrodynamics, neutron transport, and thermonuclear fusion.
Modeled estimate, not a reported figure. How we estimate →
About
NUCLEAR DESIGN BUREAU is a hardcore sandbox engineering game where you design and test nuclear devices.
You are the Chief Engineer of a secret Soviet design bureau. The State gives you a target and a warehouse full of dangerous materials. In some directives, your available materials are limited. Your task is simple: build a device that gets the job done.
THIS IS A PROTOTYPE, NOT A FINISHED GAME. THERE WILL BE BUGS, MISSING OR BROKEN FEATURES, AND BIG CHANGES COMING.
HOW IT WORKS
- Receive Directive: Read technical mission requirements and constraints, including material limits in specific directives.
- Design: Paint materials directly onto the blueprint grid. Build your primary, casing, and supporting systems.
- Simulate: Hit the red button, run full detonation physics in real time, and watch the blast in the detonation viewer.
- Refine: If it fizzles, fails, or wastes material, iterate and improve.
CORE FEATURES
- Physics-Driven Simulation: Shockwaves, heat, radiation, and chain-reaction behavior are modeled in real time.
- Detonation Viewer: Observe detonations as they unfold, with live visual feedback on blast behavior.
- Campaign Directives: Structured mission progression with technical goals and constraints.
- Sandbox Mode: Unlimited experimentation with full design freedom.
- Iterative Engineering Loop: Build, test, diagnose, and optimize.
EARLY ACCESS STATUS (CURRENT BUILD)
The current Early Access build is fully playable and includes the core gameplay loop:
- Design -> Simulate -> Refine
- Sandbox mode
- Campaign directives
- Save/load systems with ongoing stability improvements
This is an active development build. Expect balance changes, physics upgrades, and campaign iteration during Early Access.
EARLY ACCESS ROADMAP (SUBJECT TO CHANGE)
Physics and Simulation
- More realistic detonation viewer (better shockwaves, 3D-style presentation, damage radius + thermal radiation modeling, destructible buildings/targets)
- Smaller grid mode for higher realism
- Improved compression physics
- Levitated pit physics
- Fission boosting with D-T gas
Campaign and Content
- Rebuild and expand campaign around updated grid, physics, and destruction systems
- Many more directives/missions
- Voiced campaign
- Additional bomb concepts (high-radiation designs, clean bombs, neutron bombs)
- More advanced materials
Presentation and Audio
- More music tracks
Community and Meta
- Public leaderboard
DEVELOPMENT APPROACH
Early Access is where this game gets shaped in public. Player feedback will guide balancing, campaign pacing, clarity, and polish as we move toward full release.
"The atom is not a soldier, Comrade. It is a force of nature. We merely tell it where to go."
Genres
Stats
All-time low is the lowest price we've recorded for this game since we started tracking it.
Languages: English
Engagement
Reviews
Themes across 50 recent reviews and how positive each mention is. A keyword signal, not full sentiment analysis.
Reviews
This is a fun idea unfortunately largely held back by vibe coding, a buggy poorly designed editor and poor simulation accuracy. I hope that things improve through early access but releasing in this state is rough. The campaign/dossier/tutorial prompting is all clearly lightly edited LLM output, with piles of unhelpful and incomprehensible -isms like "The real X", "reproducible physical source history", and many others. Even the settings text contains direct LLM output; there is an entire section labelled "Windows 11 Alt+Tab investigation" directly from a coding agent talking about triggers clearly from the dev trying to fix a bug and not vetting the subsequent code changes at all. The editor: * You can't see k-values when placing masses because it tells you to "run transport to measure" (what does that even mean???). * Using the fill tool will break selections, making them unselectable and the simulation will use a completely different assembly than the editor shows. The undo tool does nothing and the masses disappear upon game restart and save load. * Undo/redoing too much at once often hangs and crashes the game. * There is no scale tool and if you want to slightly change something it's often easier to completely rebuild it from scratch. * Tons of useful metrics are missing from the materials, from density to reflection cross sections. * No way to separate masses of the same type without leaving air gaps (and similarly, no way to attach dissimilar metals together). * MS Paint's tools with materials for colors would honestly be a massive improvement outside of specific tools like lenses. The simulator: * Is slow as hell and single threaded. * Accuracy is not great, as evidenced by trying to make a simple recoilless gun. Anything with more than one layer you put in the barrel will jam almost immediately regardless of pressure unless you place air gaps everywhere. Masses similarly act like jello under any acceleration, well before stresses to the point they should start acting like it. I'm guessing this is an oversight in how this game tries to run FEA on single material bodies. * Has no controls (pause/investigate) nor useful metrics like gamma/x-ray density, temperature, velocity or anything aside from vague globally averaged graphs which themselves have no time or zoom controls. TL;DR this needs another year in the oven, preferably with far more attention to detail than the current vibe coding approach has shown. I really hope this gets better because there is a dearth of good sandbox-style games like this.
Media

Related
Related
Price dynamics, follower growth, revenue and sales trajectories with interactive tooltips
This and every other Pro feature is free until October 20, 2026. A free account is all it takes.
Hours played at review time, across 50 reviews with recorded playtime.
positive critical· bar length = how often the theme comes up
Disappointing. Annoying UI, bad scaling. Tutorial messages overlap with settings menu, detonation viewer lags the rest of the game. Borderless window mode can be accidentally resized if you try to drag an element close to the edge. Lots of small irritating things like that that make me not want to play the game. I understand this is a cheap game and early access, but for something that is 99% placing shapes on a grid, it is inexcusable for the 'placing shapes' part to be this fiddly and annoying.
Remember how you were making all those nukes in Powder Toy when you were 10? Yeah, this is it but for the grown-ups. The 3D detonation render is a bit laggy but otherwise, it's cool. I'm glad this game is out of the very specific kind of development hell it was in.
Great fun. Nice physics simulations, probably about as accurate as you can get them as a developer without an unfortunate visit from one or two three-letter agencies. Doesn't hold your hand too much, which is nice. Made a bomb so big I crashed my game. 10/10. Dev, if you're after feature requests: let us record the simulation and then play it back afterwards at a configurable speed, so we can watch stuff back without needing to re-run the simulation itself. Edit: re-running without the crash netted a 2.2MT bomb without looking at the manual, in a playtime of about 3.7 hours. Latent knowledge from being a teenage physics geek back in the day pays off! Plenty of replayability though, messing around optimising the designs. I'm sure the community will have fun trying to make the biggest boom they can.
First thing I did was recreate the demon core, then I did it again in the experimental first person mode. It's a pretty dry experience, but it's exactly what I wanted out of a nuclear weapon design game. 10/10 would irradiate myself again.
It's a fun simulation that gets enough right to be interesting. The UI and sort of... product in general need some work, but (appropriately) the core is here. It is extremely easy to design a weapon that gets some yield if you understand the basic principles, but appropriately difficult to make an efficient weapon. I think I'll be spending a shameful number of hours playing around with this in the coming months.
Do not buy this, especially not in its current state. The AI use nearly stopped me from buying, and I was about to excuse the AI use if it wasn't for a game-breaking bug where switching from sphere to drag then back in the editor would make clicking and dragging to draw stop working. There's also AI images in the manual. This game, at least right now, is not worth the $10USD I paid.
It's interesting, and I really want to like it, but the game needs a lot longer to cook before I could say it's worth playing. I think it's also important to point out that nearly all the text is clearly written with AI, which, along with some of the AI art, just robs this game of soul, which is a shame.
Nuclear Design Bureau is an absolute triumph of niche engineering simulation. It nails the claustrophobic tension, Cold War paranoia, and terrifying physics of early atomic testing better than anything else out there. If your ideal game night involves stare-downs with differential equations, balancing neutron transport graphs, and the looming threat of state reprimand—this is an instant buy. It scratches the exact same ruthless, uncompromising itch as Children of a Dead Earth, HighFleet, and Kerbal Space Program (with nuclear physics dialed up to 11). Score: 9.5 / 10 (10/10 would accidentally vaporize a subcritical beryllium tamper into radioactive slag again.)
Wow, is mine really the first review? Awesome I saw Scott Manley's playthrough literally yesterday and wishlisted immediately, believing it would be months before we saw the first early access release and I'm very pleasantly surprised Obviously, I've barely started but already, I love the idea, love the music, the KSP-vibes of the 3D lab you can interact with (which I don't remember Scott showing off) I really love where this is going. The fact the NNSA had to be convinced this could be published due to security concerns is a great selling point from my point of view, lol. I can't wait to get deeper into this One small critique I have is that a scalable interface would be helpful, which I'm sure is planned or might even already exist but I haven't been able to find it yet I'll revisit this review once I've played awhile but for such a one of a kind game, I'm in it for the long-haul already
Selling now
See how fast this game is selling right now — estimated copies and revenue per day from recent review growth.
Players by region
Where this game's players are, estimated from review languages.
Related
| Game | Estimated lifetime revenue | Reviews | Review score | Price |
|---|---|---|---|---|
| $2.5K | 19 | 89% | $4.99 | |
| $170.5K | 207 | 92% | $19.99 | |
| — | 0 | — | — | |
| — | 0 | — | — | |
| $199.5K | 321 | 92% | $14.99 | |
| $285.2K | 1.4K | 86% | $9.99 |
Related
| Game | Estimated lifetime revenue | Reviews | Review score | Price |
|---|---|---|---|---|
| $24.5K | 116 | 90% | $9.99 | |
| $24.5K | 172 | 78% | $5.99 | |
| $24.5K | 145 | 90% | $7.99 | |
| $24.5K | 90 | 100% | $10.99 | |
| $24.5K | 63 | 87% | $17.99 | |
| $24.5K | 144 | 76% | $7.99 |