Other place · Original Repewo Game
AI EV Powertrain Startup
A three-role startup strategy game about customer value, AI product scope, build-versus-partner decisions, validation, fundraising, strategic partners and scaling an EV powertrain optimization business.
Writer: Repewo
Can the team turn an AI powertrain optimizer into a scalable EV engineering business?
The startup is building software that receives customer priorities and constraints, then recommends the best combination of gearbox, inverter, electric motor and battery from available component options. The algorithm may balance cost, package size, weight, efficiency, range, performance, thermal limits, lifecycle and other requirements. The team has limited cash and time, so it cannot build everything at once.
- The difficult questions are strategic as much as technical: Which part must be developed in-house? What can be integrated from partners? Which customer segment should be first? When is the product mature enough to sell? When should the team expand, hire or enter another market? When should it raise money, from whom, and with what trade-offs?
Deadline: Leave the game with one focused 12-month startup strategy and clear next steps.
Challenge each other. Every proposal to build, raise money or expand must include evidence, a trade-off and a clear next milestone.
Do not merge the roles. Each player should defend their own perspective, challenge the others and negotiate one shared plan.
Exactly three players take different perspectives: Manager, Developer and Marketing & Customer.
Startup Manager
Owns strategy, money, priorities, partnerships and the decision of what the company should build now versus later.
- Keep the product scope narrow enough to survive, but ambitious enough to create a defendable advantage.
- Choose funding and partnership paths without giving away more control than necessary.
Developer
Owns the AI optimization engine, technical architecture, data, validation and integration of motor, inverter, gearbox and battery options.
- Separate core intellectual property from components or software that can be bought, licensed or integrated.
- State technical uncertainty honestly and define the next validation step.
Marketing & Customer
Represents the market, customer discovery, positioning, sales pipeline and evidence about what customers will actually pay for.
- Turn customer needs into measurable input parameters and buying criteria.
- Challenge product features that sound impressive but do not improve customer value.
A customer asks whether your system can design a better EV powertrain for its next vehicle platform.
Possible customer inputs include target vehicle type, price ceiling, packaging space, weight, required range, acceleration or torque, top speed, energy efficiency, battery size, thermal conditions, charging needs, duty cycle, production volume and preferred suppliers. The system output should recommend a component combination and explain the trade-offs, not merely produce a black-box score.
Choose one first use case. Avoid designing for every vehicle and every customer at once.
Tasks: Define the first customer and the exact parameters the customer gives the optimizer. Decide what output is valuable enough that someone would pay for it.
Role check-ins
Startup Manager
- Which customer has the strongest pain and shortest path to a paid pilot?
- Current budget and strategic limits
- Customer evidence and buying process
Final recommendation / output
Select one first customer segment and one paid-pilot goal.
Developer
- Which three customer parameters most influence the buying decision?
- Current technical capabilities and limitations
- Customer priorities and acceptance criteria
Final recommendation / output
Freeze the minimum input/output set for the first version.
Marketing & Customer
- Can we prove the recommended component combination is better than the customer baseline?
- Customer pain, willingness-to-pay assumptions and evaluation criteria
- What can be simulated, benchmarked and validated
Final recommendation / output
Turn the technical result into a customer promise that can be tested.
The team cannot develop every model, database, interface and component library itself.
Possible core assets include the optimization algorithm, decision logic, training or simulation data pipeline, validation framework and customer-specific constraint handling. Non-core work might include generic UI, cloud infrastructure, standard component databases, CAD interfaces or third-party simulation tools—unless one of these becomes a differentiator.
Create four buckets: build now, integrate/buy, partner, postpone. Keep the first release narrow enough to validate quickly.
Tasks: Decide what is core intellectual property and must be developed internally, what can be bought or licensed, what needs a partner, and what should be postponed.
Role check-ins
Startup Manager
- Which technical capability would be hardest for a competitor to copy?
- Runway and strategic priorities
- Core technology, dependencies and technical risk
Final recommendation / output
Agree on the one capability that deserves the deepest in-house investment.
Developer
- How much development time do we actually have before we need customer evidence or funding?
- Effort estimates and technical dependencies
- Runway and milestone priorities
Final recommendation / output
Cut or postpone at least one attractive but nonessential feature.
Marketing & Customer
- Which missing capability blocks a customer pilot, and which only blocks a future scale-up?
- Customer urgency and competitor alternatives
- Technical blockers and shortcuts
Final recommendation / output
Separate must-have pilot features from later scale features.
A technically elegant optimizer is not yet a business. The team needs evidence that customers trust the recommendation and will pay.
A good pilot might compare the customer baseline with two or three optimized powertrain configurations, show cost/efficiency/performance trade-offs, document assumptions and sensitivity, and quantify engineering time saved. Avoid promising production certification before the product and process are ready.
Define one pilot offer, one price or commercial model, one success metric and one reference outcome you want to earn.
Tasks: Design the first paid or strongly qualified pilot: what customer supplies, what the startup delivers, how success is measured, and what proof is needed before expanding.
Role check-ins
Startup Manager
- What commercial signal is strong enough to count as real validation?
- Decision threshold for the next investment
- Pipeline quality and buyer commitment
Final recommendation / output
Set a go/no-go milestone for the next stage.
Developer
- What evidence would make an engineering customer trust the output?
- Benchmark method, uncertainty and technical deliverables
- Customer acceptance criteria
Final recommendation / output
Create a validation package the customer can understand and challenge.
Marketing & Customer
- How much discount or free work can we afford for the first reference?
- Customer willingness-to-pay and sales friction
- Commercial boundary and runway
Final recommendation / output
Create a pilot offer that produces evidence without becoming endless free consulting.
The pilot is promising, but the startup must decide how to finance the next step without losing focus or control.
Personal capital preserves control but is limited. Customer revenue validates demand but may pull the roadmap toward one client. Loans avoid dilution but require repayment and may be hard before predictable revenue. Public grants can fund innovation but often add application and reporting work. Strategic partners can bring data, customers and credibility but may create exclusivity or dependence. Equity investors can finance rapid growth but dilute ownership and add growth expectations.
Build a financing ladder: what to use now, what milestone unlocks the next source, and what you refuse to accept.
Tasks: Compare bootstrapping/personal capital, revenue from customers, bank debt, public grants, strategic industry partners and equity investors. Choose a funding sequence, not just one source.
Role check-ins
Startup Manager
- Which funding source also improves market access rather than only adding cash?
- Runway need, ownership limits and financing constraints
- Commercial leverage of partners and customers
Final recommendation / output
Choose the first financing source and the milestone for the next one.
Developer
- Which technical work is worth financing before customer proof, and which should wait?
- Technical milestone costs and risk
- Available financing envelope
Final recommendation / output
Tie funding directly to a measurable technical milestone.
Marketing & Customer
- Would this partner accelerate sales or make us dependent on one ecosystem?
- Partner benefits, exclusivity risk and customer access
- Control limits and strategic boundaries
Final recommendation / output
Define partner criteria before approaching anyone.
Three things happen at once: one large customer wants a custom feature, another market shows interest, and your runway is getting shorter.
The custom feature could bring revenue but may consume months. Entering a new market could increase growth but needs sales effort, support and possibly localization or regulatory work. A potential investor offers capital but wants a meaningful equity stake and aggressive expansion. The team must decide what not to do.
For each issue decide: now, later after a milestone, through a partner, or not at all. Then state who owns the next action and what could make the decision change.
Tasks: Choose a response to each pressure: custom development, market expansion, hiring and fundraising. Define the trigger that would make you expand later if you decide not to expand now.
Role check-ins
Startup Manager
- What technical debt or integration risk appears if we accept the custom request now?
- Runway, hiring and ownership implications
- Engineering impact and delay risk
Final recommendation / output
Decide whether customization is core, paid exception, partner work or rejection.
Developer
- Is the requested feature unique to one customer or evidence of a broader market need?
- Engineering cost and architectural impact
- Demand evidence across customers
Final recommendation / output
Choose whether to generalize, custom-build, partner or postpone.
Marketing & Customer
- What minimum pipeline or revenue signal justifies expansion?
- Market evidence, sales cycle and partner options
- Runway and risk tolerance
Final recommendation / output
Set a measurable expansion trigger.
The three roles must agree on one 12-month direction.
What strategy should the startup commit to now?
Your startup now has a focused strategy instead of a list of possibilities.
A strong result connects customer evidence, technical scope, validation, funding and expansion in the right order. The team should know what it will build now, what it will deliberately not build, what proof comes next and which financing source fits each stage.
Task 40 · Evidence 35 · Peer 25
Startup Manager
78/100Developer
78/100Marketing & Customer
78/100The team connected customer value, technical scope, validation, financing and expansion into one coherent startup strategy.
Which decision changed most after hearing the other two roles, and what evidence would make you revisit the strategy?