Automotive industry: AI optimization of investments in e-mobility, platforms, plants, software and supply chains
Capital allocation from prioritization to mathematical optimization
Companies usually prioritize projects based on business cases, rankings and committee decisions. This approach seems rational, but does not take the entire decision-making space into account.
There are already over 1 billion possible portfolio combinations for 30 projects and over 1 quadrillion for 50 projects. Traditional methods cannot fully evaluate this space. They select a plausible solution - but not necessarily the optimal one.
Project Portfolio Optimization AI calculates the optimal project portfolio under your real constraints - including budget, resources, risk and strategic guidelines. The result is a comprehensible, mathematically sound decision-making basis for capital allocation.
For decision-makers, this means a structural difference: decisions are no longer based on approximation, but on calculated optimality.
Starting point: The complete investment list before the actual decision
The decisive difference in this new calculation method lies in the time of application: it is not used for validation after the decision has been made, but before the actual decision is made, based on the company's complete investment and project list.
Typically, there is a list of potential CAPEX projects - e.g. plant modernizations, IT transformations, product developments, Infrastructure measures or efficiency programs. At the same time, there are fixed restrictions such as a limited overall budget, limited engineering capacities, Production windows, risk budgets and strategic framework conditions.
This is precisely where the real decision-making problem arises: not all projects can be implemented. The question is therefore not which projects appear to make sense in isolation, but rather which combination of these projects forms the globally optimal overall portfolio under the given restrictions.
The new calculation method therefore does not evaluate individual projects in isolation, but calculates from the complete project list the optimal portfolio, taking into account all budget, capacity, risk and strategy limits. The result is a mathematically sound The result is a mathematically based selection of those projects that together generate the maximum overall value contribution - before the actual investment decision is made. Deviations from the calculated optimal starting position are made with explicit visibility of the resulting opportunity costs and their quantifiable impact on the overall portfolio value.
This transforms CAPEX planning from a sequential selection process to a consistent portfolio optimization, in which opportunity costs, restriction bottlenecks and portfolio effects are fully taken into account.
Projects do not disappear - they are better positioned and optimally planned over several years
In a mathematically optimized investment system, projects are not discarded. Instead, they are reprioritized, postponed or strategically repositioned, so that they make the maximum economic contribution to the overall portfolio at the optimum time under given budget, capacity and risk restrictions the maximum economic contribution to the overall portfolio.
The decisive factor here is the multi-year perspective. Investment decisions are not made in isolation for a single year, but are optimized in the context of 2-, 3-, 5- or 10-year plans.
Liquidity generated by optimization in the start year is systematically carried over to the following year transferred to the following year. This increases the available investment budget for the next period. This subsequent year is then also optimized again.
The effect: projects can be added as soon as they fit into the globally optimized portfolio under the new budget, capacity and return conditions, Capacity and return conditions fit into the globally optimized portfolio. This creates a dynamic multi-year optimization in which each optimization period Optimization period structurally improves the investment opportunities of the following years.
Automotive example:
10 projects. Fixed budget: EUR 850 million. Total investment costs: EUR 2088 million .
From mathematical model to practical application
The optimization logic can be used across all industries and can be applied to real investment, CAPEX, R&D and infrastructure portfolios. The decisive factor is not the type of project, but the structure of the decision: limited resources, competing options and clear constraints.
At the same time, the system architecture has been consistently designed for data minimization and confidentiality. Only numerical project parameters are required for the calculation. Content descriptions, strategy papers or project-specific narratives are neither required nor interpretable.
Below you can see specific use cases and the underlying data protection and data minimization architecture.
Executive Summary
Automotive manufacturers are undergoing the biggest capital allocation transformation since the invention of the internal combustion engine.
Billions invested in electromobility, software-defined vehicles, new platform architectures, battery plants and supply chains will determine which manufacturers will dominate the coming decades - and which will structurally destroy capital.
Strategic success is not determined by the quality of individual projects, but by the mathematical optimality of the entire investment portfolio under real restrictions.
The challenge is combinatorial: as soon as a selection is made from dozens or hundreds of potential investments, the number of possible combinations grows exponentially. At this point, traditional decision-making processes - even with the highest level of management expertise - can no longer fully capture the decision space.
Project Portfolio Optimization AI makes it possible for the first time to systematically calculate the globally optimal investment portfolio under real budget, resource, risk and strategy constraints.
This fundamentally changes capital allocation - from heuristic decision-making to mathematically optimal portfolio optimization.
1. Automotive manufacturers as capital allocation systems
Every OEM and supplier operates under multiple simultaneous constraints:
- CAPEX budgets for platforms, plants and software
- Engineering capacities in electronics, software and battery technology
- Production capacities and plant utilization
- Supply chain availability of critical components
- CO₂ fleet regulation and compliance requirements
- Strategic roadmap constraints (e.g. full electrification by year X)
Formally, this is a combinatorial optimization problem.
Assume a manufacturer evaluates N investment projects:
- New electric platform
- Conversion of an existing plant
- Development of a new software architecture
- Battery plant joint venture
- Vertical integration of critical components
- Autonomy software programs
- New vehicle models and derivatives
Each project has measurable parameters:
- Expected portfolio contribution (Ri)
- Investment requirement (Ci)
- Risk exposure (σi)
- Strategic contribution (Si)
- Resource requirements (engineering, production, supply chain)
The goal is to select the optimal subset of these projects:
max Σ Ri xi
s.t. Σ Ci xi ≤ Budget
xi ∈ {0,1}
2. Combinatorial reality in the automotive industry
Already exist with 50 potential investment projects:
2⁵⁰ = 1,125,899,906,842,624 possible portfolios
This corresponds to over a quadrillion possible strategic future paths for a manufacturer.
No management board, no strategy team and no spreadsheet can fully evaluate this space.
In practice, approximation methods are used instead:
- ROI ranking of individual projects
- Top-down budget allocation
- Political and organizational prioritization
- Sequential decision-making processes
- Legacy-based investment patterns
These methods do not calculate the optimal portfolio - they approximate it.
3. Typical investment decisions in the transformation to electromobility
Example 1: Electric platform vs. further development of existing platform
A manufacturer is faced with a decision:
- Investment in a completely new EV platform: €4 billion
- Further development of existing platform: €1.8 billion
- Hybrid strategy with several intermediate solutions
The optimal decision does not depend on the individual project, but on its interaction with
- planned vehicle derivatives
- Software architecture
- Production plants
- Supply chain structure
- future regulatory requirements
Example 2: Plant transformation or new construction
A manufacturer owns an existing plant for combustion engine production.
Options:
- Conversion to EV plant: €1.2 billion
- New construction of an EV plant: €2.4 billion
- Outsourcing to contract manufacturer
The optimal decision depends on the overall portfolio:
- planned model strategy
- Platform decisions
- Production volume planning
- geographical sales forecasts
Example 3: Software-Defined Vehicle architecture
Investment options:
- In-house development of software stack: €3 billion
- Partnership with tech companies
- Licensing of existing platforms
This decision has a long-term impact:
- Margin structure
- Differentiation potential
- Update and lifecycle costs
- strategic control over the vehicle
Example 4: Battery supply chain and vertical integration
Options:
- Own battery plant
- Joint venture
- External procurement
This decision influences
- Product cost structure over decades
- Supply chain risk
- Capital commitment
- strategic flexibility
4. Why classic decision-making logic is structurally suboptimal
The core problem: projects are not independent.
They interact systemically:
- A new platform enables several future models
- One plant determines production capacities for decades
- Software architecture influences the entire product strategy
- Battery strategy influences cost structure and margins in the long term
This follows:
Portfolio value ≠ Sum of isolated project evaluations
But not:
Portfolio value = f(interactions, constraints, roadmap, resources)
5. Mathematical foundation of AI-supported portfolio optimization
Formally, this is a binary integer optimization problem:
max Rᵀx
s.t. Ax ≤ b
x ∈ {0,1}
With:
- x = decision vector
- R = Portfolio contribution of the projects
- A = Constraint matrix (budget, resources, strategy, production)
- b = Restriction limits
This structure enables mathematically precise modeling of real automotive restrictions.
6. Specific automotive use cases for Portfolio Optimization AI
OEM strategy planning
- Optimal combination of platform investments
- Model portfolio optimization
- CAPEX allocation over several years
Plant network optimization
- Which plants should be transformed
- Which plants should be closed
- Where new plants should be built
Software investment strategy
- Build vs. buy vs. partner decisions
- Optimal roadmap prioritization
- Minimization of long-term architecture costs
Battery and supply chain strategy
- Optimal vertical integration
- Joint venture vs. in-house production
- Risk minimization for critical components
7. Impact on company value and competitiveness
Even small improvements in capital allocation lead to massive long-term effects.
With annual investments of:
10 billion € CAPEX
just 5% better portfolio optimization leads to:
500 million € additional value creation per year
Over 10 years, this corresponds to
€5 billion in additional enterprise value
8. Governance implications for the Executive Board and Supervisory Board
Portfolio Optimization AI fundamentally changes the role of management.
From:
- Heuristic decision-making
- political prioritization
- incremental budgeting
To:
- mathematically optimal capital allocation
- complete transparency of opportunity costs
- systematic maximization of company value
9. Strategic importance for the future of the automotive industry
The transformation to electromobility is not primarily a technology problem.
It is a capital allocation problem.
Manufacturers that allocate their investments in a mathematically optimal way will achieve structurally higher returns, faster transformation and long-term competitiveness.
Portfolio Optimization AI provides the decisive mathematical basis for this.
Conclusion
The future of the automotive industry will not be decided by individual technologies, but by the quality of capital allocation across thousands of simultaneous investment decisions.
For the first time, AI-supported portfolio optimization enables the systematic calculation of the globally optimal investment portfolio under real industrial restrictions.
This marks the transition from heuristic decision-making to mathematically optimized corporate management.