Financial models for
technology decisions
Interactive calculators built for CXOs and their teams. No login. No vendor affiliation. Change any input and the model recalculates instantly.
Cloud vs. On-Premises Cost Calculator
8-year NPV comparison with crossover year detection. Models workload growth, cloud discount tiers, hardware refresh cycles, staffing costs, and depreciation tax shield. Three preloaded scenarios.
Infrastructure · NPV · Crossover AnalysisAI GPU Infrastructure Calculator
Cloud vs. on-premises GPU economics. H100/A100/L40S/A10G pricing. Training vs. inference workload split. Utilization break-even analysis. Power cost with PUE. MLOps staffing.
GPU Economics · Training vs. InferenceEnterprise Token Economics Calculator
Model LLM consumption at organizational scale. Seven canonical workload profiles — IDE assist, agentic coding, CS policy lookup, research, data agents, and more. Heavy-tail user distribution with Pareto visualization. Governance-lever sensitivity: per-user caps, Opus→Sonnet routing, agent timeouts, and agentic intensity stress-testing. Full sourcing on every default.
Token Economics · Heavy-Tail Distribution · Governance LeversCost Per Prediction (CPP) Calculator
Decompose the unit economics of a single AI inference. Three deployment modes — self-hosted GPU, cloud GPU rental, public API — with crossover breakeven analysis between them. Energy decomposition (GPU draw, server overhead, PUE-implied cooling), hardware amortization curve, log-log scalability chart showing where one mode beats another. Sensitivity tornado on electricity, PUE, utilization, and capex. Step-by-step worked example with side-by-side run comparison. Full sourcing on every default.
Unit Economics · Breakeven Analysis · Energy DecompositionMulti-Cloud CapEx Comparison Calculator
Compare 3-year TCO across AWS, Microsoft Azure, and Google Cloud for the same workload. Commitment-strategy mix — on-demand, 1-year reserved, 3-year reserved, spot — applied uniformly for fair comparison. Per-vendor pricing cards seeded with verified Q2 2026 list rates and editable to your negotiated quote. Six Decision Context flags that reshape the verdict for what price alone cannot answer: Microsoft-stack alignment, existing committed spend, compliance gates, vendor-locked services, multi-cloud mandates, and data gravity.
Multi-Cloud TCO · Commitment Strategy · Decision ContextNeoCloud Sourcing & Fit Calculator
Decide whether you need dedicated NeoCloud GPU capacity — and then which provider fits. Stage one prices the same demand five ways (hyperscaler baseline, serverless, on-demand, reserved-at-peak, and a blended reserve-base-plus-burst portfolio) and finds the reserve-vs-burst utilization breakeven. Stage two ranks CoreWeave, Nebius, Lambda, Crusoe, IREN, and RunPod on a weighted ten-criterion matrix where counterparty financial durability is weighted automatically by commitment term. Full sourcing on every default.
Sourcing Decision · Reserve-vs-Burst Breakeven · Counterparty RiskFive tiers of AI adoption, one financial model per tier. The framing essay is published; the calculators are in development.
The Five-Tier AI Decision Taxonomy
What "going AI" actually means — five tiers, each with a different cost structure, time-to-value, and competitive moat. The framing essay that anchors every calculator below.
Published · Machine Capital Issue 001 · Mar 5, 2026Tier 1 — Packaged AI: The AI Already In Your Software
True cost of vendor-embedded AI features — Dynamics, Einstein, ServiceNow, Workday. Licensing uplift, integration, change management, and the productivity dip nobody puts in the pitch deck.
Coming SoonTier 2 — Point AI Tools: Departmental Adoption at Speed
The economics of AI sprawl. Subscription proliferation, governance overhead, security review load, and the breach exposure cost when shadow AI meets sensitive data.
Coming SoonTier 3 — Embedded ML: Where Real Competitive Advantage Begins
The full cost of custom ML on proprietary data. Model development, the MLOps tax (3–5× development cost), data pipelines, drift detection, and retraining cycles.
Coming SoonTier 4 — AI-Augmented SDLC: The Fastest ROI Nobody Is Maximizing
Productivity uplift model for GitHub Copilot, Azure Document Intelligence, automated QA, and AI-assisted code review. Adoption-curve adjusted; culture-change investment included.
Coming SoonTier 5 — ML Platform Capability: The Long Game
The CapEx case for owned AI infrastructure. H100 cluster economics, multi-year talent commitment, model refresh cycles. Build this only if your data flywheel is real.
Coming SoonComing in Issue 002: Buy vs. Build Economics — a cross-tier TCO calculator covering all five tiers, with time-to-value curves and a calculator view per tier.
All tools use standard financial modeling conventions: straight-line depreciation, compound growth, after-tax NPV discounted at WACC. Cloud pricing sourced from AWS/Azure/GCP public rate cards. GPU hardware costs from Dell/Supermicro published SKUs. Every formula is visible in the tool. These are starting points for analysis, not substitutes for vendor quotes and CFO review.
Questions on methodology or custom scenarios: contact us.