Skip to main content
Back to challenges
Challenge

Optimize AI Data Center ROI

The rapid global expansion of AI data centers presents immense investment opportunities but also significant ROI and risk management challenges. This challenge tasks you with building a sophisticated multi-agent system to analyze prospective data center projects. Your system will leverage Gemini 2.5 Pro for advanced strategic reasoning, utilizing its extended thinking capabilities to dissect complex financial models, regulatory landscapes, and energy market dynamics. The core of the system will be built with LangGraph, enabling robust, graph-based agent workflows that manage state, facilitate collaboration between specialized agents (e.g., Financial Analyst, Regulatory Expert), and allow for adaptive reasoning budgets. A critical component is the integration of the Model Context Protocol (Model Context Protocol) for secure and efficient tool integration, allowing agents to access and process mock enterprise data sources such as energy market APIs, local regulatory databases, and infrastructure cost repositories. The goal is to generate actionable investment recommendations and comprehensive risk assessments for new AI data center initiatives.

Business OperationsHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Build a graph-based multi-agent system using LangGraph and Model Context Protocol (Model Context Protocol) with Gemini 2.5 Pro for optimizing AI data center investment and operational planning.

Delivery guide

How work is evaluated

Evaluation

The evaluation will assess the agent system's ability to generate comprehensive ROI analyses and risk assessments for data center projects, demonstrating effective use of LangGraph for workflow orchestration, Model Context Protocol (MCP) for tool integration, and extended thinking with Gemini 2.5 Pro for strategic decision-making.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Learning goals

What you should walk away with

  • Develop adaptive reasoning budgets for cost-effective agent execution.

  • Integrate Model Context Protocol (Model Context Protocol) for secure enterprise data access.

  • Apply extended thinking with Gemini 2.5 Pro for strategic planning.

  • Implement graph-based agent workflows using LangGraph for state management.

  • Design multi-agent systems for complex business analysis.

Resources and assets

Reference links and supporting material

Dataset notes

Sample data for 3 tasks

Challenge setup

The evaluation will assess the agent system's ability to generate comprehensive ROI analyses and risk assessments for data center projects, demonstrating effective use of LangGraph for workflow orchestration, Model Co...

Challenge input

{ 'project_name': 'New AI DC (5GW, NoVA)', 'location': 'Northern Virginia', 'capacity_gw': 5, 'cost_per_mw': 7e6, 'energy_cost_per_kwh': 0.12, 'pro...

Agent execution

No agent tool is listed for this challenge.

Evaluated output

{ 'project_name': '...', 'location': '...', 'roi_projection': { 'years': 5, 'value': 'XX.XX%', 'confidence': 'High' }, 'risk_summary': [ { 'type':...

Evaluation checks
  • Verify all required fields in the ROI analysis and risk assessment reports are present and non-empty.
  • Verify that the log trace shows correct Model Context Protocol (MCP) tool calls and successful response han...
  • Ensure the 'recommendation' field in the ROI report is one of 'Invest', 'Monitor', or 'Reject'.
Configured evaluation evidence
  • ROI Accuracy (Deviation) target: 5
Configured runtime
  • Python execution harness

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

No tool records are attached.

Evaluation contract

  • Verify all required fields in the ROI analysis and risk assessment reports are present and non-empty.
  • Verify that the log trace shows correct Model Context Protocol (MCP) tool calls and successful response handling.
  • Ensure the 'recommendation' field in the ROI report is one of 'Invest', 'Monitor', or 'Reject'.

Metrics

  • ROI Accuracy (Deviation) · Target 5
  • Risk Identification Score · Target 0.9
  • Adaptive Reasoning Budget Utilization · Target 0.8

Workflow tasks

  1. Task 1

    Input: { 'project_name': 'New AI DC (5GW, NoVA)', 'location': 'Northern Virginia', 'capacity_gw': 5, 'cost_per_mw': 7e6, 'energy_cost_per_kwh': 0.12, 'projected_demand_growth': 0.15 }

    Output: { 'project_name': '...', 'location': '...', 'roi_projection': { 'years': 5, 'value': 'XX.XX%', 'confidence': 'High' }, 'risk_summary': [ { 'type': 'Regulatory', 'level': 'Medium', 'mitigation': '...' } ], 'recommendation': 'Invest|Monitor|Reject' }

    Execution harness: Python

  2. Task 2

    Input: { 'initial_plan_summary': 'JSON summary from previous task', 'energy_price_increase_percent': 0.25 }

    Output: { 'scenario': 'Energy Price Spike', 'impact_on_roi': 'Negative XX.X%', 'new_risks': [ { 'type': '...', 'level': '...', 'details': '...' } ], 'mitigation_strategies_update': [ '...' ] }

    Execution harness: Python

  3. Task 3

    Input: { 'location': 'Texas' }

    Output: { 'log_entries': [ { 'timestamp': '...', 'agent': '...', 'action': 'MCP Tool Call', 'tool_name': 'get_regulatory_status', 'params': { 'location': 'Texas' } }, { 'timestamp': '...', 'agent': '...', 'action': 'MCP Tool Response', 'data': { 'status': 'Green', 'incentives': ['Tax Abatement', 'Renewable Energy Credits'] } } ] }

    Execution harness: Python

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Optimize AI Data Center ROI