Macadamia

What do they actually do

Macadamia builds a customer‑installed AI engineering agent called Cooper that connects to a team’s existing project data (drives, emails, CAD/CAE outputs, spreadsheets, PDFs) and tools. Cooper ingests and indexes this data, runs checks to spot inconsistencies and likely design errors, and then surfaces flagged issues with suggested fixes; it can be configured to apply certain low‑risk fixes automatically YC profile macadamialabs.com.

Today the company delivers hands‑on, tailored deployments with a proof‑of‑concept in roughly 1–2 weeks before a paid engagement. Their go‑to‑market emphasizes integrations with existing software, enterprise security and compliance (e.g., SOC 2 Type II, GDPR), and a consultative, founder‑led sales motion macadamialabs.com YC profile.

Who are their target customer(s)

• Mechanical design lead at an OEM or product team: Spends hours manually checking CAD models/drawings for clashes or spec mismatches, so errors are caught late and cause rework and delays.
• Manufacturing/process engineer: Receives designs that don’t match tooling or assembly constraints, leading to production delays and rejects on the shop floor.
• Construction project engineer / site coordinator: Must reconcile many versions of drawings, P&IDs and PDFs from different parties; inconsistencies lead to rework, safety risks, and change orders.
• CAD/PLM administrator or engineering data manager: Data is fragmented across drives, emails and tools, making it hard to enforce design rules or trace changes across systems.
• Quality assurance / document control or supplier coordinator: Chasing updated specs and verifying supplier drawings is manual and error‑prone, leading to audit findings, delays, and cost overruns.

How would they acquire their first 10, 50, and 100 customers

• First 10: Founder‑led outreach to engineering leads at OEMs, manufacturers, and construction firms via warm intros and targeted outbound; deliver a tailored 1–2 week POC that finds concrete design issues and converts via measured impact and a referenceable case study.
• First 50: Standardize 2–3 industry‑specific POC templates, hire sales engineers to run integrations, and formalize a pilot‑to‑paid path. Use early case studies for targeted outbound and partner with CAD/PLM resellers and systems integrators.
• First 100: Productize repeatable agent flows and common connectors for faster onboarding and safe automation; add enterprise trust signals and a direct sales team. Scale through certified partners, technical content/webinars, and a referral program for upsells.

What is the rough total addressable market

Top-down context:

Closest near‑term category is engineering/design software (CAD/CAE/PLM/AEC), estimated at about USD ~$43B in 2024, which aligns with budgets for tools that prevent design errors and rework Grand View Research. Upside exists as enterprises tap broader manufacturing/industrial software budgets and emerging AI/agent spend, though these figures overlap the core market AppsRunTheWorld IoT Analytics ABI Research industry overview.

Bottom-up calculation:

Illustratively, if Macadamia sells into 10,000 mid‑to‑large engineering teams across mechanical OEMs, manufacturing, and construction with an average annual contract of $50k–$150k for error detection and rule enforcement, the initial bottom‑up opportunity is roughly $0.5B–$1.5B.

Assumptions:

• There are ~10,000 viable target teams with enough data complexity and budget in NA/EU and select global markets.
• Average ACV ranges $50k–$150k based on integrations, scope, and governance needs.
• Budgets are pulled from existing engineering QA/PLM/AEC spend rather than requiring net‑new AI line items.

Who are some of their notable competitors

• Autodesk (Construction Cloud / Model Coordination): Widely used for aggregating models, automated clash detection, and issue coordination in construction/manufacturing. Strong for model coordination but not positioned as a tailored, customer‑installed AI agent across arbitrary drives/emails and mixed docs source guide.
• Siemens (Teamcenter / Xcelerator): Enterprise PLM and digital‑thread platform for requirements, CAD, change control, and design rule enforcement. Deep PLM strength; less focused on autonomous agents that inspect mixed external data sources and propose/apply fixes source.
• PTC (Creo + Windchill): CAD/PLM stack with validation and integrated lifecycle management. Excels at CAD↔PLM workflows; not primarily an AI agent scanning emails/drives for cross‑document inconsistencies or automated fixes out of the box Creo Windchill.
• Cognite (Data Fusion): Industrial data platform that contextualizes 3D models, P&IDs, time series, and documents for analytics/AI apps. Strong data infrastructure; more a platform than a turnkey agent proposing or applying fixes on design files source.
• Avvir: Construction QA tool comparing as‑built scans/photos to BIM to find deviations. Focused on site verification rather than general CAD/CAE rule enforcement or automated design fixes across an OEM’s engineering data source background.