Contact us

Thank you for contacting us!

Your submission has been received and we'll be in contact with you shortly.

Return home
Oops! Something went wrong while submitting the form.
Dashboard showing SpaceX Merlin Engine product details with mass 630 kg, emissions, and cost $22,303.73 USD.

Register for your free Muir account

We'll give you access to the platform and five free PCFs.

Thank you for registering

We will follow up shortly with your account information.

Oops! Something went wrong while submitting the form.
Gold close icon
Right arrow
Back

November 2026

Finding the Part Before It Disappears: Muir's Air Force SBIR Phase II on Obsolescence and Supplier Risk

Muir AI's Air Force SBIR Phase II applies generative AI to find obsolescence and single-source risk across multi-tier bills of materials, without relying on suppliers to volunteer the data.

The part that no longer exists

Aircraft, ground vehicles and industrial equipment stay in service for decades. The parts inside them do not. A connector gets discontinued, a casting supplier closes, a chip goes end-of-life, and a program finds out when an order comes back unfillable.

At that point the options are expensive: a lifetime buy, a redesign, a requalification, or a search for a supplier who can still make it. Each one costs more the later it starts.

Why obsolescence is still managed after the fact

Most obsolescence management starts from a parts list someone already has. The hard cases are the ones where that list is incomplete, out of date, or stops at the first tier.

  • No full bill of materials. Older platforms often lack a current, multi-tier BoM. Sub-tier parts and materials are invisible until one fails to ship.
  • No supplier input. The organizations sustaining a platform frequently cannot get suppliers to disclose who makes what, or where.
  • Single-source risk hides in the tiers. A part with three distributors can still trace back to one fab or one foundry.
  • Manual search. When a part goes obsolete, teams hunt for alternates and suppliers by hand, part by part.

The result is a reactive process. Programs learn about risk at the moment it becomes a shortage.

What we are building with the Air Force

Muir AI has been awarded a Department of the Air Force SBIR Phase II, "Generative AI for Reverse Engineering and Supply Chain Vulnerabilities." It builds on our Phase I, which showed we could generate multi-tier BoMs and predict country of origin without supplier input.

Phase II takes that further. The work covers three areas:

  1. Part-to-part identification across a generated multi-tier BoM, so every component is matched to a real, identifiable part.
  2. Part-level supplier prediction, moving from where a part likely comes from to who likely makes it.
  3. Counterfeit and obsolescence risk metrics, quantified rather than scored high, medium or low, with an AI agent that explains the reasoning behind each flag.

The goal is to show the Air Force where obsolescence and single-source risk sit on a platform before it turns into a shortage, without relying on suppliers to volunteer the data.

How it works

The same product model that powers Muir's should-cost and tariff analysis does the work here.

  1. Start from what you have. A part number, a drawing, a spec sheet or a partial BoM.
  2. Generate the multi-tier BoM. Muir builds out components, materials and manufacturing processes down the tiers.
  3. Identify parts and suppliers. Each node is matched to real parts and the companies most likely to make them.
  4. Flag the risk. Obsolete, end-of-life, single-source and counterfeit-prone parts are surfaced with the reasoning behind them.
  5. Act on it. Teams see alternates, candidate suppliers and what a part should cost to make, so they can source or redesign before stock runs out.

Not just a defense problem

Muir is dual-use. Most of our customers are commercial manufacturers in electronics, industrial equipment and chemicals, and they face the same problem on a shorter clock.

Electronics is where obsolescence moves fastest. Component lifecycles are short, and a single discontinued part can hold up a product line that took years to qualify. Industrial OEMs supporting installed equipment for 20 or 30 years run into the same gaps the Air Force does.

The capability we are hardening for the Air Force ships in the same platform our commercial customers use today.

What's next

We will share results as the Phase II work progresses. If you sustain long-lived products, or have a list of parts you are worried about, we would like to run them through Muir.

Book a demo to see obsolescence and supplier risk on your own parts.

Blog FAQs

What is part obsolescence?

A part becomes obsolete when its manufacturer stops making it or its supply ends while the product that uses it is still in service. Common causes include end-of-life notices on electronic components, supplier exits, and changes to materials or manufacturing processes.

What is DMSMS?

DMSMS stands for Diminishing Manufacturing Sources and Material Shortages: the loss, or expected loss, of the manufacturers or suppliers of parts and materials. It is the term defense sustainment programs use for obsolescence management.

How do you find obsolescence risk without a complete bill of materials?

Generate the multi-tier BoM from what you have, such as a part number, drawing, spec sheet or partial BoM, then match each component to real parts and their likely suppliers. Risk can then be flagged at every tier, not only for parts already on a list.

Why does single-source risk matter for obsolescence?

A part sold by several distributors can still trace back to one manufacturer, fab or foundry. If that source stops production, every channel runs dry at once, so mapping each part to its actual maker matters more than counting vendors.

What can you do when a part goes obsolete?

The main options are a lifetime buy, an alternate part, a newly qualified supplier, or a redesign. Each costs more the later it starts, so finding at-risk parts early keeps more of these options open.

A line divider

Stay up to date on the latest about Muir

Don't worry, we don't spam.

Other stories

View all stories
PCF
November 2026
As Transparency Demands Rise, Chipmakers Turn to Product-Level Carbon Insight
Continue reading
Right arrow
Hand holding a white envelope against a pale teal background
No items found.
November 2026
How to Respond to a Supplier Price Increase: Test It Against a Should-Cost Model
Continue reading
Right arrow
Isometric technical line drawing of a circuit board with connected components
Should Cost
October 2026
How MCP Accelerates Cost Engineering: From the Parts You Had Time to Study to the Whole BoM
Continue reading
Right arrow

Better decisions
start with clear product intelligence