# Manufacturing-as-a-Service / Instant-Quote Platforms: Commoditization vs. Relationship

**Research question:** Xometry, Fictiv, Protolabs+Hubs, Paperless Parts, Partsimony, CADDi, RapidDirect, and PCBWay/JLCPCB all attack the "get a quote fast" problem. Do they intermediate (own the demand relationship, insert themselves as principal) or disintermediate (empower direct shop-buyer relationships)? Is there room for a matchmaker that does the opposite of Xometry - build warm, durable relationships instead of commoditizing capacity?

## Summary verdict up front

Every platform in this set that runs a two-sided marketplace with itself as principal (Xometry, and to a lesser extent Fictiv, RapidDirect) is structurally incentivized to commoditize the supply side, because its margin comes from the spread between what it charges the buyer and what it pays the shop, and that spread is easiest to protect by keeping shops interchangeable and buyers ignorant of which shop actually made their part. The platforms that avoid this are the ones that structurally cannot commoditize: Paperless Parts and CADDi Drawer sell software to the shop or the OEM and never touch the transaction, so they have no incentive to anonymize anyone. That split is direct evidence for Phil Shatkin's thesis: there is real product-market room for a matchmaker that behaves like Paperless Parts (tool-not-broker) or old-school MFG.com/ThomasNet lead-gen (introduction-not-transaction), rather than like Xometry (broker-as-principal). The catch is that this is a much smaller, services-adjacent business than a marketplace, not a bigger one - see Risks below.

## Platform-by-platform

### Xometry (brief, per assignment - deep coverage owned by another researcher)
Public company (NASDAQ: XMTR). FY2025 revenue $686.6 million; marketplace active buyers grew 20% YoY to 81,821 and active suppliers grew 17% YoY to 4,996 as of Dec 31, 2025 ([Xometry Q4/FY2025 release](https://investors.xometry.com/news-releases/news-release-details/xometry-reports-record-fourth-quarter-and-strong-full-year-2025)). Xometry is principal in every transaction: buyer pays Xometry, Xometry pays the shop net of its cut. Current effective take is roughly 20% of GMV with a stated long-term target of 30-35% gross margin ([Bowery Capital S-1 teardown](https://bowerycap.com/blog/insights/s-1-teardown-xometry); [industrialanalyst.substack.com pricing analysis](https://industrialanalyst.substack.com/p/2q24-pricing-analysis-protolabs-vs)). Xometry acquired Thomas/ThomasNet in December 2021 and now routes ThomasNet's 2.9 million inbound links into its own quoting funnel, converting what used to be a neutral directory into a lead-funnel for its marketplace - exactly the dynamic that makes Phil's "directories have the data but not the warm-intro feeling" complaint literal, not metaphorical ([Cottrill Research](https://cottrillresearch.com/xometry-acquires-thomas-these-two-companies-originating-from-different-beginnings-face-fierce-competition/); [SaveAmericanManufacturers.com](https://www.saveamericanmanufacturers.com/), an anti-Xometry advocacy site run by aggrieved former ThomasNet-adjacent shops, useful as a sentiment artifact even though it is not neutral).

### Fictiv
Raised $192M total (Series E was $100M in May 2022) before being acquired by Japanese components distributor MISUMI Group for $350 million all-cash, announced April 17, 2025, closed June 18, 2025 ([PlasticsToday](https://www.plasticstoday.com/business/misumi-to-acquire-fictiv-in-350-million-deal); [MISUMI press release](https://www.misumi.co.jp/english/news/press_250417)). That is a mediocre outcome relative to capital raised (roughly 1.8x on $192M with no disclosed profitability), consistent with the broader thesis that pure on-demand-manufacturing marketplaces struggle to hit venture-scale returns. Third-party revenue estimates (RocketReach, Growjo - not verified, treat as rough) put Fictiv around $60-70M/year at time of sale, small for the capital raised.

Structurally, Fictiv is a hybrid: it is still principal (buyer pays Fictiv, Fictiv pays the vetted manufacturing partner), but it wraps the transaction in a dedicated program manager and DFM engineering support, explicitly marketed against Xometry's "rideshare-style capacity matching" ([Fictiv vs Xometry](https://www.fictiv.com/xometry-competitor-alternative); [RapidDirect comparison](https://www.rapiddirect.com/blog/xometry-vs-fictiv/)). This buys Fictiv better buyer sentiment (Trustpilot commentary is generally positive on communication and program management) but it does not give the shop the customer relationship - Fictiv, not the shop, owns the account. Shops in Fictiv's network are still fungible capacity from the shop's point of view; the "relationship" Fictiv sells is between itself and the buyer, not between the buyer and a specific shop.

### Protolabs + Hubs (3D Hubs)
Protolabs is fundamentally different in kind from the others: it owns its own automated factories (injection molding, CNC, sheet metal) rather than brokering third-party shops, so its instant quote is a real capacity-planning quote, not a broker's guess. Protolabs acquired 3D Hubs (a 240-partner distributed manufacturing network, ~$25M 2020 revenue, >200% CAGR since 2017) for $280 million in January 2021 - $130M cash plus $150M stock plus up to $50M in earnouts - explicitly to bolt a broker network onto its owned-factory core and cover capabilities/volumes outside Protolabs' own equipment ([3D Printing Industry](https://3dprintingindustry.com/news/protolabs-acquires-3d-hubs-for-280m-to-expand-manufacturing-partner-network-182844/); [Businesswire](https://www.businesswire.com/news/home/20210119005245/en/Protolabs-Reaches-Agreement-to-Acquire-3D-Hubs-Creates-the-Worlds-Broadest-Digital-Manufacturing-Offer-for-Custom-Parts)). Protolabs later rebranded the Hubs network as "Protolabs Network" (Jan 2024, [Digital Commerce 360](https://www.digitalcommerce360.com/2024/01/12/protolabs-rebrands-its-digital-manufacturing-network/)), folding the broker layer fully into the Protolabs brand. Net effect: even the "owns its own factories" player concluded it needed a broker network for volume/capability breadth, and that broker layer commoditizes partner shops the same way Xometry's does - Protolabs Network shops are anonymous capacity behind the Protolabs brand, not named suppliers building buyer relationships.

### Paperless Parts
Structurally the cleanest disintermediation case in the set. Paperless Parts is B2B SaaS sold to the shop (quoting/CPQ software: geometry-driven part analysis, automated pricing/costing, CRM), not a marketplace - the shop's customer relationship, pricing, and brand stay entirely the shop's own. $30M Series B led by OpenView Partners (announced Sept 2021); over 800 manufacturing customers ([Businesswire](https://www.businesswire.com/news/home/20210913005163/en/Paperless-Parts-Announces-30M-Series-B-Funding-Led-by-OpenView-Partners)). Quoting-speed value prop is concrete: single components quoted in ~15 minutes vs. up to 2 hours manually, complex assemblies in ~4 hours vs. up to 4 days ([Paperless Parts](https://www.paperlessparts.com/)). Practical Machinist forum threads on Paperless Parts ("What is the deal with Paperless Parts?") exist but the forum blocks scraping (403); worth a manual read given shop owners there are candid about vendor tools. The business-model takeaway stands regardless: Paperless Parts proves shops will pay real SaaS dollars ($30M raised, hundreds of paying shops) for tools that make them faster without forcing them to compete anonymously on a marketplace. This is the closest existing analog to an "anti-Xometry" wedge, just aimed at the shop's internal workflow rather than at buyer-shop matching.

### Partsimony
Tiny by comparison: $2.27M total raised, a $2M seed closed October 2021 led by Closed Loop Partners' Ventures Group with Contour Ventures, Urban Us, Morgan Stanley Inclusive Ventures Lab, SAP.iO and others ([Closed Loop Partners](https://closedlooppartners.com/partsimony-closes-2m-seed-round-to-help-organizations-build-intelligent-manufacturing-supply-chains); [Crunchbase](https://www.crunchbase.com/organization/partsimony)). Positions itself as a "SaaS network" unifying disparate supply-chain data from prototype through production for hardware teams in automotive, aerospace, robotics, medical devices - a data/PLM layer, not a bidding marketplace. Founded 2015 (Richard and Roland Mokuolu), Brooklyn-based. The funding scale (barely $2M seed six years post-founding) suggests either a very long slow build or a thesis that hasn't found its wedge yet; worth treating as a cautionary data point on how hard the "unify hardware supply chain data" framing is to scale, not as a proof point.

### CADDi
The most instructive comparable for the anti-Xometry thesis because CADDi explicitly started as a Xometry clone and pivoted away from spot-bidding. Founded 2017 in Japan; TechCrunch reported CADDi "initially [was] a direct competitor to manufacturing service network Xometry, quoting parts for customers and sourcing work through a network of manufacturing partners" ([TechCrunch, Series C](https://techcrunch.com/2023/07/05/caddi-raises-89m-series-c-to-scale-its-b2b-supply-chain-marketplace-for-manufacturing-parts/)). It moved to a managed-capacity model: CADDi pre-negotiates pricing and reserves capacity across 600+ vetted suppliers, and its AI matches drawings to the best-fit supplier under those pre-set terms rather than running a live reverse auction - CADDi, not the buyer, stays accountable for cost/quality/delivery ([WiL blog](https://wilab.com/blog/cad-di-catalyzing-the-digital-transformation-dx-in-manufacturing)). CADDi captures ~70% of Japanese manufacturers in its network and has been expanding to the US (HQ now dual Tokyo/Chicago). Funding: $73M Series B (Aug 2021, TechCrunch), $89M Series C (July 2023, total to $164M), $38M Series C extension led by Atomico (Mar 2025) bringing the round to raise a Chicago expansion ([TechCrunch Series B](https://techcrunch.com/2021/08/23/japans-b2b-ordering-and-supply-platform-caddi-raises-73-million-series-b-funding/); [SiliconANGLE](https://siliconangle.com/2025/03/27/ai-startup-caddi-nabs-38m-help-manufacturers-optimize-supply-chains/)). Separately, "CADDi Drawer" (launched 2022) is a pure SaaS product sold to OEMs/manufacturers that makes historical 2D drawing libraries AI-searchable - again, a tool sold to one side, not a transaction broker. CADDi is the strongest existing evidence that a marketplace founder can look at Xometry's spot-auction dynamics, conclude they destroy trust and margin for suppliers, and deliberately re-architect toward pre-negotiated, durable supplier relationships instead - and still raise venture money doing it ($164M+ total).

### RapidDirect
Shenzhen-based, founded 2009, operates its own 5,000 m² factory (200+ machines) plus a broader partner network; launched an "Instant Quote" platform (June 2021) that parses CAD uploads and auto-quotes with a claimed 82% success rate on drawing recognition/pricing ([PR Newswire](https://www.prnewswire.com/news-releases/rapiddirect-announces-new-instant-quote-platform-radically-driving-transformation-in-prototype-manufacturing-301306713.html)). No funding rounds were found in public sources - it reads as a bootstrapped/founder-owned factory-plus-platform business, similar in spirit to Protolabs' original model but China-based and cheaper. Because RapidDirect owns real capacity, it is not commoditizing third-party shops the way Xometry does; it commoditizes labor/overhead inside its own factory instead. Not a strong data point either way for the relationship question - it is vertically integrated, so "relationship" is moot; the buyer's relationship is with RapidDirect itself, end to end.

### PCBWay / JLCPCB
Different category (PCB fabrication, not general job-shop CNC/sheet metal/injection molding), but instructive by contrast. Both are themselves factories (not marketplaces brokering third-party shops): JLCPCB built its business on high-volume, highly automated, low-cost production and instant online quoting; PCBWay differentiates on delivery reliability and a broader one-stop menu (assembly, 3D printing, CNC) ([Medium/RAYPCB comparison](https://medium.com/@raypcb/whats-the-difference-between-jlcpcb-and-pcbway-in-business-model-and-price-by-raypcb-acf501a8a752); PCBWay estimated $15-25M annual revenue per [ZoomInfo](https://www.zoominfo.com/c/rapiddirect/470540610)-class estimators, weak confidence). Because they are factory-direct rather than broker-of-record, there is no third-party shop being commoditized - the "instant quote" commodity here is PCB fab capacity itself, which is genuinely commodity-like (PCBs are far more standardized than custom CNC/sheet-metal parts). This is a useful negative case: instant-quote works cleanly when the underlying capability is standardized (PCBs, sheet metal to a lesser extent) and works badly when it isn't (complex CNC, casting, multi-process assemblies) - which is exactly where warm-intro matchmaking should add the most value, per Phil's BOM-upload framing.

## Take rates and shop-side sentiment (the crux)

Verified take-rate/margin data: Xometry's gross margin (its effective take, since it is principal) rose from 26% to 39% within one year per shop-owner discussion on Practical Machinist quoting Xometry's own quarterly reporting, alongside a November 1, 2024 switch to net-40 supplier payment terms (from previously faster terms), which shop owners on the same forum characterized as another margin squeeze and a sign of market-power consolidation ("[Xometry's 'take' grows to 39% last quarter](https://www.practicalmachinist.com/forum/threads/xometrys-take-grows-to-39-last-quarter.406671/)"; "[xometry net40 change and possible actions](https://www.practicalmachinist.com/forum/threads/xometry-net40-change-and-possible-actions.432746/)"). Other verified shop-side complaints surfaced in that community: job offers priced so low that "shops can barely purchase materials for what Xometry wants to pay for the final product," and the view that Xometry's 2020 acquisition of MakeTime (previously seen as a better-paying alternative for shops) folded MakeTime's suppliers into Xometry's lower-paying pricing, not the reverse. One shop owner's framing, paraphrased from the forum discussion, captures the sentiment: they do not want to compete in a market against a company happy to lose money every year to buy share. This is a direct "race to the bottom" complaint from the supply side, and it is structural, not incidental: Xometry is a public company under quarterly growth pressure, and GMV growth is easiest to buy by expanding the addressable supplier pool (cheaper, more desperate shops) and squeezing the spread, both of which look identical to a shop as "prices going down and terms getting worse."

Buyer-side sentiment is more mixed than shop-side. Public reviews (Trustpilot, SmallBizGenius aggregation) skew positive on speed and breadth, but a specific and recurring complaint is pricing/quality inconsistency across repeat orders because the buyer has no guarantee of getting the same shop twice - one quoted example: "Second order, identical to first (two months prior), was nearly double the price," which is a direct symptom of the anonymized-shop-pool model (no continuity means no learning curve, no trust, no negotiated repeat-order pricing). This is the single most important buyer-side data point for Phil's thesis: buyers who like a shop's work on order one have no lever to get that same shop on order two through Xometry. A relationship-preserving matchmaker's entire value proposition is closing exactly that gap.

## Answering the key question: is there room for an anti-Xometry matchmaker?

Yes, directionally, but the existing evidence points to a narrower and more services-like business than "the next Xometry," for three reasons grounded in the research above:

1. **The commoditizing model already visibly damages the trust it needs.** Xometry's own supply base is audibly unhappy (margin squeeze, net40, MakeTime bait-and-switch), and buyers feel the second-order effect (repeat-order inconsistency) even when they don't understand the cause. That is a real wedge: "get matched with the same 3-5 vetted shops who get to know your parts" is a coherent, differentiated pitch against "get matched with whoever bid lowest today."

2. **CADDi is the existence proof, and it is telling that it moved away from spot bidding, not toward it.** A well-funded ($164M+) team that started as literal Xometry clone concluded pre-negotiated, durable-relationship capacity was the better architecture and raised more money doing that. That is evidence the market rewards relationship-preserving models, but also evidence that doing it well requires real infrastructure (600+ vetted suppliers, pre-negotiated terms, AI matching) - i.e., it's still a hard, capital-intensive business, not a thin agent layer on top of ThomasNet data.

3. **Paperless Parts shows the cheaper, faster path to "anti-Xometry" is not matchmaking at all, it's arming the shop.** Paperless Parts never touches the transaction and still built a venture-scale ($30M+ raised, 800+ shops) business by making the shop itself faster and better at quoting, leaving the customer relationship exactly where it already was. If the goal is "shops keep their customer relationships," the fastest-validated wedge in this data set is tools-for-shops, not introductions-between-strangers.

The implication for Gabriel's "Boardy for supply chain" framing: an agent that ingests a buyer's BOM and produces 50 warm-introduced, curated suppliers is closer in spirit to CADDi's original pitch than to a truly new category, and the data suggests the winning version of that pitch evolves toward "pre-vetted, durable panel of suppliers with pre-negotiated terms" (CADDi's actual model) rather than staying a one-shot introduction engine. A pure matchmaking layer that never becomes sticky infrastructure (repeat matching, standing relationships, some transaction or quoting workflow it owns) risks becoming a one-time lead-gen tool that shops and buyers route around once they've met each other - the same fate that made ThomasNet a directory Xometry had to buy rather than compete with. The defensible version needs a reason for buyers and shops to keep coming back through the matchmaker after the warm intro has already happened.

## Risks / open questions

- Fictiv's $350M sale against $192M raised, and Partsimony's tiny $2.27M total after a decade, are two structural warnings: manufacturing marketplaces/data layers are hard to scale to venture-return outcomes even when well-executed and well-funded. A relationship-first matchmaker needs a clear answer for why it avoids both fates.
- The strongest "anti-commoditization" data comes from Practical Machinist forum threads that could not be directly fetched (403 blocked); the specific quotes above are drawn from search-result summaries of those threads, not the raw thread text, and should be verified by a human visit to the threads before being cited externally.
- Xometry's supplier-side unhappiness has not yet visibly hurt its buyer-side growth (81,821 active buyers, 20% YoY growth in FY2025) - the commoditization pain is real but has not (yet) become a growth-limiting factor for the incumbent. A wedge strategy should not assume Xometry is vulnerable on this axis in the near term; it should assume this is an underserved segment (buyers/shops who want relationship continuity) rather than a mass defection opportunity.

## Sources

- [Xometry Reports Record Fourth Quarter and Strong Full Year 2025 Results](https://investors.xometry.com/news-releases/news-release-details/xometry-reports-record-fourth-quarter-and-strong-full-year-2025)
- [Xometry Q1 revenue rises on 40% marketplace growth - Digital Commerce 360](https://www.digitalcommerce360.com/article/xometry-marketplace-sales-revenue/)
- [S-1 Teardown: Xometry - Bowery Capital](https://bowerycap.com/blog/insights/s-1-teardown-xometry)
- [2Q24 pricing analysis Protolabs vs Xometry - industrialanalyst.substack.com](https://industrialanalyst.substack.com/p/2q24-pricing-analysis-protolabs-vs)
- [Xometry acquires Thomas - Cottrill Research](https://cottrillresearch.com/xometry-acquires-thomas-these-two-companies-originating-from-different-beginnings-face-fierce-competition/)
- [Thomasnet vs Xometry - SaveAmericanManufacturers.com](https://www.saveamericanmanufacturers.com/)
- [Misumi to Acquire Fictiv in $350 Million Deal - PlasticsToday](https://www.plasticstoday.com/business/misumi-to-acquire-fictiv-in-350-million-deal)
- [MISUMI Notice Regarding the Acquisition of Fictiv Inc.](https://www.misumi.co.jp/english/news/press_250417)
- [Fictiv Vs. Xometry - Fictiv](https://www.fictiv.com/xometry-competitor-alternative)
- [Xometry vs Fictiv - RapidDirect](https://www.rapiddirect.com/blog/xometry-vs-fictiv/)
- [Protolabs Reaches Agreement to Acquire 3D Hubs - Businesswire](https://www.businesswire.com/news/home/20210119005245/en/Protolabs-Reaches-Agreement-to-Acquire-3D-Hubs-Creates-the-Worlds-Broadest-Digital-Manufacturing-Offer-for-Custom-Parts)
- [Protolabs acquires 3D Hubs for $280M - 3D Printing Industry](https://3dprintingindustry.com/news/protolabs-acquires-3d-hubs-for-280m-to-expand-manufacturing-partner-network-182844/)
- [Protolabs rebrands its digital manufacturing network - Digital Commerce 360](https://www.digitalcommerce360.com/2024/01/12/protolabs-rebrands-its-digital-manufacturing-network/)
- [Paperless Parts Announces $30M Series B - Businesswire](https://www.businesswire.com/news/home/20210913005163/en/Paperless-Parts-Announces-30M-Series-B-Funding-Led-by-OpenView-Partners)
- [Paperless Parts homepage](https://www.paperlessparts.com/)
- [What is the deal with Paperless Parts? - Practical Machinist](https://www.practicalmachinist.com/forum/threads/what-is-the-deal-with-paperless-parts.431259/)
- [Partsimony Closes $2M Seed Round - Closed Loop Partners](https://closedlooppartners.com/partsimony-closes-2m-seed-round-to-help-organizations-build-intelligent-manufacturing-supply-chains)
- [Partsimony - Crunchbase](https://www.crunchbase.com/organization/partsimony)
- [CADDi raises $89M Series C - TechCrunch](https://techcrunch.com/2023/07/05/caddi-raises-89m-series-c-to-scale-its-b2b-supply-chain-marketplace-for-manufacturing-parts/)
- [Japan's B2B ordering and supply platform CADDi raises $73M Series B - TechCrunch](https://techcrunch.com/2021/08/23/japans-b2b-ordering-and-supply-platform-caddi-raises-73-million-series-b-funding/)
- [AI startup CADDi nabs $38M - SiliconANGLE](https://siliconangle.com/2025/03/27/ai-startup-caddi-nabs-38m-help-manufacturers-optimize-supply-chains/)
- [CADDi Catalyzing the Digital Transformation in Manufacturing - WiL](https://wilab.com/blog/cad-di-catalyzing-the-digital-transformation-dx-in-manufacturing)
- [RapidDirect Announces New Instant Quote Platform - PR Newswire](https://www.prnewswire.com/news-releases/rapiddirect-announces-new-instant-quote-platform-radically-driving-transformation-in-prototype-manufacturing-301306713.html)
- [What's the Difference Between JLCPCB and PCBWay - Medium/RAYPCB](https://medium.com/@raypcb/whats-the-difference-between-jlcpcb-and-pcbway-in-business-model-and-price-by-raypcb-acf501a8a752)
- [Xometry's "take" grows to 39% last quarter - Practical Machinist](https://www.practicalmachinist.com/forum/threads/xometrys-take-grows-to-39-last-quarter.406671/)
- [Xometry net40 change and possible actions - Practical Machinist](https://www.practicalmachinist.com/forum/threads/xometry-net40-change-and-possible-actions.432746/)
- [Xometry Prices - Practical Machinist](https://www.practicalmachinist.com/forum/threads/xometry-prices.411079/)
- [MFG.com - Wikipedia](https://en.wikipedia.org/wiki/MFG.com)
- [Xometry Reviews - Trustpilot](https://www.trustpilot.com/review/xometry.com)