Velo3D: Metal Additive Manufacturing and the Systems Engineering of the Machine That Makes the Parts

Metal additive manufacturing for aerospace flight hardware is one of the most demanding contexts in which any engineering organization can operate. The physics are difficult. The certification landscape is fragmented. The customers — defense primes, launch vehicle manufacturers, gas turbine OEMs — operate under regulatory and contractual frameworks that were written for processes that look nothing like laser powder bed fusion.

Velo3D sits in the middle of that environment and has chosen to compete precisely where the difficulty is highest: printing metal parts for flight-critical applications using their Sapphire platform. Understanding how they engineer the systems that make those parts reveals something important about what systems engineering actually looks like when the product is a manufacturing process, not a vehicle or a weapon.

What Velo3D Actually Builds

The Sapphire printer is a laser powder bed fusion (LPBF) system capable of processing nickel superalloys, titanium alloys, and tool steels at high precision and, critically, with very low support structure requirements. That last point is not a marketing claim — it is an engineering differentiation that matters for flight hardware. Reducing support structures reduces post-processing, reduces material removal, and reduces the number of process steps that can introduce variation. For turbine components and propulsion hardware where internal geometries are complex and tolerances are tight, this matters enormously.

But Velo3D does not simply sell machines. They sell what they call a “supported manufacturing solution” — the printer, the print management software (Flow), the process parameters, and a quality framework that customers can leverage when seeking their own qualification approvals. This vertical integration is deliberate, and it shapes every dimension of how Velo3D’s engineering organization is structured.

The Coupled Problem: Machine Requirements and Part Requirements

In conventional subtractive manufacturing, the machine requirements and the part requirements are largely decoupled. A CNC machining center has specifications — spindle speed range, linear positioning repeatability, axis travel — and as long as the machine operates within those specifications, the part quality is governed by the process program and the operator’s decisions. The machine vendor and the part manufacturer can maintain separate quality systems that intersect only at the machine acceptance test.

Additive manufacturing breaks that clean separation. In LPBF, the part properties — porosity, microstructure, residual stress, surface finish — are direct functions of the laser power, scan speed, hatch spacing, layer thickness, and atmosphere control that the machine delivers. Change the machine firmware, and you potentially change the part. Replace a laser module, and the part qualification data generated on the previous configuration may no longer be applicable. The machine is not a tool that removes material to a drawing; it is a process system whose outputs are inseparable from its current state.

This creates what is, from a systems engineering standpoint, a requirements coupling problem of unusual severity. Velo3D has to maintain requirements on the machine system that are defined, at least in part, by what parts the machine is expected to produce. A customer printing Inconel 718 turbine blades for a turbofan application is placing requirements on machine laser power stability, atmosphere oxygen content, and powder delivery consistency that are derived from the metallurgical requirements on the part. Those requirements flow backward through the interface between the customer’s part specification and Velo3D’s machine specification.

Managing that interface rigorously — not as a contractual afterthought but as a living technical relationship — is one of the central systems engineering challenges Velo3D faces.

AS9100 and the Quality System Around a Process

AS9100 is the aerospace quality management system standard, based on ISO 9001 but with additional requirements tailored to the safety and reliability demands of aviation, space, and defense. Achieving and maintaining AS9100 certification for a manufacturing process like LPBF requires demonstrating not just that you have documented procedures, but that those procedures are effective, that deviations are controlled, and that the product is traceable to its production records.

For Velo3D, AS9100 compliance applies at two levels. First, Velo3D itself operates under AS9100 as a manufacturer of the Sapphire systems. Second, customers who use Sapphire systems to produce flight hardware must have their own AS9100-compliant processes, and those processes depend on the machine performing consistently with its specifications.

The documentation burden is significant. Every machine configuration — software version, optical assembly state, powder delivery calibration — must be documented such that a quality auditor can determine whether a given set of printed parts was produced on a qualifying machine configuration. For Velo3D’s engineering team, this means that change management is not just a development discipline; it is a quality obligation. A firmware update that improves print performance is also a potential qualification event that requires careful assessment against the machine’s approved configuration baseline.

NADCAP adds a further layer. The National Aerospace and Defense Contractors Accreditation Program provides third-party accreditation for special processes, and additive manufacturing is now within its scope. NADCAP accreditation for additive manufacturing requires demonstration of process control at a level of detail that is unusual even by aerospace standards. The auditors are specifically interested in how process parameters are established, validated, and locked down — and in what happens when a parameter drifts or a machine component is replaced.

For Velo3D, this means that the process parameters embedded in their Flow software are not just operational settings. They are controlled configuration items with qualification data attached. The engineering organization must maintain traceability from a specific set of process parameters to the mechanical property datasets that justify those parameters for a given alloy and application. That traceability chain — from the part requirement, through the process parameters, to the machine specification, to the qualification test data — is what NADCAP auditors are looking for, and what Velo3D’s systems engineering infrastructure must support.

The Interface Control Problem

Velo3D’s customers bring part files, material requirements, and often their own process requirements derived from their own engineering organizations or from their regulatory submissions. The interface between what the customer specifies and what Velo3D’s machine can deliver is complex and not fully standardized.

This is, in the language of systems engineering, an interface control problem. In aircraft development, interface control documents (ICDs) govern the physical and functional interfaces between systems from different suppliers. The equivalent in additive manufacturing is still evolving. There is no industry-standard ICD format for the interface between a part designer’s requirements and an LPBF system’s process capabilities. Velo3D has had to develop their own approach.

The practical implication is that Velo3D’s applications engineering function — the team that works with customers to translate part requirements into validated print parameters — is doing systems engineering work whether or not it is labeled as such. They are capturing requirements, identifying constraints, resolving conflicts between what the customer needs and what the machine can deliver, and producing a process specification that both parties can agree on and that can be defended in a quality audit.

This work is difficult to scale. Every new customer, every new part geometry, every new material, potentially reopens aspects of the interface definition. Aerospace customers have long design cycles and complex approval chains; a change to the print parameters that improves yield may require months of re-qualification work before it can be implemented in production. Velo3D’s ability to manage this complexity is a direct function of how well their engineering organization captures and maintains requirements across the machine-to-part interface.

Configuration Management at Machine Scale

One of the less-discussed challenges of Velo3D’s position is the configuration management complexity of a deployed fleet of Sapphire systems. As machines are installed at customer sites, they accumulate service history, component replacements, and software updates at different rates. A machine at a customer facility that was installed eighteen months ago may be running a different firmware version and may have had its optical assembly replaced once since installation.

From a qualification standpoint, this creates a potentially fragmented landscape. The qualification data that the customer relied on to certify their parts was generated on a specific machine configuration. If that configuration has changed — even through a sanctioned maintenance event — the customer and Velo3D together must assess whether the change is qualification-neutral or whether re-validation is required.

This is configuration management in its most operationally serious form. The stakes are not just functional — a misconfigured machine that produces structurally deficient flight hardware represents a potential safety event. Velo3D’s engineering processes must make the current configuration of each deployed machine legible to both internal quality teams and to customer quality engineers at any given moment.

This is also where requirements management at the machine level connects directly to customer part certification. A customer submitting data to the FAA or to a defense program office is representing that their parts were produced on a qualified process. If Velo3D cannot provide documentation that the machine used to produce those parts was in the qualified configuration at the time of production, the certification case becomes difficult to defend.

What the Qualification Gap Actually Looks Like

The additive manufacturing qualification gap is real, but its nature is often mischaracterized. It is not primarily a question of whether LPBF can produce high-quality parts — it demonstrably can. The gap is procedural and evidentiary. The existing qualification frameworks for aerospace manufacturing were developed around processes with different failure modes and different variability signatures than additive.

For Velo3D, closing the qualification gap means building an evidence base that satisfies conservative regulatory and customer expectations, not just demonstrating technical capability. This involves:

Process capability data at volume. Qualification requires statistical confidence in property distributions, which requires large sample sets. Velo3D’s investment in generating and maintaining this data — across alloys, geometries, and machine configurations — is a long-term engineering program, not a one-time test campaign.

Equivalency arguments for machine changes. When a laser module is replaced, or a new software version is released, Velo3D needs an engineering methodology for arguing that parts produced on the updated configuration are equivalent to those produced on the qualified configuration. This is non-trivial. It requires both physical testing and a documented analytical approach that regulators and customers will accept.

Customer-side qualification support. Velo3D’s customers often need to take Velo3D’s process data and use it as supporting evidence in their own certification submissions. This means Velo3D’s documentation has to meet standards set by someone else’s quality organization, not just Velo3D’s own internal requirements. The content and format of qualification packages must be usable by customers navigating their own AS9100 audits and, in some cases, FAA or DoD review.

The Operational Reality

Velo3D’s engineering organization is doing something that most capital equipment companies are not: they are maintaining technical accountability for the downstream output of their machines in ways that are normally reserved for the part manufacturer. This is not a marketing positioning choice. It follows from the physics and the qualification framework.

In practice, this means that Velo3D’s systems engineering function has to hold requirements at multiple levels simultaneously — machine system requirements, process parameter requirements, and the interface requirements that connect the machine to the customer’s part. These levels interact. A requirement on layer thickness derived from part surface finish requirements has implications for the powder delivery system specification. A requirement on atmosphere oxygen content derived from titanium oxidation limits has implications for inert gas handling system design.

The engineering teams that do this work well will be those that maintain live, navigable models of these requirement relationships rather than static documents that reflect the system as it was understood at a point in time. The volume of interactions — across materials, machine configurations, customer part families, and qualification states — makes document-based approaches progressively untenable as the machine fleet grows and the customer base diversifies.

Honest Assessment

Velo3D has made a coherent technical and business bet: that the aerospace and defense market will pay for the combination of low-support-structure printing capability, a qualified process framework, and the vertical integration that makes qualification tractable. The engineering organization that supports this bet is operating at the intersection of capital equipment development, process engineering, and aerospace quality — a combination that demands unusual depth in both the physics and the procedural.

The qualification landscape for additive manufacturing is still maturing. NADCAP’s additive manufacturing checklist is relatively new. The FAA’s approach to additively manufactured flight hardware is still evolving. Velo3D is navigating a regulatory environment that does not yet have stable, widely accepted answers for some of the hardest questions about process equivalency and machine configuration control.

What distinguishes the organizations that navigate this well from those that struggle is the rigor of their requirements management — specifically, their ability to maintain coherent traceability from the mechanical property requirement on a flight part backward through the process parameters, through the machine specification, to the qualification evidence. That chain, maintained and auditable, is what turns a technically capable printing system into a defensible aerospace manufacturing process.

Velo3D’s position in the market depends on making that chain reliable for their customers. The engineering organization’s job is to make it real.