Stratasys presented the F870 at IMTS 2026 in Chicago as a large industrial FDM platform for tooling, fixtures, inspection gauges, robotic end effectors and full-scale prototypes. Its defining dimension is a 1,000 × 610 × 610 mm build volume. The more consequential proposition, however, is the combination of that envelope with a fully heated chamber, active material conditioning, automatic spool changeover and software intended for multi-shift factory use.

Toyota Production Engineering in Georgetown, Kentucky, and Rivian Automotive in Plymouth, Michigan, are evaluating the platform before commercial availability, according to Stratasys' September 9 announcement. The company says the work covers factory tooling, manufacturing aids and advanced prototyping. These are meaningful launch-customer disclosures, but they are not completed independent durability or return-on-investment studies.

The work the F870 is meant to do

The target is not casual desktop prototyping. Stratasys lists assembly fixtures, alignment jigs, check gauges, drill guides, forming tools, ergonomic lift assists, protective housings and end-of-arm tooling. Such parts are often unique to one station and revised when the vehicle, process or operator feedback changes. Rapid internal iteration can therefore matter more than the unit economics of a mass-produced consumer part.

A one-metre build length can allow a fixture that was previously segmented, bolted or bonded to print as one body. Fewer joints may reduce assembly labour and stack-up error. The same scale also makes shrinkage, flatness and inspection more demanding. A model fitting in the machine is not evidence that every point along a metre-long tool meets the required datum and tolerance.

A 95°C fully heated chamber

The product page specifies a chamber temperature of 95°C. A controlled hot environment is intended to reduce temperature gradients, warping and weak interlayer bonding in large thermoplastic parts, particularly carbon-fibre-filled polyamide tooling. Actual chamber uniformity and part distortion remain application questions that require measurement across representative geometry.

Stratasys publishes XY accuracy of ±0.254 mm or ±0.002 mm per millimetre, whichever is greater. Z accuracy is ±0.200 mm or ±0.002 mm per millimetre, whichever is greater, plus one layer height. The company states that achievable accuracy is geometry- and material-dependent and derives the specification from statistical data at 95 percent dimensional yield. On a metre-long feature, the relative term can imply millimetre-scale permissible deviation, so the headline accuracy must be applied correctly.

Three heads for different production priorities

The N14 head processes ASA and ABS-M30 at 0.18, 0.25 and 0.33 mm layer resolutions, with manufacturer-listed throughput up to 25 cm³/h. The hardened N20H is assigned to abrasive Nylon 12CF at 0.25 mm and up to 21 cm³/h. The N25 high-speed head prints ASA at 0.33 mm and carries a stated maximum of 66 cm³/h.

Stratasys markets the N25 as delivering up to twice the N14 throughput for ASA. That is a head- and material-specific claim, not a promise that every complete part finishes in half the time. Supports, acceleration, minimum layer time, geometry and preparation also affect cycle time. Tool-free head exchange may reduce changeover, but each head and material combination still requires an approved profile and first-article check.

Nine launch material variants

The current list comprises FDM Nylon 12CF, black ABS-M30, six ASA colours and grey FDM ABS Draft, with SR-35 soluble support for the model materials. ASA can suit UV-exposed aids, ABS-M30 general functional models and tools, and Nylon 12CF applications demanding a high stiffness-to-weight ratio. The count of nine includes colour variants rather than nine unrelated polymer families.

This is a controlled material ecosystem. Its advantage is coordination between machine, profile, support and service. Its trade-off is dependence on the supplier's catalogue, spool format and commercial terms. A purchaser should assess regional availability, shelf life, support-material handling and the process for qualification when a formulation changes.

Integrated drying and automatic changeover

The F870 uses two dryer bays and the same 4,100 cm³, or 250 in³, spool design as the Stratasys F3300. Automatic changeover can prevent a long build from stopping at the end of a spool. Regenerative drying is intended to control the state of moisture-sensitive polyamide. Stratasys says the bays can actively dry wet material rather than merely preserve a dry spool.

The trade-show report cites six to ten hours for drying a wet spool, but this should not be turned into a universal cycle. Starting moisture, material, bay loading and the acceptance threshold matter. The published claim of up to two weeks of unattended production likewise assumes high-capacity spools and successful automatic changeover; it does not mean every geometry and facility can be left without a risk process.

What two weeks unattended really means

The product page says the material system can support up to two weeks of continuous printing without intervention for material changes. In an operating factory, unattended production additionally depends on fire policy, access controls, maintenance state, remote alert handling and quality procedures. Mobile monitoring helps observe status; it does not automatically accept the manufactured part.

The platform lists MTConnect, an emergency stop, GrabCAD Print and Insight software, and mobile monitoring. MTConnect can provide manufacturing data to plant systems, but buyers need to establish which data are exposed, how users are authenticated, how records are retained and how the printer is segmented on the network. Connectivity is not the same as validated traceability.

Toyota, Rivian and the automotive examples

The official release describes Toyota and Rivian as early evaluators ahead of commercial availability. The on-site report adds a 42-inch-wide Nylon 12CF robotic end-of-arm tool and a Toyota cobot gripper interface reportedly printed in 20 hours. These examples demonstrate intended use, but the sources do not publish a full comparison with the previous metal tools' production time, mass, service life and audited cost.

Lower end-effector mass can increase available robot payload or reduce dynamic loads. A polymer composite differs from aluminium or steel in thermal expansion, conductivity, creep and attachment behaviour; inserts may be necessary at fasteners. Impact, heat, chemicals and repeated cycles therefore require an application-specific risk assessment and inspection interval.

A printed inspection gauge introduces another concern: it may be quick to replace, but its dimensional state must remain linked to calibration and wear controls. A tool that touches painted surfaces may need a particular finish; a drill guide may require a metal bushing; an ergonomic aid may need rounded, cleanable contact surfaces. The printer supplies geometry, not automatic fitness for every station.

Price and installation are only the entry costs

The trade-show report gives a US starting price of $149,000. That is not a delivered European or Hungarian price and does not necessarily include freight, installation, training, service, software, initial materials or infrastructure. Stratasys lists three-phase 208 V, 30 A power, double-door access, forklift handling for placement and 99 cm of clearance on each side.

A total-cost model should compare the real tools being displaced. It includes design, build time, soluble-support removal, inserts, measurement, scrap, downtime, maintenance and material. CNC-machined or welded metal is not always more expensive: high load, heat, abrasion and long service may still favour metal. The F870 case is strongest for large, frequently changing, low-volume aids where lead time has economic value.

Shared spools with the F3300 can simplify inventory for a site already using that platform, but create no such saving for a new fleet. Soluble support also adds washing or removal workflow, fluid handling and drying considerations. Those activities belong in labour, facility and environmental assessment rather than being hidden outside the printer purchase.

What a procurement trial should prove

A buyer should submit its own representative fixture rather than rely on a trade-show sample. The trial should record build time, model and support consumption, active labour, critical dimensions, surface, mass and functional load. Large parts should be measured at several locations and across repeat builds; repeatability matters more than one selected success.

The test should exercise dryer recovery, automatic spool changeover, weekend fault handling, reproducibility after a head change and software permissions. A Nylon 12CF tool belongs on the actual robot or station under expected cycles and environment. Supplier data is a selection basis, not an application approval.

Acceptance criteria need to be written before the build. Define datums, maximum distortion, insert pull-out requirement, permitted finishing, measurement method and replacement interval. Without them, a lighter and visually impressive tool can be called successful even when it fails the process that justified the investment.

Verified facts, claims and editorial assessment

It is verified that Stratasys announced the F870 on September 9, 2026 and exhibited it at IMTS in Chicago from September 14 to 19. Official specifications list a 1,000 × 610 × 610 mm envelope, 95°C chamber, three N-series heads and nine launch material variants. Toyota and Rivian evaluation is also disclosed by Stratasys.

Two-times throughput, two weeks unattended, shorter lead times and lower total ownership cost remain manufacturer or customer claims within the public material. They are plausible operational goals, but the sources provide no complete independent comparison across competing systems. LAYERTIMES' assessment is that the F870's most important feature is not size alone but the integration of a long heated chamber with material conditioning and operational automation.

The machine may be valuable where a manufacturer has a recurring family of large, frequently revised, moderately loaded tools and an engineering team capable of exploiting fast iteration. Where demand is rare, polymer is unsuitable, or no process owner and inspection system exist, a large printer can become underused capital. Procurement should therefore begin with qualified part families and a total-cost model, not build volume alone.