Biohybrid robotics: sourcing living actuators

Nobody sells a living actuator. One supplier does now sell finished contractile muscle tissue — myriamed lists myrTissue-Skeletal at EUR 1,400 for one tissue, ambient shipping, certificate of analysis per batch, 7–11 weeks — but a tissue is not an actuator, and myriamed publishes no force acceptance criterion for it. So if you are building a biohybrid device you will still assemble it from a cell source or a bought tissue, a mould, several weeks of culture and a force measurement, and you will do the integration yourself or pay a contract lab to do the parts of it they are willing to do. What changed in 2026 is that the make-versus-buy comparison finally has a number on both sides; what has not changed is that nothing is sold against a force spec.

The published side is much healthier than the commercial side. Skeletal-muscle-powered walkers, swimmers, pumps and grippers have been demonstrated repeatedly, with real force numbers, since about 2012. This page maps who built what, from which cells, at what measured output.

The landscape

Read that table for its shape rather than its rows. Two cell sources dominate — the C2C12 mouse myoblast line for skeletal work and neonatal rat cardiomyocytes for cardiac work — and both are chosen for convenience rather than for relevance. Absolute forces span three orders of magnitude, from single-microtube µN measurements to the low mN range. And the only entry that operates without an incubator and without stimulation is the insect one, which is the whole reason insect tissue keeps appearing in this field despite its unfamiliarity.

What the numbers actually mean

Three framing figures are worth carrying around.

Native adult human skeletal muscle produces roughly 150–250 mN/mm². That figure appears in Khodabukus and colleagues’ 2019 work as the comparison against which engineered tissue is judged. Sakar and colleagues quote 260 kPa for native adult skeletal muscle performance in the same spirit.

The best engineered human muscle reported reaches 19.3 mN/mm². Khodabukus and colleagues achieved this specific force after one week of chronic 1 Hz electrical stimulation of human myobundles — roughly a three-fold increase in tetanic force over unstimulated controls, and described in the paper as the highest specific force reported for engineered human muscle. That is about a tenth of native.

Most C2C12 biohybrid actuators produce under 1 mN. Chen and colleagues state this directly in their 2026 comparison: most C2C12 muscle tissues generate less than 1 mN of active force, with only a few reports exceeding 2 mN, and corresponding stresses consistently below 2 mN/mm². Their own 7.05 mN result is presented as an outlier achieved through a self-training maturation protocol.

So the honest performance envelope for a skeletal-muscle actuator today is: sub-mN to low-mN absolute force, single-digit mN/mm² specific force in most published work, up to roughly 19 mN/mm² at the very top of the human-tissue literature. The full normalised comparison, with cross-sections and stimulation protocols, is on the force specification reference.

The supply-side map

Here is what you can actually buy, which is a much shorter list than what has been published.

That last pair of rows is the interesting part and it is worth being precise about why. You can buy a finished contractile tissue: myriamed will ship you one. What you cannot buy is that tissue released against a stated twitch or tetanic force, which is the term an engineering purchase order needs. Several organisations build force-generating muscle constructs as a routine step and sell the resulting data: Curi Bio’s service programme engineers customer cells into 3D tissues and runs tetanic force, fatigue and damage assays; eNUVIO advertises contract research explicitly listing 3D muscle cultures; Creative Biolabs advertises customised in vitro 3D skeletal muscle models. The object exists inside all three workflows. The deliverable is a report. Whether any of them will contract to hand over the physical part against a force acceptance criterion is a commercial question, not a scientific one.

Why this field has no procurement literature

Searching for supplier-side content on biohybrid actuators returns academic papers and nothing else. That is not an accident of indexing; it reflects a market with perhaps a few hundred active research groups worldwide, no standardised part, and no repeat-purchase consumable other than generic cell culture reagents. Every group solves sourcing by emailing a colleague.

The practical consequence for anyone entering the field: budget for the integration, not for the parts. The cells cost hundreds to low thousands. The device costs $55 an experiment at the catalogue end. The instrument, if you buy one, is capital. What actually consumes the budget is the four to six weeks of skilled culture labour per iteration, and the fact that you will run several iterations before the construct does what you want.

Where to start, by starting position

Engineering group with no wet lab. Do not build a culture facility to test a concept. Contract the tissue construction and the force measurement to a lab that already has both, take the data, and decide afterwards whether the capability is worth owning. The CRO directory has the shortlist.

Biology group with culture capability and no device. Start with a catalogue casting device. The OMEGA-MP at $440 for eight experiments is the cheapest credible entry point in the field, and its published specification — 2.7 mm pillar height, 1 mm pillar width, 25–30 µL seeding volume, open-top for electrode access — is enough to plan against.

Group with both, needing a specific geometry. This is the custom fabrication path. Get the RFQ checklist right before you commission tooling, and put post stiffness in the specification with a tolerance — published data show day-10 twitch force falling from 1156 µN to 624 µN purely on moving post stiffness from 1.92 to 0.96 N/m.

Group whose actuator is the wrong metaphor. If what you want from living tissue is computation rather than movement, the supply picture is different again — a small number of organisations sell access to cultured neural networks on high-density electrode arrays, some of it as a cloud subscription rather than as hardware. The biological computing hardware landscape covers who those are, what they charge, and what they will not tell you.

Anyone specifying a force requirement. Read the force specification reference first. A requirement stated in µN without a cross-section, a stimulation protocol and an anchor stiffness is not a specification, and it will produce quotes that cannot be compared.

Reading this hub

Choosing a substrate and a route

• Cardiac vs skeletal muscle actuators — the decision is about control, not force. Cardiac contracts by itself and never stops; skeletal does nothing until commanded, then does exactly that.
• Muscle actuator sourcing guide — the four cell sourcing routes compared on cost, maturation time, force, culture burden, ethics and supplier.
• Optogenetic muscle control — driving tissue with light instead of electrodes, what it costs in force, and the two licensing chains it creates.
• Neuromuscular junction actuators — putting motor neurons in command, and the curare test that proves a junction is real.

Specifying and budgeting

• Engineered muscle force specifications — the normalised table across published papers, with cross-sections and stimulation protocols.
• Actuator operating envelope — 37 °C, buffered media, sterility, and the published lifetimes nobody advertises.
• Bill of materials — every line item for one actuator, priced where a price exists and marked quote-only where it does not.

Building the capability

• Cell culture for robotics labs — what an engineering group must acquire, learn and staff, against the cost of not doing it.
• 3D muscle culture devices — the hardware, with the one public price in the category.
• Tissue engineering CRO directory — who will quote the wet work.

How we can help

We source the cells, the device and the contract culture capacity against one specification and return the quotes normalised. Where a requirement is stated in force units, we will tell you which published work supports it and which does not, because a force spec that is outside the demonstrated envelope is worth knowing about before the purchase order rather than after.