Questioning Assumptions and (Inoculum) Potential

The Challenge

If you work in private R&D, outside of patent disclosures, very often the bulk of your work is invisible outside of the commercial results. This is one of the stranger tensions of industrial science. The most formative work, the work that actually changes your judgment, often cannot be fully shown. It remains embedded in patents, products, bruises, failed batches, exhausted teams, and the odd sentence in a review paper where years of difficulty get compressed into a few technically defensible words.

A few years ago I had the opportunity to be part of a team that plowed new ground in solid-state fermentation systems, meaningfully challenging previous state-of-the-art assumptions. We were able to share part of that work, at least to a reasonable level of detail, in the review paper The Amazing Potential of Fungi: 50 Ways We Can Exploit Fungi Industrially (Hyde et al., 2019, section 41). That paper describes a mycelium composite cultivation methodology that combined a reimagined propagation strategy with an actively aerated solid-state bioreactor system (Hyde et al., 2019). The work was technical, biological, physical, ecological, and organizational all at once. The discussion that follows is limited to information already disclosed in the public literature and to my own non-confidential interpretation of the project’s broader lessons; it does not disclose proprietary methods, parameters, data, or business strategy.

The system addressed the inherent limitations of passively aerated tray-based fermentation by incorporating programmable control over airflow, temperature, and humidity through modular air pretreatment and distribution infrastructure. The bioreactor enabled uniform colonization of lignocellulosic substrates in 0.7 m³ blocks while significantly reducing reliance on aseptic processing and supporting consistent material formation across depth gradients (Hyde et al., 2019; Mueller et al., 2022). Engineering features included pressure-rated humidification chambers, custom nozzle arrays for more homogeneous gas distribution, and vessel geometries designed to mitigate airflow bypass events (for example, the wonderful and awful phenomenon of “burping”) under increasing back pressure during colonization (Mueller et al., 2022).

Development of the system integrated quantitative image-based analysis of epoxy-embedded thin sections to characterize inter-particle hyphal morphology and correlate structural features with environmental parameters and mechanical outcomes (Hyde et al., 2019). Alongside the pilot system, we developed a scaled-down bioreactor system that allowed recursive learning between bench and pilot scale. In other words, the project forced us to treat biology, scale, propagation ecology, imaging, mechanics, and experimental learning as one problem.

Questioning Assumptions from the Previous State of the Art

The goal was a deep bed (>60 cm) static fermentation targeting high inter-particle colonization density and uniform mechanical characteristics along the depth of the bed; a single monolithic block of myceliated material that could meet relevant mechanical performance throughout its volume.

If you have ever grown a passively aerated tray of densely myceliated substrate deeper than about 16 cm, you can appreciate how ambitious this goal was. Deep-bed mycelium biomaterial production poses significant challenges because fungal growth consumes oxygen and produces heat and CO₂, while simple tray-based growth is limited by diffusion of heat and gases beyond relatively shallow depths (Mueller et al., 2022). As mycelium grows and inter-particle porosity decreases, back pressure rises. Airflow channels and oxygen gradients form. Metabolic heat and CO₂ accumulate, intensifying vertical temperature and gas differentials. Pressurized air can lift sections of the substrate, disrupting the bed and turning uniformity into wishful thinking.

There were a tremendous number of reasons to expect that pursuing a deep-bed reactor would be untenable. Airflow channeling, oxygen and temperature gradients, metabolic heat accumulation, physical deformation of the bed, uneven colonization, escalating control complexity; the known challenges alone were enough to warrant caution. And caution would have been reasonable.

The real question was whether those objections fully described the opportunity. They described a deep-bed reactor as viewed through conventional assumptions about trays, asepsis, bed depth, airflow, and spawn. What remained open was what might become possible if the cultivation strategy, physical reactor, and ecological management of colonization were redesigned in concert around the material target itself.

Sometimes the strongest argument for experimentation is a calculated confidence in success. Other times the strongest argument is an appreciation for how incomplete the prevailing assumptions might be when viewed through the lens of a different target within a different context. This was the latter kind of problem. The system’s success was uncertain, but the existing assumptions clearly did not exhaust the design space.

Rethinking Mycelium Propagation Conventions

The notion of building monolithic bulk myceliation systems for consumer products that required strict aseptic control was not particularly realistic. That premise can be made to sound clean on paper. In practice, it collides quickly with feedstock reality, process economics, reactor volume, material handling, and the basic fact that a large mass of wet lignocellulosic substrate is an ecological object before it is an industrial input.

This is where propagation ecology, and within it inoculum potential, became useful.

Inoculum potential, classically defined in plant pathology as the growth energy available for infection or colonization, can be adapted here to mean the capacity of an introduced fungal propagule to successfully colonize a given substrate or environment, taking into account its biological vigor and the ecological context in which it is deployed (Lockwood, 1986). It integrates the quantity and viability of the inoculum, the geometric distribution of that inoculum, compatibility with the substrate and conditioning, timing and staging of resource availability, and the competitive landscape. Functionally, inoculum potential is an ecological concept directed to managing the evolution of priority effects, particularly under conditions where the system cannot be perfectly closed.

The trap is to reduce inoculum potential to spawn rate. More spawn. Better spawn, cleaner spawn, finer distribution. All of those matter, but none of them capture the full inoculum solution space.

As an artist and former teacher, I appreciate the flexibility of positive and negative space in visual design; that moment when perception flips and the negative space becomes the thing you are actually seeing. Suddenly the complexity around the focal object becomes the information. Something similar happened for me in rethinking mycelium propagation. Spawn was the obvious object. It was visible, countable, and easy to name as the intervention. But the more important design space was often around the spawn: the temporal ecological context into which the spawn was introduced.

That shift mattered. The creative management of the ecological field around spawning became central: substrate history, resource timing, moisture and thermal trajectory, aeration logic, biological vigor, competitor pressure, and the window during which the desired fungus could establish exclusivity. The details of these strategic tools are largely wrapped up in patent disclosures, but the transferable principle is simple enough: propagation ecology functions as a negotiated continuum between inoculum, substrate, competitors, timing, and physical context.

By working in concert across cultivation strategy and physical bioreactor design, the team was able to reimagine what a managed propagation ecology might tolerate. The result was a system in which the biology and engineering made one another more plausible, exceeding a simple workaround for contamination risk or a clever reactor configuration. Ultimately, this holistic approach helped navigate the inherent physical problems of deep-bed bioreactors while enabling operation with reduced reliance on aseptic processing and with substrate and propagation strategies designed to reduce process cost and complexity (Hyde et al., 2019; Winiski et al., 2018; Mueller et al., 2022). Just as importantly, it reduced dependence on expensive physical controls by shifting part of the control burden into the managed propagation ecology itself.

Cutting to the Chase

When you have competing development targets, you are often best served by cutting to the chase. In this case the competing targets were wonderfully concrete: maximize the mechanical strength of mycelium binding between substrate particles while preserving enough air passage for respiration and heat dissipation. Too little binding and the material fails. Too much density in the wrong way and the bed suffocates, overheats, or becomes mechanically uneven. The target lived in the mycelium's physical relationship between particles (binding vs. blockage).

So the tissue between the particles became the thing to see.

In this project, quantitative imaging was a central design tool. As described in the review, we used epoxy-embedded thin sections coupled with high-resolution image analysis to directly characterize the hyphal network bridging substrate particles (Hyde et al., 2019). This approach captured both the quantity and the microstructural organization of inter-particle mycelium, offering a window into the physical architecture that ultimately governed composite integrity.

This is the kind of measurement that matters because it sits close to the value-creating structure. Biomass proxies can be useful. Visual coverage can be useful. Endpoint mechanical testing can be useful. But if the product and the physics producing it depends on mycelium acting as a binding tissue, then the physical logic of that binding tissue deserves direct attention. Otherwise, you risk optimizing around shadows cast by the structure instead of the structure itself.

Critically, we integrated high-throughput image analysis and predictive modeling to link hyphal density and network organization with reactor performance and mechanical performance outcomes (Hyde et al., 2019). By quantifying structural features and mechanical properties across environmental conditions, we could model how variation in hyphal network complexity translated into changes in bulk material behavior.

This allowed us to move past speculation about how a process choice might influence the final product. The question became more concrete: did this set of conditions produce the inter-particle mycelial architecture that the material needed? Did that architecture persist through depth? Did it correlate with mechanical performance? What was the associated propagation and reactor behavior?

The Value of Bi-Directional Scaling

Previously I described bi-directional scaling as the practice of developing both larger and smaller versions of a system in parallel; treating scale as a recursive learning space rather than a linear sequence. This project provided a very real lesson in what that philosophy looks like in practice.

While developing the pilot-scale actively aerated bioreactor, we faced classic deep-bed challenges: airflow dynamics, metabolic gradients, heat accumulation, changing back pressure, mechanical variability, and spatial heterogeneity through the bed. Conventional bench systems did not easily simulate these specific spatial and temporal heterogeneities. Direct use of the pilot-scale system for complex multi-dimensional characterization and optimization was also unrealistic. It was too slow, too expensive, and too blunt an instrument for the amount of learning we needed.

So we built a tailored bench-scale analogue: an array of more than 50 individual units at less than 100 mL operating volume per unit, each designed to simulate gas exchange and temperature dynamics according to substrate depth conditions and growth timepoint of the full reactor (Hyde et al., 2019). This was a deliberately specific learning tool built to preserve the critical scale-dependent behaviors that mattered.

The system was first validated using mechanical and hyphal network models developed from the full-scale reactor. Once validated, it allowed rapid structured optimization at the bench scale, including response surface design, to refine operational parameters within the larger system (Hyde et al., 2019). It also allowed us to understand and order critical time-dependent sensitivities driving ultimate mechanical performance relative to spatial distribution through the depth of the bed.

The result was true recursive learning. Mechanical and biological insights from the pilot scale informed bench-scale design. Bench-scale optimization accelerated pilot-scale improvement. Each addressed distinct and complementary regions of the total design space; scale itself became an instrument.

A Holistic Lesson in Mycelium R&D

In the end, this was a reminder that real progress in mycelium R&D often comes from the willingness to let assumptions become unstable long enough to be redesigned.

A deep bed looked physically unreasonable under one set of assumptions. A large open-top monolithic process looked ecologically unreasonable under another. Inoculum looked like a spawning parameter until the negative space around spawning became the real design field. Bench scale looked inadequate until it was rebuilt as a targeted analogue for the right scale-dependent behaviors.

Each shift required multidisciplinary teaming, direct observation, ecological imagination, and a willingness to make the learning system as designed as the reactor itself. In mycelium R&D, questioning assumptions means changing the shape of the problem until biology, engineering, and measurement can meet in a more useful arrangement.

That is where inoculum potential becomes more than a propagation term. It becomes a way of seeing how ecological possibility is constructed; how a fungus is given the chance to arrive early, occupy meaningfully, and build the material world we are asking it to make.

References

Hyde, K. D., Xu, J., Rapior, S., Jeewon, R., Lumyong, S., Niego, A. G. T., Abeywickrama, P. D., Aluthmuhandiram, J. V. S., Brahamanage, R. S., Brooks, S., Chaiyasen, A., Chethana, K. W. T., Chomnunti, P., Chepkirui, C., Chuankid, B., de Silva, N. I., Doilom, M., Faulds, C., Gentekaki, E., ... Stadler, M. (2019). The amazing potential of fungi: 50 ways we can exploit fungi industrially. Fungal Diversity, 97(1), 1–136. https://doi.org/10.1007/s13225-019-00430-9

Lockwood, J. L. (1986). Soilborne plant pathogens: Concepts and connections. Phytopathology, 76(1), 20–27.

Mueller, P. J., Winiski, J. M., & O’Brien, M. A. (2022). Process and apparatus for producing mycelium biomaterial (U.S. Patent No. 11,343,979). U.S. Patent and Trademark Office. https://patents.google.com/patent/US11343979B2/en

Winiski, J., Van Hook, S., Lucht, M., & McIntyre, G. (2018). Process for solid-state cultivation of mycelium on a lignocellulose substrate (U.S. Patent No. 9,914,906). U.S. Patent and Trademark Office. https://patents.google.com/patent/US9914906B2/en