Why are sophorolipids the biosurfactant with the best commercial economics?

Three reasons. First, S. bombicola natively produces sophorolipids at very high titer (200-400 g/L in research strains, 120-180 g/L in typical industrial processes), which is an order of magnitude higher than most other microbial surfactants. Second, sophorolipids phase-separate from broth by gravity in low-water media, which means the DSP can skip solvent extraction entirely. Third, the organism is a wild-type yeast that uses food-grade feedstocks (glucose + rapeseed or palm oil), so the regulatory path for food/cosmetic applications is straightforward.

What titer do I need to hit $5/kg or $2.50/kg COGS?

At 100-150 g/L titer and standard solvent extraction DSP, expect $15-30/kg. At 150-250 g/L with gravity phase separation (Holiferm-style), $5-12/kg is realistic. To reach $2.50/kg you need a high-titer strain (400+ g/L), phase separation DSP, and a solvent-free product finish. That combination is what Holiferm has been publicly reporting and it's the state-of-the-art for industrial sophorolipid economics.

Why does gravity phase separation work for sophorolipids specifically?

Sophorolipids accumulate as a lactonic and acidic form in the fermentation broth and at high enough concentration (typically above 100-150 g/L) they coalesce into a dense immiscible phase. In low-water fed-batch operation where vegetable oil is part of the feed, this phase settles to the bottom of the fermentor and can be drained directly. It looks like a honey-colored thick fluid. This is fundamentally different from rhamnolipid or MEL recovery, where the product stays emulsified in broth and requires solvent extraction or foam fractionation.

What's the difference between lactonic and acidic sophorolipids for scale-up economics?

Lactonic sophorolipids (cyclic form) have better surfactant performance and are preferred for most applications, but the native fermentation produces a mixture (typically 60-80% lactonic, 20-40% acidic). Separating the two forms requires additional DSP (chromatography or selective crystallization) which adds $3-8/kg. Most industrial processes skip this and ship the mixture, which is fine for industrial cleaner applications but not for cosmetics or personal care.

How do I switch from glucose-only to a glucose+oil feed in Augur's predictions?

In the scenario builder, select S. bombicola as the organism. The default feed is mixed glucose/rapeseed oil at a 40:60 mass ratio, which is industry-standard for sophorolipid production. You can override via the feedstock parameters: pure glucose runs at lower titer (80-120 g/L) because oil is a direct precursor to the lipid tail. Waste oil feedstock (waste cooking oil, animal tallow) can be substituted at 20-30% lower cost per batch, with a ~5-10% yield penalty due to oil composition variability.

Can the platform model Holiferm's specific process (inverted microbubble sparger + gravity separation)?

Partially. The inverted microbubble sparger improves kLa in low-water conditions, which you can capture in Augur by overriding the kLa parameter on the scenario. Gravity phase separation is a registered DSP unit operation (phase_separation) with tunable parameters (separation efficiency, drain yield, re-extraction yield). We can't match Holiferm's proprietary numbers without their internal data, but we can reproduce the general shape of their TEA (sub-$5/kg COGS at high titer with phase separation DSP) and show how different process choices move the number.

Sophorolipid production in S. bombicola: scale-up economics from lab to 50,000 L

Sophorolipids from Starmerella bombicola (formerly Candida bombicola) have the best commercial economics of any microbial biosurfactant, and they're the one product in the glycolipid category where sub-$5/kg COGS has been credibly reported. This post walks through the scale-up from lab to 50,000 L: what drives titer at scale, why gravity phase separation changes the DSP economics, and what combination of strain and process is needed to reach the $2.50/kg number that Holiferm has publicly cited.

What makes S. bombicola special

Most microbial biosurfactant producers top out at 20-80 g/L titer. S. bombicola natively produces sophorolipids at 120-180 g/L in typical industrial processes, and research strains have reported 200-400 g/L. The organism is naturally oleaginous, uses rapeseed oil (or other vegetable oils) as a lipid precursor, and tolerates very high product concentrations in the broth. Combined with the fact that sophorolipids physically separate from low-water broths by gravity, this gives the category a structural cost advantage over rhamnolipids or MEL.

Lab-scale baseline

A representative lab fermentation (5 L, fed-batch, 30°C, pH 3.5, rapeseed-oil + glucose feed at 40:60 mass ratio):

Duration: 200-300 h
Titer: 120-180 g/L at harvest
Glucose consumption: ~250 g/L total
Oil consumption: ~150 g/L total
Biomass: 40-60 g/L dry weight
Product yield on carbon: 0.35-0.50 g/g
Typical lactonic:acidic ratio: 70:30

The long duration is characteristic. Unlike P. putida rhamnolipid fermentations (3-5 days) or mAb CHO runs (10-14 days), sophorolipid processes run for 8-14 days to reach commercial titer. This long cycle time affects DSP scheduling and facility utilization. Fewer batches per year at a higher titer per batch can still come out ahead on cost per kg.

Scale-up physics: the kLa problem

Scaling from 5 L to 50,000 L introduces the usual hydrodynamic challenges, but sophorolipid fermentations have a specific twist: the broth viscosity rises sharply as product accumulates. At 120 g/L sophorolipid the apparent viscosity can be 10-20x that of water, and this crushes kLa in conventional stirred-tank bioreactors.

Typical kLa values across scale:

5 L lab (early batch, low viscosity): 300-400 h^-1
5 L lab (late batch, high viscosity): 80-120 h^-1
50,000 L tank (early batch): 80-120 h^-1
50,000 L tank (late batch, high viscosity): 20-40 h^-1

At a 20-40 h^-1 kLa, an unoptimized process is oxygen-limited for the last 30-50% of the fermentation, which caps achievable titer. Three mitigations exist. First, oxygen-enriched air (30-40% O2) can restore effective mass transfer. Second, inverted microbubble spargers (the Holiferm approach) improve kLa under low-water conditions without expensive O2 enrichment. Third, the reactor geometry can be designed for higher tip speed and power input at the cost of some shear.

DSP: solvent extraction vs gravity phase separation

Two DSP routes dominate sophorolipid production and the choice drives COGS more than any other single decision.

Solvent extraction (the traditional route)

Ethyl acetate or similar at 2:1 solvent-to-broth ratio
Single-pass yield: 85-92%
Requires solvent recovery (distillation) for economic viability
At 95% solvent recycling: DSP cost $18-28/kg at 150 g/L titer
At 85% recycling or worse: DSP cost balloons to $40-80/kg

Gravity phase separation (the Holiferm route)

Requires low-water fed-batch operation (water content below 40-50% of broth mass)
Sophorolipids accumulate as a dense immiscible phase at the bottom of the reactor
Phase drained directly, no solvent needed
Yield: 85-92% depending on re-extraction steps
DSP cost $3-8/kg at 150+ g/L titer, solvent-free

The gravity route is roughly 3-5x cheaper on DSP and eliminates solvent handling entirely. The tradeoff is tighter operational constraints: the fermentation has to run in low-water mode, and this doesn't work for every strain. It's the right choice for industrial cleaner and biodegradable surfactant applications. For cosmetic or personal-care applications where regulatory requirements favor a clearly-defined product specification, a second polishing step (chromatography or crystallization) is usually added regardless of the primary DSP route, eroding some of the advantage.

COGS at 50,000 L: scenario comparison

Three realistic scenarios at 50,000 L, 25 batches per year:

Baseline industrial: 150 g/L titer, solvent extraction at 95% recycling. 3,600 kg product per batch. Total COGS: $22-28/kg.
Phase separation upgrade: 150 g/L titer, gravity phase separation. 3,450 kg product per batch (slightly lower yield). Total COGS: $8-14/kg.
High-titer strain + phase separation: 300 g/L titer, gravity phase separation. 6,900 kg product per batch. Total COGS: $3-6/kg.

To hit $2.50/kg you need 400+ g/L titer, phase separation DSP, and a solvent-free product finish. That combination is the frontier of industrial sophorolipid economics today and is achievable only with purpose-engineered strains.

Feedstock sensitivity

Feedstock cost is 40-55% of COGS at industrial scale. Three feedstock strategies:

Food-grade rapeseed + glucose: $0.80-1.20/kg rapeseed oil, $0.40/kg refined glucose. Clean product spec, regulatory-friendly.
Industrial-grade rapeseed + corn syrup: 20-30% cheaper feedstock, minor titer hit (5-8%), needs additional DSP polishing for cosmetic applications.
Waste cooking oil + corn syrup: 40-50% cheaper feedstock, titer hit can be 10-15% due to oil composition variability, requires tighter DSP spec control. Can drop COGS to $2-4/kg at 200+ g/L titer.

What Augur predicts for your strain

If you upload 5-10 lab runs from your S. bombicola strain with glucose+oil fed-batch operation, the platform will predict production-scale titer, rate, yield, and COGS across any of the three DSP routes above. Switch the DSP template between solvent_extraction and phase_separation to see the cost delta directly. Enable the overflow metabolite tracking to surface any acetate or ethanol side-products that would indicate oxygen limitation.

Pilot customers working on sophorolipid production have seen prediction intervals of 15-25% on titer and 25-40% on COGS at 50,000 L scale. The intervals widen for phase separation DSP because the unit operation is newer and the empirical correlations are less established than for solvent extraction.

If you're developing a sophorolipid process and want to see production-scale economics for your specific strain and DSP choices, request access. We have capacity to help pilot customers configure phase separation DSP chains and custom feedstock profiles during onboarding.