☰

Emulsi Biotech AB

Scaling lipid-based drug delivery from pilot to market

Industry
Biotechnology — Drug Delivery
Headquarters
Lund, Sweden
Public information as of
February 2026

A4BEE prepared this analysis from publicly available sources. It reflects our own reading of Emulsi Biotech AB's published strategy and is not endorsed by, or produced in cooperation with, Emulsi Biotech AB. Company website

Strategic priorities

Emulsi Biotech is moving its NanoLIPID platform from laboratory verification to pilot-scale production. The company runs lipid modification at 2 kilograms per day at the Lund University Pre-Pilot Plant (LUPPP) and is targeting a 10x scale-up so clinical trial material can be produced under pharmaceutical partner standards. The strategic direction has been laid out in a 2024 Vinnova award (SEK 824,344) and a 2025 Vinnova Innovative Startups grant focused on the platform's use in obesity and diabetes.

The platform uses enzymatic rearrangement of vegetable oils to make lipid excipients that self-assemble into nanostructures on contact with water, with the stated aim of delivering biologic drugs orally. Validation work runs at 0.3 g to 3 kg scale and is split across three separate sites: process data at LUPPP, analytical work at the SmiLe Venture Hub, and scale-up validation with RISE Research Institutes of Sweden. The pattern is a small, IP-heavy spin-out whose commercial case rests on proving the platform scales reliably.

Funding in early 2026 is still non-dilutive and grant-driven. The next capital event is a Series A round that company materials describe as contingent on validated pilot data, and the longer-term plan is to license the production process as a modular unit to pharmaceutical partners. Both paths pass through the same data and documentation infrastructure that the company has not yet had the scale to build.

The growth constraint sits at the interface between process, analytical and validation data. A unified record of what was made, how it was tested and what the result was is the prerequisite for the partner audits, the Series A diligence and the regulatory file that all three capital paths depend on.

Challenges we see

  • Operations Manufacturing

    Engineering reactors to withstand lipid process chemistry

    Enzymatic modification of vegetable oils operates in pH and moisture ranges that attack standard 316L stainless steel. A 2024 RISE report on the project described severe corrosion signs in the reactor used for lipid mixture production at LUPPP.

    Where the reactor material is the limiting factor on campaign length, the practical question becomes which alloys, coatings or process windows let a campaign run through without unscheduled intervention.

  • Operations Manufacturing

    Holding self-assembly conditions steady across larger batches

    The NanoLIPID platform depends on precise lipid ratios and on temperature and mixing profiles during enzymatic rearrangement. The process is operated at 2 kg per day or 3 kg per batch, ten times the 0.3 g laboratory scale at which the original characterisation was performed.

    As the working volume grows, the same nominal setpoint sits on a different surface of the parameter envelope, which moves parameter control and not just parameter target onto the critical path.

  • Digital Integration

    Connecting process, analytical and validation data across sites

    Process data is generated at LUPPP, analytical data (HPLC and GC-MS) at the SmiLe Venture Hub, and validation data at RISE. The deep research report describes reconciliation of these sources as a manual step performed in Excel.

    Where the same batch lives as three different records, optimising the formulation requires assembling it from parts, which limits how quickly a result can be acted on across the sites that produced it.

  • Digital Integration

    Linking laboratory analytics to reactor controls in real time

    OT systems controlling the LUPPP reactors (pressurized air, cooling, vacuum) are air-gapped from the IT systems used for lipid profiling and analytical work at SmiLe.

    In a process where the analytical result could feed back into a setpoint within the same shift, an air gap turns an information loop into a paper one, which keeps the operators working from the last reading rather than from the current state.

  • Compliance Regulatory

    Building GxP-ready data trails for partner audits

    Pharmaceutical partners in the SmiLe ecosystem, including sponsors such as AbbVie, expect comprehensive audit-ready data trails. Current process and analytical records are kept in manual in vitro study files and techno-economic calculations that do not yet meet GAMP5 expectations.

    If the data trail is compiled on request, every partner audit repeats the assembly work; if it is produced by the process itself, the work is done once and reused.

Opportunities, by urgency and business impact

Each bubble is one opportunity, numbered to match the list below. Further right means it bites sooner; higher means a bigger effect on the business. A bigger bubble means a bigger implementation effort.

Source: A4BEE analysis of public sources

  1. Unifying process, analytical and validation data across sites

    Process data at LUPPP, analytical data at SmiLe and validation data at RISE are recorded in separate systems and reconciled manually. Each formulation cycle pays a fixed cost in data assembly before a result is usable.

    An ontology-based data platform that defines assay, lot, batch, equipment and result once, and then loads each site's output against that model, makes the formulation cycle visible without manual reconciliation.

    • Vinnova project 'Novel Solvent-Free Excipients for Enhanced Drug Solubility and Delivery', reference 2024-01516
    • Emulsi Biotech deep research report, 4 February 2026

  2. Simulating the enzymatic process before committing a batch

    Scaling enzymatic lipid modification from 0.3 g to 2 kg per day introduces temperature, mixing and pH dynamics that are not fully characterised, and a failed batch at pilot scale costs material and campaign time.

    A digital twin of the enzymatic rearrangement process lets the team test reactor conditions and lipid ratios in software and only commit a physical batch to the parameter windows the simulation predicts will hold.

    • Vinnova project report on solvent-free excipients, 2024
    • Emulsi Biotech deep research report, 4 February 2026

  3. Standing up a GxP-compliant data foundation for partner audits

    Pharmaceutical partners and regulatory submissions require audit-ready data trails, but current records rely on manual in vitro study files and techno-economic calculations. The gap shows up first in a partner audit and then again in the regulatory file.

    A GAMP5-aligned data foundation with electronic records, audit trails and role-based access produces the evidence chain once, in a form that holds up across partner qualification and regulatory submission.

    • Emulsi Biotech deep research report, 4 February 2026
    • SmiLe Venture Hub portfolio entry for Emulsi Biotech

  4. Standardising the pilot unit as a licensable module

    The pilot process relies on manual operation of decanters, separators and filters downstream of the reactor. The same manual steps at the partner site become a support liability rather than an installed base.

    An MTP-based (Module Type Package) module specification wraps the lipid unit into a vendor-neutral package with documented service interfaces, so a pharmaceutical partner can integrate it into their own control system without re-engineering the internal steps.

    • Lund University 'Smart mini factory testbed for small-scale production of biologics'
    • Emulsi Biotech deep research report, 4 February 2026

  5. Reducing campaign interruption from reactor corrosion

    Lipid process chemistry attacks 316L stainless steel. Unscheduled reactor repair pulls campaign time out of the schedule, and the available mitigation (changing alloys or adding coatings) is itself a development task that needs process data to inform it.

    Process data and material history in one record lets the team see which operating windows are associated with shorter campaign life and decide on alloy, coating or process-window interventions on evidence rather than on calendar.

    • RISE project record on lipid mixture production, 2024
    • Emulsi Biotech deep research report, 4 February 2026

What we'd propose

  • Enterprise AI

    Source-to-scientist data platform for the NanoLIPID programme

    An ontology-based data platform that defines assay, lot, batch, equipment and result once, and loads process data from LUPPP, analytical data from SmiLe and validation data from RISE against the same model, so the team queries one record rather than reconciling three.

    • Shared lipid programme ontology

      One agreed vocabulary

      Define the entities that cross sites (batch, lot, process run, analytical result, equipment) and the relationships between them, so a query written once returns comparable answers from each site instead of three dialects of the same table.

    • Pipelines from process, lab and validation systems

      Loading every site

      Build ingestion for LUPPP reactor data (OPC UA, Open Platform Communications Unified Architecture), SmiLe analytical instruments (HPLC, GC-MS) and RISE validation systems, with schema validation at the boundary so a bad record fails loudly rather than silently.

    • Real-time formulation analytics

      Results the team can act on

      Expose the model through dashboards and a retrieval layer so scientists see process conditions and analytical results together, with the lineage needed to follow a result back to the instrument that produced it.

    • Each formulation cycle saves the manual reconciliation step that today sits between experiment and decision.
    • The Series A data room and the partner audit read from the same record, so diligence stops being a separate data build.
    • New assays and new sites attach to the model rather than triggering another integration project.

    We've done this before

    See how we do Enterprise AI
  • Digital Lab

    Digital twin of the enzymatic lipid modification process

    A simulation of the enzymatic rearrangement reactor, calibrated against the pilot runs at LUPPP, that the team uses to test lipid ratios, temperature and mixing profiles in software before committing a physical batch.

    • Reactor simulation calibrated to LUPPP

      Process model in software

      Build a physics-based model of the rotating bed reactor that captures temperature, mixing, pH and moisture dynamics, then calibrate it against the existing pilot runs so the model behaves like the unit it represents.

    • Material and corrosion interaction

      Operating window vs reactor life

      Layer a material-interaction model on top of the process model so the team can see how candidate operating windows interact with 316L stainless steel and with candidate alloys or coatings, supporting a campaign-life decision on evidence.

    • Batch parameter workspace

      Test before commit

      Provide a workspace where scientists compare simulated outcomes for different lipid ratios and process setpoints, with the predicted outcome traceable back to the model parameters and to the pilot data those parameters were calibrated against.

    • Failed batches at 2 kg or 3 kg scale drop because the parameter window has been narrowed in software first.
    • The case for a campaign-life intervention (alloy, coating, operating window) is supported by the model rather than asserted in a meeting.
    • Pilot data is captured in a form that can be replayed against the model for any future scale-up question.

    We've done this before

    See how we do Digital Lab
  • Digital Lab

    GxP-ready data foundation for partner audits and regulatory files

    A GAMP5-aligned data layer (the Good Automated Manufacturing Practice guideline used in pharmaceutical software qualification) with electronic records, audit trails and role-based access, so partner qualification, regulatory filing and internal record-keeping all read from one audit-ready store.

    • GAMP5-aligned data layer

      Records that hold up to audit

      Implement electronic records and audit trails to ALCOA+ principles (Attributable, Legible, Contemporaneous, Original, Accurate, plus Complete, Consistent, Enduring and Available), with role-based access and electronic signatures where required.

    • Validated cloud environment

      Where the records actually live

      Stand up the data layer on a validated cloud environment with encryption, access logging and change control, sized for a small team's traffic and reviewed for the controls a partner audit expects.

    • Electronic batch record

      One record per batch

      Capture process parameters, analytical results and release decisions in a single per-batch record, so the evidence behind a release decision is generated as the batch is made and is read directly during an inspection.

    • Partner qualification stops being a one-off data build and reads from the same store as the rest of the work.
    • Regulatory submissions start from a record that has been kept for its own reasons, not assembled at the end.
    • Every batch carries an evidence chain that an inspector can read without reconstruction.

    We've done this before

    See how we do Digital Lab
  • Digital CDMO

    MTP-based modular production package for licensing

    A Module Type Package specification of the lipid pilot unit, including the downstream decanters, separators and filters, so a pharmaceutical partner can install the unit as a documented process module in their own control environment.

    • Process Equipment Assembly specification

      Module definition

      Define the lipid unit as an MTP-compliant Process Equipment Assembly with documented service interfaces, so the partner's orchestration layer can call it as a unit without re-engineering the internal steps.

    • Automated downstream unit operations

      Removing the manual bottleneck

      Wrap the decanter, separator and filter steps in the same module specification, with the control logic that today sits in operator hands expressed in the module's own service interface.

    • Partner integration guide

      How to install the module

      Document the network, security and orchestration assumptions a partner needs to satisfy, so a deployment at the partner site is a configuration task rather than an integration project.

    • The licensing conversation moves from selling a process to selling a module with a documented interface.
    • Downstream steps that today depend on operator skill become part of the package.
    • Future pilot iterations are absorbed into the module spec instead of triggering a new integration.

    We've done this before

    See how we do Digital CDMO

Digital maturity: today and target

Scored out of 100 across six dimensions. The target is what Emulsi Biotech AB's own published ambition implies — not a perfect score.

Source: A4BEE analysis of public sources
Data Integration 25 → 80
Process data at LUPPP, analytical data at SmiLe and validation data at RISE are reconciled manually today. A shared model across the three sites is the prerequisite for the rest of the programme.
IT/OT Convergence 15 → 75
OT systems at LUPPP are air-gapped from the IT systems used for analytical work, so a feedback loop between analytical result and reactor setpoint is a manual step. Connecting them within a controlled security boundary is what makes that loop automatic.
Process Automation 30 → 85
Downstream decanters, separators and filters are operated manually, which caps throughput at pilot scale and makes the process dependent on operator skill. Documenting these steps as automated unit operations is the basis for the licensing case.
Predictive Analytics 20 → 75
Reactor campaign life is managed reactively, and the recent RISE record of corrosion signs in the lipid reactor shows the cost of that posture. Process data captured today is the input the team will need to forecast campaign life in future.
Regulatory Compliance 35 → 90
Current records are kept in manual in vitro study files and techno-economic calculations that do not yet meet GAMP5 expectations. Producing audit-ready records as part of the normal workflow is the change that puts the partner and regulatory paths on the same footing.
Cybersecurity 25 → 80
Public-facing signals point to basic, shared hosting. For a company whose value sits in a lipid platform being prepared for partner qualification, an enterprise-grade security and access-control posture is part of the audit conversation.

Check this yourself

Our Service Portal has free self-assessments and market comparisons. These are the ones that line up with what we've read above — no sales call required.

Think we've read this right?

Talk to us

Related reading

This is an independent analysis prepared by A4BEE from publicly available information as of February 2026. It reflects A4BEE's own interpretation and opinion, is not affiliated with, endorsed by, or verified with Emulsi Biotech AB, and may be incomplete or inaccurate. All company names and trademarks are the property of their respective owners. To request a correction or removal, contact [email protected].