SoftMining / Series A 2026

The operating system for AI + automated drug discovery

From fragmented R&D to a closed learning loop: evidence → design → integrated robotic synthesis → assay → learning.

$10MSeries A

19algorithms

5+2patents

Integrated discovery loop — Scientific evidence, models, robotics and experimental feedback in one system.

Problem

Drug discovery is not slow because AI is missing. Learning is fragmented

10–15 yrsto bring a new drug to market

$1–2Bper approved drug when failures are included

~90%of clinical candidates fail

AI, chemistry and biology still run in weak feedback loops. Context is lost between target reasoning, design, assays, failures and decisions.

The value leak is the repeated loss of scientific learning.

ADAPT / solution

Many specialised tools. One programme-level intelligence

ADAPT works holistically: each model contributes to a shared decision, and every result returns to the same scientific memory.

ONE SCIENTIFIC MEMORYADAPTGlobal AI core

DeepSageEvidence

Target reasoningBiology

Structural models3D context

SIMILISSimilarity

Pareto optimisationTrade-offs

ADMETDevelopability

SYNTHEXSynthesis

AssaysValidation

PharmacologyTranslation

Expert reviewDecisions

Technical ADAPT deep dive ↗

Global drug discovery~$150–175BLarge, durable R&D expenditure base

AI-enabled discovery~$1.9B → ~$6.9B2024 → 2029

Market / why now

Pharma is moving from point AI to validated discovery partnerships

Workflow shiftWinners will own feedback and execution context—not only predictions.

CRO pressureDemand is increasing for faster, cheaper and better-connected validation.

Initial wedgeInnovative mid-size pharma and venture-backed biotech across Europe, North America and Asia-Pacific.

Category positioning

SoftMining bridges AI prediction and experimental proof

AI-ONLY DISCOVERY

Predict

Models and scoring with weak experimental closure.

Limited reusable learning

CLASSICAL CRO / CDMO

Execute

Operational capacity detached from decision context.

Learning remains project-specific

SOFTMINING ADAPT

Learn

Evidence, computation, synthesis, assays and feedback in one system.

Compounding scientific memory

up to 20×faster workflows

~80%fewer false positives

up to 90%lower costs

Product & moat

The moat is a connected stack—not a single algorithm

Evidence intelligenceStructured signals, provenance and target reasoning.

Molecular computationStructural, similarity and multi-objective methods.

Robotic feedbackDirect connection to synthesis and validation.

Scientific memoryEvery project improves the next workflow.

ADAPT orchestration networkMatisse · Similis · Pareto · Flocking · specialised tools · one programme-level intelligence

Product pipeline

Agnostic in capability. Concrete in molecules

The platform can address different targets and therapeutic areas while advancing a real internal and co-development pipeline.

M01Glioblastoma

New therapeutic strategies
for aggressive brain tumours

M02Breast cancer

Discovery programme for
molecularly diverse disease

M03 · patent pendingNASH

Therapeutic programme for
metabolic liver disease

M04 · patent pendingAntibacterial I

First discovery programme
focused on bacterial infection

M05 · patent pendingAntibacterial II

Second programme expanding
the antibacterial pipeline

M06 · patent pendingADAR

Target-driven programme for
ADAR-mediated RNA editing

8products in pipeline

6molecules entering preclinical

Public communication intentionally withholds targets, indications and chemical structures.

SYNTHEX / physical discovery engine

Robots are not the story. The integrated learning loop is

Illustrative automated chemistry laboratory — Illustrative environment; the differentiation lies in customisation and integration.

Custom-configured robots

Automation configured around programme chemistry and handling.

Functionalised magnetic nanoparticles

Parallel synthesis, washing, separation and multi-step handling.

Integrated selection

ADAPT chooses high-information syntheses and tests.

SELECT → MAKE → SEPARATE → TEST → LEARN

Technical SYNTHEX deep dive ↗

Business model

Services fund the platform. SaaS scales it. Co-development creates upside

01 / SCIENTIFIC SERVICES

Resolve a bottleneck

Target intelligence, screening, toxicity and developability.

Entry revenue + customer proof

02 / SAAS + API

Embed ADAPT

Configured workflows, modules and enterprise access.

Recurring platform economics

03 / CO-DEVELOPMENT

Build assets

Milestones, licensing, shared IP and proprietary programmes.

Portfolio-level upside

Go-to-market

Land with a scientific win. Expand into infrastructure

2026 H2Preclinical pilots, robotic laboratory and Singapore market entry.

2027 H1First design → synthesis → preclinical workflow; ISO 27001 pre-audit.

2027 H2ADAPT SaaS/API pilots with pharma and Asia launch.

2028Licensing, recurring services and co-development deals.

International discovery infrastructure

Traction

Proof across science, IP, funding, users and preclinical assets

5+2patents filed / filing

$2.1Mfunding raised

19proprietary algorithms

8products in pipeline

6European projects

>120kbiomedical app users

6molecules entering preclinical

Top100IRCAI / UNESCO solution

Selected collaborations & relationships

University of Salerno

Laminar Pharma

AWS

Bayer

Microsoft

ETH Zurich

Management forecast

Move from project revenue to platform-scale economics

Revenue ($M)Operating costs ($M)

$0.5

$0.3

2026

$2.2

$1.0

2027

$17.5

$3.8

2028

$78.0

$12.0

2029

438%revenue CAGR 2026–2029

84.6%operating margin 2029

6.5×revenue / cost ratio 2029

Management plan from the 2026 investor deck. Revenue: $0.5M · $2.2M · $17.5M · $78.0M. Costs: $0.3M · $1.0M · $3.8M · $12.0M.

Team

Scientific depth with international execution

Stefano Piotto · CEOAI-driven drug discovery · computational chemistry

Rajesh Pahwa · CBOFinance · VC/PE · M&A · Singapore

Jeffrey Seah · CDOManagement · commercial operations

Luigi Di Biasi · CTOSoftware architecture · deployment

Simona Concilio · CSOMaterials science · synthesis · validation