Skip to content
Sign in
SWC 2026 FinalistGrand Final: Lunch Block

Ethermind

UK/Bulgarian start-up turning top experts into on-premise “AI twins” for training and know-how retention in nuclear, pharma and defense.

London, United Kingdom, EuropeAI & software, AI agents for knowledge management / corporate trainingBulgaria – Startup World Cup Bulgaria (Crossroads 2025, NDK, Sofia; article dated 21 Oct. 2025), winner
Stage
Pre-seedAI at the core
Product
PilotPilots / LOIs
Funding raised
$635Kdisclosed
Founded
20242 years ago
Founders
1Teodor
Global score
4.5Grade DRank 63 of 151

Overview

Ethermind creates “AI twins” of an organization's best experts to capture their know-how and make it available 24/7 to teams (training, upskilling).

Traction

Pilots / LOIs

Six anonymized use cases (global school chain, mining, telecom, pharma, software publisher, training provider); targets energy, pharma, mining, telecoms; no named customers. Recognition: EdTechX Awards, Europass European Tech Startup Awards, Training Industry.

Market

Enterprise knowledge management software

Industrial and regulated organisations face retirement of senior experts and slow onboarding; they buy knowledge-management and training tools to retain know-how. Generative AI makes conversational access to expert knowledge possible, but regulated sectors require on-premise deployment and traceability. Buyers are L&D, operations and knowledge-management heads; sales cycles are long and pilot-driven. Ethermind differentiates on classified/on-prem deployment.

Addressable market

2025
$14.7B
2030Not estimated
+14.3% a year

Enterprise knowledge management software: $14.73bn in 2025 (Spherical Insights) to $33.5bn (MRFR), ~14% annual growth.

Customers

Energy, nuclear, pharma, mining, telecom and defense organisations; training providers

Who signs the cheque

Heads of L&D, operations excellence and knowledge management

Geographies

  • United Kingdom
  • Bulgaria
  • Europe

Drivers

  • Retirement of experienced workforce
  • Onboarding and upskilling costs
  • Demand for on-premise AI in regulated sectors

Headwinds

  • Anonymous customers
  • Competition from enterprise copilots
  • Long procurement in defense/nuclear

Global knowledge management software market estimated at between $14.7bn (Spherical Insights) and $33.5bn (Market Research Future) in 2025, growing ~14% per year (Technavio).

Scalability 8/10. English-language B2B SaaS deployable in any country, but long enterprise sales cycles and per-expert customization.

Market source: sphericalinsights.com

Technology

Category

Enterprise AI agents (knowledge management)

AI at the coreAI agentsGenerative AI / LLM

Maturity

  1. R&D
  2. Prototype
  3. Pilot
  4. On the market
  5. Scaling

pilot

How it works

Knowledge agents are trained on an organisation's internal expertise to act as 24/7 “AI twins” of its best experts. They can be deployed on premises, including in classified environments.

Moat

Moderate: on-prem/classified deployment capability and captured expert content create switching costs; base models are commodity.

Differentiation

Expert-centric capture and classified deployment versus cloud-only copilots.

Integration points

  • LMS / training platforms
  • Maintenance and field-service tools
  • Document repositories and knowledge graphs
  • Wearables for frontline workers

Depends on

  • Open or licensable LLMs for on-prem use
  • Expert time for knowledge capture
  • Security accreditation

Use of AI

Expert “AI twins”: knowledge agents trained on internal know-how, available 24/7.

Product in detail

AI knowledge agents trained on internal expertise, deployable on premises (including classified environments) for nuclear, pharma and defense.

Business

Model
Enterprise licences and deployments (pricing not disclosed).
Revenue logic
Per-twin or per-organisation licence (assumed).
Go-to-market
Direct enterprise pilots; awards (EdTechX, Europass, Training Industry).

Next milestones

  1. Named reference customer in a regulated sector
  2. Security accreditation for classified use
  3. Seed round

Key risks

  • No named customers
  • Competition
  • Long sales cycles

Opportunities

4 found in the network

Concrete moves for Ethermind inside the network: who to sell to, sell through, build with or raise from, each with the reasoning, the evidence and the first step.

  • Co-development 3
  • Investors 1

Co-development3

An integration or joint product

Deliver expert twins hands-free to technicians via smart glasses

Viz Ai builds hands-free assistants for mechanics with real-time sync to an expert; Ethermind's AI twin can be that expert when humans are unavailable. Ethermind gains a frontline delivery channel; Viz gains captured expert know-how.

30%confidence

Evidence from the records

  • “AI knowledge agents trained on internal expertise”
  • “hands-free diagnostics, automatically generated repair logs and real-time sync with an expert”
If it works
A joint pilot in one Viz dealership or repair shop where the expert is an Ethermind twin.
First step
Hold a joint demo session connecting an Ethermind twin to Viz's glasses app.

Combine expert twins with auditable knowledge graphs for pharma

Both target pharma; Deepentix provides deterministic, source-traceable knowledge graphs and small language models, which addresses regulators' traceability concerns about AI twins. A joint stack is more credible in GxP environments than either alone.

25%confidence

Evidence from the records

  • “deployable on premises (including classified environments) for nuclear, pharma and defense”
  • “"CLAIM Graph" and vertical small language models (SLMs) providing a deterministic chain of evidence traceable back to the sources”
If it works
A joint pharma pilot proposal with traceable expert answers.
First step
Map both companies' pharma prospects and draft a joint pilot scope.

Embed expert maintenance know-how into InnoKeep’s CMMS

InnoKeep digitises equipment manuals and service reports; Ethermind captures tacit expert reasoning. Linking the two gives maintenance teams both documented and tacit knowledge, and opens Korean manufacturing accounts to Ethermind.

15%confidence

Evidence from the records

  • “Automatic extraction of equipment data from PDF manuals, voice transcription of service reports”
  • “targets energy, pharma, mining, telecoms”
If it works
An integration demo for one shared manufacturing prospect.
First step
Request InnoKeep's API documentation and propose an integration prototype.

Investors1

A fund whose thesis fits

Pitch Seedcamp for a European seed round

Seedcamp leads angel and seed rounds across Europe in AI & software; Ethermind's ~GBP 500K pre-seed and UK base fit its entry point, but a named customer is needed first. No network warm path identified.

15%confidence

Evidence from the records

  • “Community-led early-stage investing across Europe: leads angel and seed rounds, can join rounds up to Series B.”
  • “~GBP 500K (~$646K)”
If it works
A seed lead to fund regulated-sector pilots.
First step
Prepare a one-page with one referenceable customer and apply through Seedcamp's open submission.

Analyst reading of 21 Nov 2026, from the public record of the network. Method: analyst-ai.

Investors that fit

London early-stage investor ($180M Fund VI, 460+ companies, nine unicorns) leading pre-seed and seed rounds across Europe.

London 'AI twins' of an organisation's experts with use cases in pharma, mining and telecom: enterprise AI from the UK at pre-seed.

50%confidence

Evidence from the records

  • “AI agents for knowledge management / corporate training”
  • “United Kingdom”
Intro path
Sofia Angels Ventures (Ethermind investor).
First step
Pitch Ethermind once it converts an anonymised use case into a named customer.

London seed firm of Robin and Saul Klein ($115M seed + $180M Latitude growth fund) behind Wise, Citymapper and Improbable.

UK pre-seed AI agents for knowledge management — UK AI at LocalGlobe's entry stage.

45%confidence

Evidence from the records

  • “"Pre-seed","Seed"”
  • “AI agents for knowledge management / corporate training”
  • “Six anonymized use cases (global school chain, mining, telecom, pharma, software publisher, training provider)”
First step
Send memo via a UK angel network.

UK EIS fund manager giving private investors access to early-stage tech, life-science and renewable-energy companies; backed Ocuwell.

UK AI agents for knowledge management at pre-seed with six enterprise use cases: EIS-eligible UK tech (assuming UK entity).

45%confidence

Evidence from the records

  • “AI agents for knowledge management / corporate training”
  • “United Kingdom”
Intro path
Ocuwell (Deepbridge portfolio, UK peer)
First step
Confirm EIS advance assurance, then pitch.

Liverpool investor with no profile in our data; participated in Ocuwell's 2025 seed.

UK pre-seed AI knowledge management — UK company.

25%confidence

Evidence from the records

  • “Ocuwell”
  • “AI agents for knowledge management / corporate training”
  • “Six anonymized use cases (global school chain, mining, telecom, pharma, software publisher, training provider)”
Intro path
Founder-to-founder intro via Ocuwell Limited (network portfolio company of LYVA Labs).
First step
Qualify via Ocuwell.

Manchester/Leeds/Sheffield university-founded investor (£312M) backing spin-outs and IP-rich businesses in the North of England.

UK AI knowledge management — UK; Northern base unknown.

25%confidence

Evidence from the records

  • “"United Kingdom"”
  • “AI agents for knowledge management / corporate training”
  • “Six anonymized use cases (global school chain, mining, telecom, pharma, software publisher, training provider)”
Intro path
Founder-to-founder intro via Ocuwell Limited (network portfolio company of Northern Gritstone).
First step
Check Northern England link.

Investment analysis

4.5D

Grade D

Rank 63 of 151

Disruption5.0/10
Business model3.0/10
Market6.0/10
Risk(lower is better)7.0/10

Global score = mean of disruption, model, market and (11 − risk). Last updated 21 Nov 2026.

Too early.

Disruption

5/10

Expert AI twins for training: innovation in an emerging category.

Business model

3/10

Anonymized use cases, no named customers.

Market

6/10

TAM enterprise knowledge management software ~$14.73bn (base 5), high scalability 8/10 (+1), no proven international expansion (0) → 6/10.

Risk

7/10

Lower is better

High risk: intense competition, anonymized data.

Strengths

  • Varied use cases
  • Clear positioning

Risks

  • Anonymous customers
  • Competition

Funding

$635K, disclosed
  1. 2024-07

    seed / pre-seed (depending on source)~GBP 500K (~$646K)

    Sofia Angels Ventures

Amount used in the study. Round of about £500K converted; range $593-646K depending on sources.

Note. ≈ GBP 500K (Preqin/Dealroom); $593K (Techleap) or ~$646K (VCBacked) – conflicting figures

Regional identified: Ethermind won Startup World Cup Bulgaria (Crossroads 2025, Sofia); company of Bulgarian origin based in the United Kingdom (Techleap). Conflicting funding amount ($593K Techleap vs ~$646K VCBacked).

Team

  • Teodor V. Panayotov

    Founder & CEO

    Bulgarian entrepreneur, speaker on AI and corporate training (Capital.bg, Bloomberg TV Bulgaria).

Sources

Regional final
Bulgaria – Startup World Cup Bulgaria (Crossroads 2025, NDK, Sofia; article dated 21 Oct. 2025), winner
Identification confidence
medium
  1. 1ethermind.ai/
  2. 2preqin.com/data/profile/asset/ethermind-ltd/650808
  3. 3app.dealroom.co/companies/ethermind
  4. 4vcbacked.co/company/ethermind
  5. 5businessnovinite.bg/bg-biznes/ethermind-specheli-startup-world-cup-bulgaria-i-z
  6. 6finder.techleap.nl/news/feed/ethermind-raises-593k-in-funding

From the open study Startup World Cup 2026 by TIBOK, based on public sources, verified in November 2026. Scores are an analyst's judgment on an explicit grid, not the jury's.

Wall

Wall

No posts yet