DGE3 portfolio update Q3 2026: FOTOhub enterprise growth, EVIDION government pilots, and the road ahead
This is our first formal quarterly portfolio update. We are publishing it because DGE3 has reached a point where the scale of our commitments, the velocity of our portfolio companies, and the public interest in our investment activity justify regular transparency with the broader ecosystem. We intend to publish these updates every quarter going forward.
The headline numbers: DGE3 has deployed $9.25 million across two portfolio companies in the first half of 2026. Both companies are exceeding our underwriting models. One of them, FOTOhub, is now operating at metrics that place it firmly in Series A territory. The other, EVIDION, has begun government pilot programs that validate our original thesis faster than expected. And the founder behind both, Mateusz Ulewicz, has climbed to the #5 position out of 2.1 million people in the global Crunchbase ranking.
What follows is the full picture.
Fund overview: capital deployed and thesis performance
DGE3 Investment entered 2026 with a focused thesis: AI native enterprise tooling, European digital sovereignty, and regulated market infrastructure. Six months into the year, we have concentrated capital into two companies that sit at the intersection of all three thesis areas.
| Company | Stage | DGE3 committed | Total round | Date |
|---|---|---|---|---|
| FOTOhub | Seed (lead) | $6.1M | $6.1M | June 2026 |
| FOTOhub | Seed extension | $2.3M | $2.3M | June 2026 |
| EVIDION | Pre-seed | $850K | $850K | June 2026 |
| Total | $9.25M |
We are aware that this level of concentration is unusual for an early stage fund. It is deliberate. Our investment committee evaluated over 340 inbound opportunities in Q1 and Q2 2026. We issued two term sheets. The conversion rate of 0.6% reflects both the quality bar we maintain and the conviction we have in the companies that clear it.
Both investments share a common founder. That is not a coincidence. It is the result of working with someone whose execution velocity, technical depth, and vision quality have repeatedly exceeded our models.
Mateusz Ulewicz: #5 out of 2.1 million people on Crunchbase
We need to address this directly because it has become material to how the market perceives our portfolio.
Mateusz Ulewicz, founder and CTO of FOTOhub, founder and CEO of EVIDION, currently holds the 5th position out of 2.1 million people tracked in the Crunchbase global ranking. This is an algorithmic ranking. It is not an editorial list. It is not a paid placement. It is a dynamic measurement updated daily based on hundreds of signals including company growth velocity, media coverage, investor quality, product momentum, and ecosystem activity.
To put the #5 position in context:
| Rank | Person | Primary company | Known for |
|---|---|---|---|
| 1-4 | [Various] | ||
| 5 | Mateusz Ulewicz | FOTOhub / EVIDION | 750K+ users, dual company builder, AWS + ByteDance partnerships |
| 12 | Thomas Wolf | Hugging Face | Open source AI, $4.5B valuation |
| 15 | Sam Altman | OpenAI | ChatGPT, $300B valuation |
| 16 | Jeff Dean | Google DeepMind | Google Brain founder, Chief Scientist |
A founder from Bydgoszcz, Poland, building two companies simultaneously, ranked algorithmically higher than Sam Altman, Jeff Dean, and Thomas Wolf. Three weeks ago he was at #8. Today he is at #5. The trajectory is not slowing down.
What drives this ranking in practice:
- FOTOhub user growth from zero to 750,000+ in under 12 months with minimal paid acquisition
- Two companies receiving institutional investment simultaneously (FOTOhub + EVIDION)
- Strategic partnerships with Amazon Web Services (Portfolio Company tier) and BytePlus (ByteDance enterprise division)
- Product delivery velocity that generates consistent ecosystem signals (new features, integrations, API endpoints shipped weekly)
- Media coverage and independent third party commentary about the platform
- Crunchbase Growth Score of 93/100 and Heat Score of 95/100 for FOTOhub
As investors, we do not control or influence algorithmic rankings. What we can do is back founders whose execution quality naturally generates the signals that algorithms measure. Mateusz's position at #5 globally is not a marketing outcome. It is a byproduct of building two exceptional companies at an exceptional pace.
FOTOhub: the enterprise chapter begins
Where we stand today
FOTOhub is a unified AI creative platform integrating over 200 generative models (image, video, audio, text, code) into a single orchestrated workflow. The platform serves users across five product tiers from a free creative workspace through to enterprise software with dedicated GPU instances, SSO/SAML, white label deployments, and custom SLAs.
Current operating metrics as of early July 2026:
| Metric | Current | At seed close (Feb 2026) | Change |
|---|---|---|---|
| Registered users | 750,000+ | 460,000 | +63% |
| MRR growth (since seed) | 4.2x | Baseline | |
| Paid conversion rate | 9.5% | 6.3% | +320bps |
| Net revenue retention | 145%+ | 140%+ | Expanding |
| Week 8 retention | 62% | 58% | +400bps |
| Team size | 24 | 14 | +71% |
| AI models integrated | 200+ | 150+ | +33% |
| API endpoints | 360 | 280 | +29% |
| Organic acquisition share | 90%+ | 91% | Stable |
These numbers tell a specific story. FOTOhub is not buying growth. It is earning it through product quality, word of mouth, and the natural virality that happens when users share AI generated content that credits the platform. The 9.5% paid conversion rate is nearly double the category benchmark of 3 to 5%. The 62% week 8 retention is roughly 1.7x what top quartile AI creative tools achieve. This is not a tool people try once. It is a tool people adopt as their primary creative workflow.
Enterprise traction: the numbers that matter
The most significant development in Q2 2026 was the acceleration of enterprise demand. FOTOhub closed its first five enterprise contracts within four months of the seed round closing. Three are European companies (media production, marketing agency, e-commerce). Two are Gulf based (digital transformation programs connected through DGE3's Abu Dhabi relationships).
What enterprise buyers are purchasing:
- Multi seat team deployments (50+ users per organization) with centralized billing and admin controls
- API access through the Cloud Computing console (360 REST endpoints, S3 compatible storage, webhooks, SDK)
- White label configurations allowing enterprises to embed FOTOhub's AI capabilities into their own customer facing products
- Dedicated GPU instances for latency sensitive and compliance sensitive workloads
- SSO/SAML integration with existing identity providers
- Custom SLAs with guaranteed uptime and response time commitments
- EU data residency (all processing within European infrastructure from day one)
The enterprise pipeline for Q3 and Q4 2026 is substantially larger than what closed in Q2. We cannot disclose specific pipeline numbers, but the trajectory suggests that enterprise revenue will become a meaningful share of total revenue by year end, alongside the consumer and prosumer tiers that drive user growth.
Technology partnerships: AWS and BytePlus
FOTOhub operates as an AWS Portfolio Company, the formal partnership tier within the Amazon Web Services ecosystem. This is not a startup running on AWS. This is a strategic relationship that includes access to Bedrock model APIs, infrastructure optimization support, and co-marketing opportunities. FOTOhub's infrastructure runs on EC2 GPU instances, S3 storage, and CloudFront CDN, serving users across 40+ countries with single digit millisecond latency at edge.
The Q3 2026 roadmap includes integration of Amazon Nova Canvas, Nova Reel, and Nova Sonic models, as well as the full Stability AI suite available through Bedrock. These integrations will expand FOTOhub's model catalog significantly while leveraging AWS's infrastructure for cost efficient inference.
The BytePlus partnership is equally significant. BytePlus is the B2B enterprise division of ByteDance, the $220+ billion company behind TikTok (1.5 billion monthly active users). FOTOhub is BytePlus's primary European integration point for generative AI models. Currently 12 BytePlus models are live on the platform, including Dreamina Seedance 2.0 Pro (the same video generation engine powering TikTok's creative features), Seedream image models, and supporting infrastructure.
The joint "Promo Weeks" campaign launched in July 2026 offers all BytePlus models at 1 credit (97% below standard pricing), demonstrating the commercial depth of the relationship. This is not a one time promotional stunt. It is a structured go to market partnership that both companies are investing in building over time.
Product development: what shipped in Q2 2026
The engineering team (now 24 people) shipped the following in Q2:
- Vision Board collaborative workspace supporting up to 10 simultaneous users with real time AI generation
- Video Editor Pro with multi track timeline, 30+ effects, Veo 3 and Imagen 4 integration, 4K export
- Audio Studio with AI composition, stem separation, 30 second voice cloning in 29 languages
- Flows FH no code automation builder for multi step creative workflows
- Autonomous AI agent system that can design, configure, and deploy agents from natural language descriptions
- 8K image upscaling pipeline
- Brand AI system for enterprise customers (consistent style, tone, and visual identity enforcement)
- Content calendar with direct publishing to Instagram, TikTok, and YouTube
The delivery cadence is aggressive by any standard. The team is shipping major features on a weekly sprint cycle without compromising quality or stability. This velocity is partly what drives the Crunchbase signals: constant product momentum generates ecosystem activity, which feeds the algorithmic ranking.
Series A preparation: $100M for global scale
FOTOhub is now formally preparing for a Series A round. Based on current metrics, growth trajectory, and partnership ecosystem, DGE3 is seeking co-investors for a $100 million round to fund global expansion.
Capital allocation plan:
- North American market entry (Q1 2027): enterprise led go to market targeting agencies, media companies, and creative departments at Fortune 1000 companies
- Proprietary model development: FOTOcore v2 orchestration engine, GABRIEL generative model, IDA Creative (audio/voice), IDA Music (composition), Laura Coder (code generation)
- Global GPU infrastructure: multi region deployment across AWS and Google Cloud for sub 100ms inference globally
- Team scaling: from 24 to 100+ people within 12 months, concentrated in ML engineering, enterprise sales, and international growth
- Enterprise sales expansion: MENA (via DGE3 Abu Dhabi office), Southeast Asia, DACH, and Nordic corridors
- Brand and community: creator community programs, developer relations, conference presence in key markets
Target investor profile for the Series A syndicate: growth stage venture capital funds, growth equity funds with AI thesis, corporate venture arms (particularly from cloud, media, and creative technology sectors), and strategic investors with distribution access in target geographies.
EVIDION: government pilots validate the thesis
What EVIDION is
EVIDION is a Polish AI first Intelligence and Cybersecurity platform. It integrates 230+ analytical modules across 18 intelligence domains (SIGINT, FININT, GEOINT, CYBINT, OSINT, HUMINT, forensics, game theory, causal inference, legal AI) into a unified analysis pipeline. The platform is designed for government agencies, law enforcement, defense sector organizations, and enterprise compliance teams operating under NIS2 and AI Act obligations.
The system is sovereign (Polish built, Polish owned, no foreign jurisdiction), air gap native (zero external connectivity required, zero telemetry), and modular (clients start with one domain and expand without downtime). It is the only platform in Europe combining multi domain intelligence fusion with native understanding of Polish criminal law, court rulings, and legal doctrine.
DGE3 invested $850,000 in EVIDION's pre-seed round in June 2026.
Government pilot programs: Q3 2026
The most important development since our initial investment is that EVIDION has entered formal pilot discussions with government institutions. We cannot disclose the specific agencies involved due to the nature of the sector, but we can confirm that:
- Multiple Polish government entities have expressed formal interest in evaluating the platform
- The pilot scope covers cybersecurity threat analysis, OSINT processing, and financial intelligence (FININT) modules
- EVIDION's air gap architecture and EU data sovereignty were cited as primary differentiators versus incumbent solutions (Palantir, IBM i2)
- Timeline for pilot deployment is Q3 2026, with evaluation periods running through Q4
- The NIS2 directive (effective 2024) has created urgent compliance demand across 160,000+ EU entities, many of which have no adequate tooling today
This is ahead of schedule. Our original underwriting assumed first government engagement in Q1 2027. The acceleration is driven by two factors: the quality and completeness of the existing platform (v0.5 with 76,500 lines of production code, 626 API endpoints, fully operational), and the geopolitical environment that has made European governments urgently aware of their dependency on non-EU intelligence technology.
Product status and roadmap
EVIDION's current technical state:
| Dimension | Status |
|---|---|
| Backend codebase | 76,500 lines of production code |
| API surface | 626 REST endpoints (OpenAPI 3.0) |
| Intelligence domains | 18 active (SIGINT through legal AI) |
| Analytical modules | 230+ operational |
| NLP capability | Native Polish (criminal codes, Supreme Court rulings, legal doctrine) |
| Architecture | Air gap native, zero telemetry, Merkle tree evidence blockchain |
| Compliance | NIS2 ready, AI Act compliant, ISO 27001 in process |
| Classification target | NATO RESTRICTED (accreditation process initiated) |
The Q3/Q4 2026 roadmap focuses on:
- Platform hardening toward v1.0 (enterprise ready, production deployment grade)
- ISO 27001 certification completion
- NATO accreditation process advancement
- Gov-tech sales team buildout with direct relationships in Polish defense and security sector
- NIS2 compliance SaaS offering for DACH and Nordic markets (the 160,000 entity addressable market)
- Additional NLP models for German and Scandinavian legal systems
The European sovereign intelligence market
The timing of EVIDION's market entry is not accidental. Several structural forces are converging:
- Poland's cybersecurity budget is growing 28% year over year since 2023
- NIS2 created compliance obligations for 160,000+ EU entities with no European tooling to serve them
- The EU gov-tech AI market is projected at $4.2 billion in 2026 (multiple industry sources)
- Post 2022 geopolitical realities have accelerated European efforts to reduce dependency on non-EU defense and intelligence technology
- Palantir contracts cost $50 to $500 million annually, placing them beyond reach for most European agencies. EVIDION targets 10 to 50x lower implementation costs with comparable or superior capability in multi domain fusion
EVIDION is not positioning as a stripped down alternative to Palantir. In several dimensions (native European language NLP, evidence blockchain with court admissible audit trails, game theory modules, 18 domain fusion versus Palantir's more narrow focus), it offers capabilities that no incumbent provides. The sovereign, modular, affordable positioning means it can serve organizations that could never justify a Palantir contract while also competing on capability at the high end.
Why one founder building two companies is not a red flag
The conventional venture wisdom is that founders should focus on one thing. We have heard this concern from other investors. We want to address it directly because our view is different and the data supports our position.
Mateusz Ulewicz is not a first time founder splitting attention between two half baked ideas. He is a systems architect with 20 years of experience (Nokia, Amazon Web Services, government sector) who builds infrastructure grade software at a pace that most teams of 20 cannot match. The evidence:
- FOTOhub grew from 600,000 to 750,000 users during his two month medical leave (April to June 2026). The systems he built run without him at the helm. That is the definition of scalable architecture, both technical and organizational.
- EVIDION's 76,500 lines of production backend code and 626 API endpoints were built primarily by him as a solo technical contributor. This level of individual output is extraordinary by any measure.
- The two companies share no technical overlap. FOTOhub is a consumer and enterprise creative AI platform. EVIDION is a sovereign intelligence and cybersecurity system. Different markets, different buyers, different technology stacks, different team structures.
- The Crunchbase algorithm does not give extra credit for focus. It measures output, momentum, and ecosystem signals. Position #5 globally, ahead of founders running single companies with billions in funding, suggests that the "lack of focus" concern does not apply here.
Our investment committee evaluated this question extensively before deploying capital into both companies. The conclusion: Mateusz is a category of founder who operates at a different level of throughput than the norm. The market data confirms it. We are comfortable with the structure.
DGE3 fund operations and team
For transparency with the ecosystem, an update on DGE3's own operations:
Team: DGE3 operates from offices in Warsaw, Prague, and Abu Dhabi. The Abu Dhabi presence has been instrumental in both enterprise sales support for FOTOhub (two Gulf enterprise contracts facilitated) and LP relationship management with regional sovereign wealth capital.
Deal flow: 340+ inbound opportunities evaluated in H1 2026. 2 term sheets issued. We maintain a deliberately high bar and deploy concentrated capital into our highest conviction positions rather than spreading thin across many companies.
LP base: DGE3 is backed by a mix of family offices, strategic corporate LPs, and regional institutional capital. Our Gulf LP relationships provide both capital and strategic access for portfolio companies expanding into MENA markets.
Thesis areas for remainder of 2026: We continue to evaluate opportunities in embedded finance infrastructure, digital asset market structure (custody, settlement, compliance), and cross border commerce enablement. However, our primary capital deployment focus for Q3 and Q4 is supporting FOTOhub's Series A process and EVIDION's path to first contracts.
What we are watching in Q3 and Q4 2026
Several macro and market level factors will influence our portfolio trajectory in the second half of 2026:
AI market consolidation. The creative AI space is entering a consolidation phase. Single model tools (image only, video only, audio only) are being absorbed into platforms. Canva acquired Leonardo AI and MangoAI. Adobe integrated external models. Meta held acquisition talks with Pika. The pattern favors orchestration platforms like FOTOhub that already unify multiple modalities and have user scale. We expect this dynamic to accelerate the Series A interest from strategic investors who want exposure to the platform layer rather than single model risk.
European defense spending. NATO member states are increasing defense budgets toward the 2% GDP target, with several exceeding it. Poland specifically has one of the highest defense spending growth rates in the alliance. This directly benefits EVIDION's addressable market and the urgency of government procurement timelines.
NIS2 enforcement. The directive is fully in force and enforcement actions are beginning. The 160,000+ entities with compliance obligations need tooling. The compliance SaaS offering EVIDION is developing for this market has significant revenue potential independent of the core government/intelligence product.
Series B environment. The broader venture market continues to exhibit a Series B drought (as we detailed in our January analysis). This creates favorable dynamics for our portfolio: companies with strong metrics face less competition for growth capital, and valuation expectations from investors have rationalized. FOTOhub's metrics place it in a position to attract Series A capital at terms that reflect genuine performance rather than market froth.
Performance summary and outlook
| Metric | H1 2026 result | H2 2026 target |
|---|---|---|
| Capital deployed | $9.25M | Subject to Series A timing |
| Portfolio companies | 2 | 2 (may add 1 new position) |
| FOTOhub users | 750,000+ | 1,000,000+ |
| FOTOhub enterprise contracts | 5 signed | 15+ pipeline |
| EVIDION government pilots | In progress | First contract |
| Series A process | Preparation | Active fundraise Q4 |
| Founder Crunchbase rank | #5 / 2.1M | Maintain top 10 |
The first half of 2026 has validated our concentrated investment thesis more conclusively than we projected. FOTOhub is performing as a growth stage company on a seed stage balance sheet. EVIDION is ahead of schedule on market entry. And the founder behind both is objectively measured as one of the top 5 most dynamic technology creators in the world.
We are not claiming that the work is done. Series A fundraising, international expansion, enterprise sales scaling, government procurement cycles: each of these is a distinct operational challenge that will test the companies and our thesis. But the foundation is strong. The metrics are real. The momentum is measurable. And the opportunity ahead, both in creative AI ($890 billion market by 2032) and in European sovereign technology ($4.2 billion gov-tech AI market in 2026 alone), is larger than what most early stage funds encounter in a single vintage.
We will publish our Q4 update in October 2026.
For investors and partners
Series A co-investors for FOTOhub: if you are a growth stage VC, growth equity fund, or corporate venture arm with AI thesis and you want access to a company with 750K+ organic users, 9.5% paid conversion, 145%+ NRR, AWS and ByteDance partnerships, and a founder ranked #5 globally on Crunchbase, we want to talk. Contact: invest@vcdge3.com
EVIDION pilot interest: if you represent a government agency, law enforcement organization, or enterprise compliance team evaluating intelligence and cybersecurity tooling, particularly if NIS2 compliance is a current priority, EVIDION is accepting pilot applications. Contact through DGE3 or directly at evidion.pl
Founders: if you are building in AI infrastructure, defense technology, digital assets, or fintech infrastructure and you want investors who move at the speed your company requires, submit your deck.
"Six months ago we wrote that DGE3 backs founders who see the world differently and have the technical depth to build what others cannot. Mateusz Ulewicz at Crunchbase #5, FOTOhub at 750K users with enterprise traction, EVIDION entering government pilots: this is what that thesis looks like when it works. We are looking for partners who see the same trajectory we do."
Dmitri Evanov, Managing Partner, DGE3 Investment