• Skip to main content
  • Skip to secondary menu
  • Skip to footer
information matters logo

Information Matters - Agentic AI News and Market Forecasts

The Agentic AI Revolution: what it means for business and the rules of competition

  • Home
  • About
    • The Team
    • About Us
    • Our Methodology
  • Contact
  • Subscribe
  • Downloads
  • Agentic AI Company Tracker
  • Agentic AI Sector Analysis

Black Forest Labs

COMPANY PAGE

Black Forest Labs

German frontier image-and-video generative foundation-model lab — the FLUX family of text-to-image and text-to-video models distributed through a hosted BFL API, partial open-weights on Hugging Face, and inference-marketplace partners (Fal.ai, Replicate, Together AI) — founded by the original Stable Diffusion research team and positioned as the European counterweight to US frontier image-and-video labs.

Founded 2024
Private — Series B
Foundation Model Provider
bfl.ai

Last Updated: 28 May 2026
Fact-checked: 2 June 2026
Coverage: Tracker · Category Report #IM109
← Back to AI Tracker

The Business

Black Forest Labs is a German frontier image-and-video generative foundation-model lab — the maker of the FLUX family of text-to-image and text-to-video models, distributed through three main surfaces: the hosted BFL API (developer and enterprise inference), partial open-weights releases on Hugging Face (the canonical open-weights distribution surface for FLUX components, with 400M+ cumulative downloads by the FLUX 2 release cycle), and inference-marketplace partners (Fal.ai, Replicate, Together AI) that host FLUX endpoints alongside other generative-image models. The company was founded in 2024 by Robin Rombach, Andreas Blattmann and Patrick Esser — co-authors of the original Latent Diffusion / Stable Diffusion research at Stability AI and the single most-cited generative-image research lineage in the industry. BFL is headquartered in Freiburg, Germany (the Black Forest region the company is named for) and is the only European frontier image-model lab at this capital base in the FMP cohort. The product family covers text-to-image generation (FLUX.1 and successor releases), text-to-video generation (the extension that anchored the FLUX 2 release cycle), and a partial open-weights / partial hosted-API distribution model that is itself the brand differentiator versus closed-model rivals like Midjourney and OpenAI Sora.

Customers and Distribution

Black Forest Labs has assembled the most extensive platform-partner stack of any independent image-model lab in the FMP cohort: named integrations include Meta (multi-year contract reported at approximately $140M in named-press coverage), Adobe, Canva, Snap, xAI (Grok image generation), Microsoft, Picsart, ElevenLabs, VSCO and Vercel. The collective contracted-revenue stack from the Meta contract plus the Adobe, Canva and Snap deals reported in named press is in the hundreds of millions of dollars against a sub-100 headcount at the December 2025 Series B announcement. Distribution sits across three channels: the BFL hosted API (developer and enterprise direct), the Hugging Face open-weights distribution surface (400M+ cumulative downloads of FLUX open-source components by the FLUX 2 release cycle), and inference-marketplace partners (Fal.ai, Replicate, Together AI) that host FLUX endpoints. BFL does not publish a clean annualised run-rate revenue figure and the IM tracker labels the headline as “contracted-revenue stack (not recurring ARR)” rather than ARR; we decline-to-publish a precise stand-alone ARR number pending a primary disclosure.

Model Strategy

FLUX is a Frontier-first generative-image-and-video play with an open-weights / hosted-API hybrid distribution model: the strategic bet is that frontier text-to-image and text-to-video capability, shipped in part as open weights on Hugging Face and in part as a hosted API alongside inference-marketplace partners, beats closed-only rivals on enterprise integration and developer mindshare. The infrastructure backstop is the NVIDIA strategic investment that landed in the December 2025 Series B alongside multi-cloud inference partners — supplier alignment with the primary GPU vendor combined with distribution across Fal.ai, Replicate and Together AI rather than a single hyperscaler relationship. The open-weights design itself is intentional and lineage-driven: Rombach, Blattmann and Esser authored the Latent Diffusion research that became Stable Diffusion at Stability AI, and BFL was founded explicitly to extend that open-weights research tradition into a commercially-funded frontier lab. The trade-off is structural customer portability — API customers can self-host the open-weights variants at low switching cost, and the strategic bet relies on continued capability cadence and platform-partner contract retention rather than lock-in.

At A Glance

Annualised revenue
●
2026-04-30 as-of

API tokens
●
2026-03-31 as-of

2024-12-312025-12-31

Headcount
55 ●
2026-04-30 as-of

2024-12-312026-04-30

Funding to date
$450M ●
2026-03-31 as-of

2024-12-312026-03-31

The Numbers

API tokens

2.2B 300M 2024-12-31 — 0.3 2025-06-30 — 1.0 2025-12-31 — 2.2 2024-12-31 2025-12-31

Headcount (FTE)

55 20 2024-12-31 — 20 2025-06-30 — 35 2025-12-31 — 45 2026-04-30 — 55 2024-12-31 2026-04-30

Funding to date

$450M $31M 2024-12-31 — 31 2025-06-30 — 31 2025-12-31 — 231 2026-03-31 — 450 2024-12-31 2026-03-31

Leadership Team

CEO & Co-founder
Robin Rombach
Co-founded Black Forest Labs in 2024 with Andreas Blattmann and Patrick Esser. Previously co-author of the High-Resolution Image Synthesis with Latent Diffusion Models paper that became Stable Diffusion at Stability AI — the single most-cited generative-image research lineage in the industry. Public lead on the FLUX model family and on the Series B announcement; widely viewed as the central figure on European frontier image-model research.

Co-founder & Research
Andreas Blattmann
Co-founder; co-author of the original Latent Diffusion / Stable Diffusion research alongside Rombach and Esser. Anchors model-architecture research on the FLUX family. Continues to publish under the Black Forest Labs research banner.

Co-founder & Research
Patrick Esser
Co-founder; co-author of the original Latent Diffusion / Stable Diffusion research alongside Rombach and Blattmann. Public lead on the open-weights research releases that established the FLUX family’s developer mindshare on Hugging Face.

Black Forest Labs is unusually founder-concentrated for a frontier model lab — the Rombach / Blattmann / Esser trio carries the Stable Diffusion research lineage into the FLUX family, and the company has so far chosen to scale headcount slowly (sub-100 employees at the Series B announcement) rather than build out a conventional executive bench. Operations are headquartered in Freiburg, Germany (Black Forest region) with a growing US commercial presence. CFO, CRO and CTO roles have not been publicly disclosed as separate appointments at the time of writing; Anjney Midha (Andreessen Horowitz general partner) co-led the Series B and sits closest to a board-level commercial voice in public coverage.

IM Framework Scoring

IM’s structured assessment of Black Forest Labs’s competitive position. The summary below is the headline; expand “Show the full analyst-grade analysis” near the bottom for the per-dimension reasoning and evidence. Methodology →

Competitive Position
Disruptive Challenger
Foundation Model Providers sector

The Information Matters Compass

5 7.5 10 5 7.5 10 Defensibility → Disruption Potential →Disruptive Challengers Dominant InnovatorsEmerging Players Established Incumbents Black Forest Labs © Information Matters

Strategic Bet
Frontier capability wins in image and video — FLUX holds the open-weights quality leaderboard against closed-model rivals on text-to-image and is extending to text-to-video
Plus: Plus: vertical reach (Adobe, Canva, Meta, Microsoft, xAI, Picsart, ElevenLabs distribution) converts open-weights quality into a contracted-revenue platform-partner stack at a scale no other independent image-model lab has matched

Watch: The release cadence of FLUX 2 and successor frontier image models against OpenAI Sora and Google Veo 3; the conversion of the platform-partner stack (Meta, Adobe, Canva, Snap, xAI, Microsoft) from contracted to recurring API revenue; the trajectory of the NVIDIA strategic investment into deeper supplier alignment; and the regulatory posture on deepfake and right-of-publicity enforcement in both Germany and the United States — any of these can shift the disruption profile in either direction inside a quarter.

Funding History

Date Round Raised Post-money Lead investor(s)
Dec 2025 Series B $300M $3.25B Salesforce Ventures & Anjney Midha co-led (Andreessen Horowitz, NVIDIA, Northzone, Creandum, Earlybird, BroadLight, General Catalyst, Temasek, Bain Capital Ventures, Air Street, Visionaries Club, Canva, Figma Ventures participating)
2024-2025 Series A (previously undisclosed) ~$119M — Previously undisclosed per Black Forest Labs Series B announcement (Dec 2025)
Aug 2024 Series Seed $31M — Andreessen Horowitz (with General Catalyst, MatchVC, Brendan Iribe, Garry Tan, Michael Ovitz, others)

Cumulative external equity is approximately $450M+ through the December 2025 Series B at a $3.25B post-money valuation, co-led by Salesforce Ventures and Anjney Midha with Andreessen Horowitz, NVIDIA, General Catalyst and Temasek participating. The August 2024 founding round (Series Seed / later characterised as Series A) of $31M was led by Andreessen Horowitz alongside a roster of named angels including General Catalyst, MätchVC, Brendan Iribe, Garry Tan and Michael Ovitz. NVIDIA’s participation as a strategic investor in the Series B is the supplier-alignment milestone Black Forest Labs has most often referenced in commercial coverage. Round dates and lead-investor identities are taken from BFL’s own announcements and corroborated by named-author TechCrunch, VentureBeat and TechFundingNews coverage.

Competitive Landscape

Black Forest Labs’ competitive set sits in three concentric rings: frontier multimodal labs that ship image-and-video as part of a wider model family (OpenAI Sora, Google Imagen / Veo, Adobe Firefly through partnership rather than head-on rivalry), open-weights image specialists (Stability AI — the immediate predecessor research lineage), and closed-API image-and-video pure plays (Midjourney for image, Runway for video, Ideogram for image). BFL is unusual in the set because the competitive frame is open-weights-plus-platform-partner-distribution rather than headline consumer brand — the strategic bet is that frontier image-and-video capability, shipped in part as open weights and in part as a hosted API, beats closed-only rivals on enterprise integration and developer mindshare.

Competitor Positioning Distribution edge Threat profile
Sora
(OpenAI)
Frontier consumer and developer text-to-video generation embedded inside ChatGPT and the OpenAI API; OpenAI also ships GPT image as the closed-model image-generation counterpart. The closest generalist substitute for FLUX on the consumer surface. ChatGPT (900M+ weekly active users on the March 2026 funding-round disclosure), OpenAI API and Microsoft Foundry / Azure OpenAI distribution. High — brand reach and developer mindshare structurally larger than FLUX’s; the open question is whether developer and enterprise creative procurement values open-weights flexibility and per-token cost over OpenAI’s frontier capability bundle.
Stable Diffusion / Stable Video
(Stability AI)
The original open-weights image-and-video model family — founded around the same Latent Diffusion research that the BFL founders authored. Widely used by creative developers and self-hosters; the brand-and-research lineage BFL was founded to succeed. Hugging Face plus direct download; commercial API via Stability AI and via hyperscaler marketplaces. Medium — brand-recognition advantage in the open-weights segment but Stability has not matched FLUX on either commercial-partnership cadence or release cadence since the BFL founders departed.
Midjourney Category-defining consumer brand for high-aesthetic image generation; the brand most often paired against FLUX in creative-professional procurement discussions. Web-app and Discord-native; subscription-only. Web app and Discord-native (legacy); subscription-only consumer pricing. Medium-high — aesthetic-quality lead in consumer-creative segments and no developer-platform overlap with FLUX, but no open-weights story and no platform-partner distribution stack comparable to BFL’s.
Runway Gen-4 Specialist creative-video generation positioned at the film and motion-graphics segment; the most direct vertical competitor to FLUX video. Direct subscription plus a small enterprise sales motion; partnerships with select Hollywood studios. Medium — narrower lane (professional video creation) but stickier inside it; flanking risk on FLUX video specifically as BFL extends from image to video.
Ideogram Independent text-to-image model with strong typography rendering and a consumer-and-developer API; the closest pure-play image-model competitor to FLUX on the closed-API surface. Direct consumer subscription plus a developer API; selective platform integrations. Medium — narrower than FLUX on the platform-partner stack but a credible direct API competitor on capability and price for image-only workloads.

Pricing benchmark: FLUX models are priced per-generation on the BFL API and on inference-marketplace partners (Fal.ai, Replicate, Together AI) in line with Stable Diffusion derivatives and within roughly a 2x band of Ideogram and Midjourney on equivalent quality settings. Open-weights distribution on Hugging Face removes the per-generation cost entirely for self-hosters — the structural pricing differentiator versus closed-model rivals. The competitive frame is therefore open-weights-and-platform-integration rather than headline per-generation price.

Potential Risks

The case for Black Forest Labs at IM Framework 7.76 rests on the FLUX open-weights research lineage carried over from Stable Diffusion, the platform-partner contracted-revenue stack (Meta, Adobe, Canva, Snap, xAI, Microsoft, Picsart, ElevenLabs, VSCO, Vercel), the December 2025 $300M Series B at a $3.25B post-money capital base, and NVIDIA’s participation as a strategic investor. The case against splits into five risks of differing magnitude — with customer portability the most structural (open-weights design is intentional) and regulatory exposure the most active (image-and-video generation carries the highest-active legal-risk profile in the FMP cohort).

Customer portability at the open-weights layer is structural

FLUX models are partially open-weights and distributed on Hugging Face alongside the hosted BFL API — the open-weights design is itself the brand differentiator versus Midjourney and Runway, and the lineage that the BFL founders carried over from Stable Diffusion. The trade-off is that API customers can self-host the open-weights variants at low switching cost, and developer users can substitute FLUX for Stable Diffusion or Ideogram with limited friction at the inference layer. The sub-rubric score on D1c portability was held at 5 on this evidence: substitution risk is real and structural by design, and the strategic bet relies on continued frontier capability cadence and platform-partner contract retention rather than lock-in.

Regulatory and copyright exposure — image-and-video generation has the highest-active legal-risk profile in the FMP cohort

Image-and-video generation models face the most active regulatory-and-litigation environment of any frontier-model category — deepfake legislation in the United States, right-of-publicity claims on synthetic likenesses, copyright suits on training-data composition, and the EU AI Act’s general-purpose-AI obligations binding from a German jurisdiction. None of these is fatal to FLUX, but collectively they cap how aggressively BFL can lean into consumer-facing generative use cases versus enterprise-platform-partner channels. The sub-rubric score on D4c regulatory exposure was held at 4 on this evidence.

Geopolitical exposure — German HQ in a US-dominated FMP cohort

Black Forest Labs is headquartered in Freiburg, Germany — the only European frontier image-model lab at this capital base in the FMP cohort. The bull case is sovereign-AI procurement preference inside the EU and a positioning premium as the named European counterweight to US frontier labs. The bear case is that the lion’s share of frontier-model revenue today still routes through US enterprise procurement and US hyperscaler distribution channels (AWS, Azure, Google Cloud), where US-headquartered rivals have structural distribution advantages. The sub-rubric score on D4f geopolitical exposure was held at 5 on this evidence.

Founder concentration and bench depth at sub-100 headcount

BFL is unusually founder-concentrated — the Rombach / Blattmann / Esser trio carries the Stable Diffusion research lineage that anchors the brand, and the company has so far chosen to scale headcount slowly (sub-100 employees at the Series B announcement) rather than build out a conventional executive bench. No CFO, CRO or CTO has been publicly disclosed as a separate appointment at the time of writing. A coordinated founder departure of the magnitude seen at Stability AI (when the same trio left to found BFL) would be a material disruption to the research roadmap. The sub-rubric score on D4e key-person dependency was held at 5 on this evidence.

Frontier-capability cadence in image-and-video is contested every quarter

FLUX’s strategic bet is open-weights-frontier image-and-video capability sustained release-by-release against OpenAI Sora, Google Imagen / Veo and the closed-API image-and-video specialists. The leaderboard position changes monthly — OpenAI, Google and the closed-API specialists have all shipped frontier image-or-video releases at a cadence BFL needs to match release-by-release, and the open-weights segment itself faces Chinese frontier labs at credible capability. A sustained capability slip against Sora or Veo would compress the pricing power that the FLUX API and the platform-partner stack currently enjoy.

Recent IM Coverage

  • Foundation Model Providers — Category Report May 2026.

Show recent press coverage of Black Forest Labs
  • Dec 2025 — Black Forest Labs raises $300M Series B at $3.25B valuation — Salesforce Ventures and Anjney Midha co-lead; NVIDIA participates as strategic investor.
  • Dec 2025 — Black Forest Labs raises $300M for image and video AI — Series B announcement.
  • Nov 2025 — FLUX 2 release and open-source component download milestones — tens of millions of FLUX cumulative downloads on Hugging Face.
  • 2025 — Black Forest Labs FLUX revenue and platform-partner stack — Meta multi-year contract; Adobe, Canva, Snap deals.
  • Aug 2024 — Black Forest Labs launches with $31M Series Seed led by Andreessen Horowitz — ex-Stability AI founders.
  • Aug 2024 — Black Forest Labs founding announcement — FLUX.1 release on Hugging Face.
  • 2024 — xAI Grok image generation partnership — FLUX integrated into Grok consumer surface.

Curated feed of named-source coverage — BFL’s own newsroom and product announcements, Hugging Face model-release pages, named-author tech and business press. Excludes paywalled article bodies of The Information, WSJ, FT and Bloomberg (headline + free-snippet only), wire-aggregator reposts and unsourced AI round-up pieces.

Show the source register for the figures on this page

IM operates a primary-source-where-possible discipline. The figures above come from:

  • Revenue (basis-disclosure note): Black Forest Labs is private and does not file. The most directly comparable disclosure is the contracted-revenue stack referenced in named-press coverage — a Meta multi-year contract reported at approximately $140M, plus Adobe, Canva and Snap deals reported in the same coverage cycle — collectively a contracted-revenue stack in the hundreds of millions of dollars. We label this “contracted-revenue stack (not recurring ARR)” rather than “annualised run-rate revenue” because the contracts are multi-year and BFL does not publish a clean ARR figure. We decline-to-publish a precise stand-alone ARR number pending a primary disclosure.
  • Usage — FLUX open-source component downloads: BFL referenced 400M+ cumulative downloads of FLUX open-source components on Hugging Face by the FLUX 2 release cycle per VentureBeat’s FLUX 2 coverage, against a platform-partner integration list that includes Adobe, Canva, Meta, Microsoft, Picsart, ElevenLabs, VSCO and Vercel per BFL’s own product page. Cumulative-downloads on Hugging Face is the canonical FLUX open-weights usage disclosure BFL has used since launch.
  • Headcount (basis-disclosure note): Black Forest Labs does not publish precise headcount. Named press around the December 2025 Series B referenced a sub-100 employee base at the announcement. We decline-to-publish a precise figure pending a primary disclosure and reference the sub-100 range only with the caveat that it is press-triangulated, not company-disclosed. The Tech Funding News Series B coverage is the closest published reference.
  • Funding to date: Cumulative external equity approximately $450M+ through the December 2025 Series B at a $3.25B post-money valuation. References: Tech Funding News Series B $300M / $3.25B post-money, co-led by Salesforce Ventures and Anjney Midha with Andreessen Horowitz, NVIDIA, General Catalyst and Temasek participating; TechCrunch Series Seed $31M, August 2024 led by Andreessen Horowitz; Dakota Series B confirmation.

Methodology & Disclaimer

For metric definitions, source-tier hierarchy, and decline-to-publish rules, see the tracker methodology. Confidence dots (• green / • amber / • red) follow the same convention as the AI Tracker.

Spotted a figure you believe is wrong? Send corrections to info@informationmatters.net.

Information Matters Framework scores are the considered opinion of the IM team — human and AI — applied to publicly-available evidence under a disclosed methodology. They are not statements of fact about the companies scored and they are not investment advice.

Footer

  • LinkedIn
  • YouTube

Copyright © 2026 · Information Matters

Manage Consent
To provide the best experiences, we use technologies like cookies to store and/or access device information. Consenting to these technologies will allow us to process data such as browsing behavior or unique IDs on this site. Not consenting or withdrawing consent, may adversely affect certain features and functions.
Functional Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes. The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
  • Manage options
  • Manage services
  • Manage {vendor_count} vendors
  • Read more about these purposes
View preferences
  • {title}
  • {title}
  • {title}