Filmustage

🚨 Disclaimer

All research on the startup and the whole investment memo has been done for personal use of Akim Arhipov and is not investment advice.

General Overview

Filmustage is an AI pre-production platform for film and TV: upload a screenplay and the system breaks it down, schedules and budgets the shoot, and is now extending the same data model into storyboards and AI-generated video ("screenplay to screen"). Positioning: autopilot or full control — the agent can take a script all the way to video, or a director can step in at any point and change character, costume, location, props or camera in a persistent 3D scene.

Quick facts:

• Founded 2020, Delaware C-corp (HQ Santa Monica / Wilmington), team of ~10 (2 founders + 8)
• B2C live since 2020, B2B product launched 2024, focus shifted to enterprise in 2025–2026
• $360K trailing-12-month gross revenue (Stripe, Nov 2025–Oct 2026); $328K ARR on an MRR basis at Oct 31, 2026; ~410 paid subscribers
• 17 tier-1 studios engaged, 4 in contract negotiation, 0 enterprise contracts signed yet
• SOC 2 Type II + TPN Blue Shield certified (the entry ticket to studio scripts)
• $1.965M raised to date across 20 SAFEs (caps $1.5M → $3M → $5M → $7M); now raising up to $500k Pre-Series A SAFE at $7M post-money cap

Why now

AI has reached the film set and Hollywood is buying — but pre-production is where the industry has decided AI is acceptable (no actor likeness issues, no union conflict, pure planning). Signals from the deck (company-cited, Oct 2026):

• Netflix used generative AI on ~300 titles in 2026; reportedly paid ~$587M for InterPositive (AI filmmaking)
• Google reportedly put ~$75M into A24 for AI (June 2026)
• Higgsfield: $700M annualised revenue, $5.4B valuation (Aug 2026); Runway: $200M ARR, $5.3B valuation; Kling: ~$500M ARR
• Preview: $12M raised, seed led by Sequoia (Aug 2026) — Sequoia calls it "Cursor for video"
• Other recent entrants: Flick ($6M seed, True Ventures/GV/YC), Mito AI ($4.5M pre-seed, Lightspeed/Sequoia scout), Wonder Studios ($12M)

Problem

Video models can generate a shot. They can't make a film.

• AI video starts from a prompt, not a screenplay
• Characters, locations and wardrobe drift from shot to shot — no persistent 3D space to block a scene in
• Productions still plan in 11 disconnected tools from the 1980s–90s (Movie Magic, StudioBinder, spreadsheets)
• Nobody sees the film, or its real cost, before the shoot. A lost day on set costs $50K–$500K+. Example from Filmustage's own blog: on Rampage one unplanned close-up would have added ~$60K to VFX.

Solution & Product

Idea or script in. Film out. The pipeline: Idea → Screenplay → Breakdown → Storyboards → Video — all on one data model.

1. Pre-production core (live, revenue-generating since 2020)

• Twelve modules on one data model: breakdown, schedule, budget, storyboards, call sheets
• Rewrite the script → breakdown, schedule, budget and previz update together
• Exports to Movie Magic, Gorilla Budgeting etc.

2. AI Agent in pre-production (live)

• Writes a screenplay from an idea; edits live breakdown/schedule/budget with preview before every change
• Works across the studio's whole slate, catches conflicts across cast, locations and crew hours; voice input in any language
• Next: MCP, Slack and Microsoft Teams integration ("the agent goes where the crew works")

3. Filmustage Studio — screenplay to screen (beta, release Q4 2026)

Business model: one engine, two worlds

Pricing (self-serve): Free plan (core non-AI tools free) + paid plans (Professional, Studio; "Director's Cut" per current help centre); AI features powered by an internal credit currency; a $999 Lifetime Deal sold since Oct 2026. ARPU has climbed from ~$38 to ~$67/month over two years as plans were repriced. Self-serve is a ~$30K MRR business; the enterprise motion is what has to work.

How a studio account grows: 1 department (~10 seats, first contract) → 4 departments (engaged at one major studio today) → studio-wide seats + generation usage on top. First tier-1 contract targeted at $100K ARR.

Traction & Validations

Launched: B2C 2020, B2B 2024

ARR: $360K trailing-12-month gross revenue / $328K MRR-based ARR (Stripe, Oct 2026) — see the metrics section below

Studio pipeline (names under NDA, from deck appendix):

• Paying corporate subscriptions: Leonine Studios (Germany's #1 by revenue), Netflix Thailand & Netflix Taiwan (co-developed Asian-language support), HBO
• Contracts in progress: Major US studio — 1 department contract in progress, 4 departments engaged; another major US studio — full security review passed, contract in progress
• Paid pilots ended, next steps in progress: Major US studio + its animation arm (2 contracts in progress), US film studio, Hollywood film studio, US media group, Big Tech studio, streaming studio, UK public broadcaster, UK TV network
• Plus 4 more tier-1 studios in early talks

Academic channel: 10+ institutions with Filmustage in the syllabus, 30+ instructor licences, inbound (near-zero CAC). UNCSA School of Filmmaking, Mater Dei High School (77 students, 10 films), Storytellers Film Academy — all paid academic subscriptions.

Industry presence: exhibited at the Directors Guild's AI event in Atlanta (2026); testimonials from Roger Christian (Oscar winner, set decorator, Star Wars) and Callan Green (DoP, The Gentlemen / Netflix).

Financial & customer metrics (data room, Stripe, Nov 2024 – Oct 2026)

Two-year summary

Month by month (gross = Stripe charges; net = after refunds, disputes, fees; MRR = Stripe Billing at month end, annual plans at 1/12)

How to read the curve

• 2025 was the growth year for self-serve: MRR went from $17K to $30K (+69%), driven by new subscribers (peak 558 in June 2025) and a mid-2025 repricing that lifted ARPU from ~$42 to ~$64.
• Nov–Dec 2025 spike ($46K / $39K gross) is annual plans charged in Q4 ($28.8K and $22K of annual revenue in those two months, likely a Black Friday push), not a step change in run-rate — MRR stayed flat at ~$30K.
• 2026 is flat-to-down on self-serve: MRR has drifted from $30.9K (Feb) to $27.3K (Oct), subscribers from 558 (peak) to 410, and gross revenue in Aug–Oct 2026 is below the same months a year earlier (−12%, −6%, −4% YoY).
• Company's explanation: from late 2025 the team deliberately moved focus and selling effort from the B2C channel to enterprise — security certification, pilots and contract negotiations with large studios and publishers — so acquisition spend and product attention on self-serve dropped. The data is consistent with that story (new paying customers per month fell from 85–95 in spring 2026 to 42–52 by autumn), but the enterprise revenue that is supposed to replace it is not yet visible in the numbers. That is the central thing to diligence.

Retention and churn (Stripe Billing)

• Monthly subscriber churn runs at 12–18%; gross MRR churn 17–32%. Over the last 12 months the company added $60K of new MRR and lost $80K to churn.
• Cohort retention (average of monthly cohorts started Nov 2025–Oct 2026): subscribers retained M3 = 40%, M6 = 23%, M12 = 11%; MRR retained M12 = 10%.
• This is normal for a prosumer tool used per production (filmmakers subscribe for a project, then leave), and the data room shows 270 customers who came back after a gap of 2+ months (median gap 7 months) — but it means self-serve is a treadmill, not compounding SaaS.

Concentration: top 10 customers = 7.2% of revenue; no dependency on any single self-serve account.

Debt: Filmustage uses Stripe Capital revenue-based loans — eight loans since 2021, ~$155K principal in total. Outstanding at Oct 31, 2026: $43,501 (three loans taken Jul–Oct 2026), repaid automatically as 22.5% of Stripe sales. Repayments took $49.7K out of Stripe cash flow in the last 12 months. Taking three loans in the four months before a fundraise suggests cash was tight going into this round.

Team

Founders

• Egor Dubrovsky — CEO. 10+ years in film industry; founded two media productions; partnered with Uber, TikTok and the UN.
• Andrei Karalkou — Co-founder, CTO & COO. 10+ years in development and engineering management; previously at Intel in California.
• team of 8 across engineering, marketing and sales.

Advisors

• Barry Loudis — media executive, former VP at WarnerMedia; 20+ years at WarnerMedia and CBS
• Joe Jarvis — Senior VP at Final Draft (industry-standard screenwriting software) — a potential distribution/partnership channel
• Byron Saltysiak — media executive, former VP at WarnerMedia

Former third co-founder: Ruslan Khamidullin (CTO / Chief Technology Innovation Officer, board member) left on December 31, 2024. A separation agreement and general release was signed in July 2025 (amended July 2026). See the cap table below for the share treatment.

Assessment: founder-market fit is credible (film operator + strong engineer), the advisory board is Hollywood-native, and the team has survived 6 years on under $2M — capital-efficient. Weakness: a 10-person team is now building three products (pre-pro SaaS, agent, AI video/3D engine) and selling to two customer types. No obvious enterprise-sales leader on the team.

Competitors

From the deck (public product information, Oct 2026):

* Filmustage video and 3D camera: in final testing, release Q4 2026.

What puts Filmustage apart: the only product where every generated shot is grounded in real production data and stays in sync with the schedule and budget. What doesn't: on the generation side itself they are a consumer of other people's models, late (Q4 2026 release vs. competitors already shipping), and outspent.

Investments so far

Stage: Seed → raising Pre-Series A

Raised to date: $1,965,000 across 20 post-money SAFEs (Carta, Nov 2, 2026). No priced round yet, so no preferred stock exists and none of the SAFEs has converted. Carta buckets: Pre-Seed SAFEs $95K · Seed SAFEs $690K · Extension Seed SAFEs $1,180K.

Investors, amounts and terms (from signed SAFEs and the Raw Ventures term sheet):

Cap table (Carta, fully diluted, Nov 2, 2026 — before SAFE conversion)

Departed co-founder's shares: under the July 2025 separation agreement, 500,000 of Khamidullin's shares (5.0%) stay fully vested; the remaining 2,113,750 shares (21.1%) are subject to an unconditional company repurchase option at par ($0.00001/share), in exchange for "special payments" to him totalling $75,600 in instalments. The July 2026 amendment extended the deadline to August 1, 2027. The repurchase had not been exercised as of the Carta export. Once it is, the active founders move from 58.9% to ~74.6% FD and the overhang disappears — but the company has to find ~$76K of cash for it before Aug 2027, and the shares sit on the cap table until then.

Pro-forma on conversion (rough): SAFEs at the $5M cap convert to ~18% (915K / 5M) and at the $7M cap to ~15% (1.05M / 7M) of post-money; this round at the $10M cap adds up to 10%. So at a priced Series A roughly a third to 40% of the company would sit with SAFE holders before the Series A investor's own stake, with the founders holding the balance net of the pool — within the normal range if the Khamidullin repurchase goes through, tight if it does not.

Corporate, security & IP (data room)

• Entity: Filmustage, Inc., Delaware C-corporation, incorporated February 6, 2020 (certificate and EIN on file); bylaws, board and stockholder consents, founder restricted stock purchase agreements and CIIAs dated February 2022.
• Equity plan: 2022 Stock Plan, 1,500,000 shares; 16 option grants with board consents on file, including advisor grants to Loudis, Saltysiak and Jarvis (Sep 2026).
• Security: SOC 2 Type II — report dated October 6, 2025 and a second report for 2026; TPN (Trusted Partner Network) Best Practice Questionnaire completed Sep 2025 (Blue Shield status); third-party penetration test report on file; cyber-insurance certificate for 2026 (carriers: Chubb, The Hartford). For a company selling into studios this is the moat that is actually documented.
• IP: USPTO trademark registration for "Filmustage"; invention-assignment agreements from founders and all contractors on file. Team is engaged on consulting agreements + CIIAs rather than employment.
• Investor rights in place: Raw Ventures board observer and spend-approval right (> $100K) until SAFE conversion; Geek Ventures management-rights letter; pro-rata letters for Raw and Geek.

The Ask

Round: Pre-Series A, up to $1,000,000

Instrument: Post-money SAFE

Valuation cap: $7,000,000 post-money

Discount / MFN / lead: not stated in deck — request

Timing: "invest before the screenplay to screen release" (Q4 2026)

Use of funds:

1. Release screenplay-to-screen and scale generation (GPU and model costs)
2. Launch generation for AI content studios (consumption pricing already live)
3. Close studio contracts (17 tier-1 studios engaged)

Stated next milestones: first tier-1 studio contract signed ($100K ARR) → screenplay-to-screen release (autopilot + 3D camera) → studio contracts at scale → agent inside Slack/Teams → AI video for ad agencies, vertical drama and AI studios → multi-language.

Implied valuation math: $10M cap on $328K MRR-based ARR = ~30x ARR (28x on trailing gross revenue). Rich for a pre-pro SaaS; cheap if you believe the AI-video story (Preview at $40–60M EV on no disclosed revenue). The price is essentially a bet on the Q4 release landing and one or two studio logos converting to $100K contracts in the next 12 months.

Burn / runway: not disclosed in deck or data room (report headings present, tables empty) — request. Known: $43.5K of Stripe Capital debt outstanding, repaid at 22.5% of Stripe sales; three loans taken Jul–Oct 2026.

Key risks

1. Product not shipped. The round is sold on screenplay-to-screen; it is in closed beta with no public release date. Slips are the norm for generative pipelines.
2. Enterprise conversion unproven. Six years in, enterprise ARR is ~0. Studio sales cycles are long and security reviews are only the start.
3. Two customers, one small team. Studios want control and planning; AI creators want cheap generation volume. Different sales motions, different roadmaps, 10 people.
4. Gross margin on generation. Pay-per-generation revenue is passed through to GPU/model suppliers (Kling, Runway etc.). Margin structure is unknown.
5. Capital asymmetry. Preview ($12M, Sequoia), Flick ($6M), Mito ($4.5M) are all raising bigger rounds for adjacent products. $1M buys ~12 months at a 10-person burn.
6. Platform risk from both sides. Model companies can move up into planning; Entertainment Partners (Movie Magic) can add AI breakdown down into Filmustage's core.
7. Small round, no lead signalled. SAFE up to $1M with no named lead — check whether Raw Ventures / Geek are following.
8. Self-serve is churning, not compounding. 12–18% monthly subscriber churn, M12 retention ~11%, MRR down 7% YoY. The $360K figure is trailing gross revenue, inflated by Q4 annual-plan billing; run-rate is closer to $328K and drifting down.
9. Cash position going into the round. Three revenue-based loans in the four months before raising, with 22.5% of Stripe receipts going to repayment. Runway without this round is probably short.
10. Departed co-founder holds 26% until the repurchase is executed (deadline Aug 2027, ~$76K of payments).
11. Senior terms above the new money. Raw Ventures' $1M carries a 1.5x non-participating liquidation preference. What’s here for new investors?

Questions for founders (before committing)

• Enterprise contracts paid outside Stripe: list, annual value, term — for Leonine, Netflix APAC, HBO and any others. What exactly do they use?
• Opex, net burn and cash balance by month (the Mercury tables missing from the report); runway today and after $1M; headcount plan
• Gross margin including AI inference and GCP (the PostHog / GCP tables missing from the report); GPU/model cost per minute of generated video; which models are under the hood
• Why three Stripe Capital loans in Jul–Oct 2026, and will the 22.5% withholding be cleared from round proceeds?
• CAC on self-serve and what was deliberately cut when focus moved to enterprise; is the plan to let self-serve coast?
• Who is leading the round? Are Raw Ventures (pro-rata), Geek Ventures and Flyer One participating? Is the new SAFE plain YC post-money with no liquidation preference or side letters?
• Timing and funding of the Khamidullin share repurchase
• Realistic date and definition of the "screenplay to screen" release; how many beta users and what retention
• Which of the 4 studios in contract negotiation is closest to signature and at what ACV
• Series A plan: target metrics and timing