The AI video generation market crossed $3 billion in cumulative startup funding in 2025 alone, up nearly 95% from 2024, according to Crunchbase data. In 2026, that capital has kept moving. New rounds are closing at higher valuations, new infrastructure players are emerging from stealth, and the competitive map has been permanently reshaped by one of the most expensive product failures in recent tech history: OpenAI pulling the plug on Sora in April 2026 after burning $15 million per day against $2.1 million in total lifetime revenue.
What’s left is a leaner, sharper competitive field. Runway, Luma AI, and Synthesia are the dominant funded players by valuation. Pika, Mirage (formerly Captions), and HeyGen are building profitable second-tier businesses. And a new infrastructure layer, represented by Reactor’s May 2026 stealth exit, is now attracting serious capital.
Here is every major funded AI video startup worth tracking in 2026, ranked by last known valuation.
1. Luma AI
- Total Funding: $1.07B
- Latest Valuation: $4B+
- Latest Round: $900M Series C (November 2025)
- Lead Investor: HUMAIN (Saudi Arabia’s Public Investment Fund)
- Founded: 2021
- HQ: San Francisco, CA
- Key Investors: HUMAIN, Andreessen Horowitz, AMD Ventures, Amplify Partners, Matrix Partners, Amazon
Luma AI sits at the top of the funding leaderboard in pure video generation, having closed the largest single round in the category when it raised $900 million in November 2025 at a $4 billion valuation. The round was led by HUMAIN, Saudi Arabia’s sovereign AI company under the Public Investment Fund, with participation from AMD Ventures, Andreessen Horowitz, Amplify Partners, and Matrix Partners. Alongside the funding, Luma and HUMAIN announced Project Halo, a joint plan to build a 2-gigawatt AI supercluster in Saudi Arabia, adding a sovereign compute dimension to what began as a video generation story.
Luma’s flagship product is Dream Machine, a consumer-facing text-to-video and image-to-video platform with over 30 million users. The company’s competitive claim is its Ray3 model, released in September 2025 and positioned as the world’s first reasoning video model. CEO Amit Jain said Ray3 benchmarks above OpenAI’s Sora 2 and comparably with Google’s Veo 3. Luma frames itself not as a video tool but as a multimodal world model company that trains on text, video, audio, and images simultaneously to simulate physical environments.
Total funding now exceeds $1 billion across six rounds. The company’s prior Series C (December 2024) was $90 million and included Amazon. The November 2025 raise was a step-change in both scale and strategic direction, pulling Luma into sovereign AI infrastructure territory.
2. Synthesia
- Total Funding: $530M+
- Latest Valuation: $4B
- Latest Round: $200M Series E (January 26, 2026)
- Lead Investor: Google Ventures (GV)
- Founded: 2017
- HQ: London, UK
- Key Investors: GV, NVentures (NVIDIA), Accel, Kleiner Perkins, NEA, PSP Growth, Air Street Capital, FirstMark, MMC Ventures
Synthesia has become the most funded enterprise AI video company in Europe and the most valuable generative AI media startup in the UK. The London-based company raised $200 million in a Series E round led by Google Ventures on January 26, 2026, nearly doubling its valuation from $2.1 billion to $4 billion in twelve months. The round included NVIDIA’s NVentures arm, Accel, Kleiner Perkins, NEA, PSP Growth, Air Street Capital, FirstMark, MMC Ventures, Evantic, and Hedosophia. Alongside the raise, Synthesia ran an employee secondary sale in partnership with Nasdaq at the same $4 billion valuation.
The company had previously raised $180 million in Series D funding in January 2025. The trajectory tells a consistent story: two consecutive years of near-doubling in valuation, both accompanied by nine-figure rounds. Adobe reportedly held acquisition talks with Synthesia at a $3 billion valuation but walked away over price. The company turned down the offer and closed its Series E instead.
Synthesia builds AI-powered video with synthetic avatars and voice in 140-plus languages, used primarily for enterprise training, onboarding, and internal communications. Over 90% of Fortune 100 companies are now customers. ARR crossed $150 million by the time of the Series E close and is projected to exceed $200 million by end-2026. The new capital is being deployed toward what Synthesia calls “conversational AI agents” for corporate learning, pushing the product beyond video generation into interactive knowledge delivery.
3. Runway
- Total Funding: $1.05B
- Latest Valuation: $5.3B
- Latest Round: $315M Series E (February 10, 2026)
- Lead Investor: General Atlantic
- Founded: 2018
- HQ: New York, NY
- Key Investors: General Atlantic, NVIDIA, Adobe Ventures, AMD Ventures, Google, Salesforce Ventures, Coatue, QIA, Fidelity
Runway is the most-funded and highest-valued pure-play AI video company in the world. The New York-based startup closed a $315 million Series E on February 10, 2026, led by General Atlantic, with participation from NVIDIA, Adobe Ventures, AMD Ventures, AllianceBernstein, Fidelity, Mirae Asset, Emphatic Capital, Felicis, and Premji Invest. The round nearly doubled the company’s valuation from $3.3 billion at its April 2025 Series D to $5.3 billion. Total funding now stands at approximately $1.05 billion since founding in 2018.
Runway was founded by Cristóbal Valenzuela, Alejandro Matamala-Ortiz, and Anastasis Germanidis at NYU’s Interactive Telecommunications Program. Its Gen-4 model, and the subsequent Gen-4.5 release, established Runway as the quality benchmark for professional video generation, particularly for advertising, marketing, and film production workflows. Gen-4.5 introduced high-definition output, native audio, multi-shot generation, and advanced editing controls.
The Series E signals a strategic pivot beyond video. Runway has framed its next phase around world models: AI systems that construct internal representations of 3D environments to simulate physical reality, not just generate clips. The company plans to expand into gaming, robotics, medicine, and climate science using world model architectures. In March 2026, Runway launched a $10 million venture fund with checks up to $500,000, targeting pre-seed and seed startups in AI, media, and world simulation, along with a Builders program providing API credits to eligible companies.
4. Pika Labs
- Total Funding: $115M+
- Latest Valuation: ~$900M
- Latest Round: $80M Series B (June 2024)
- Lead Investor: Lightspeed Venture Partners
- Founded: 2023
- HQ: Palo Alto, CA
- Key Investors: Lightspeed Venture Partners, Spark Capital, Greycroft, Nat Friedman
Pika Labs built one of the fastest-growing consumer AI video platforms on record, scaling to over 500,000 active users generating millions of videos weekly. Founded in 2023 by Demi Guo and Chenlin Meng, both former Stanford AI PhD students, the company raised $80 million in a Series B in June 2024 at a valuation approaching $700 million, with Lightspeed Venture Partners leading. Total funding stands at approximately $115 million across three rounds.
The product has evolved significantly from its text-to-video origins. Pika 2.0, released in December 2024, introduced scene editing and character consistency, and the platform expanded into consumer social features with a mobile app launched in July 2025 that focuses on lip-sync, selfie-to-video, and viral short-form content. The company also secured distribution through Adobe Firefly, where it is available as an integrated third-party video model alongside the global Firefly Boards product launch.
Meta held acquisition discussions with Pika in July 2025, according to sources familiar with the matter, signaling that platform companies view specialized video AI as a strategic asset worth acquiring rather than replicating. No deal materialized. Pika continues to operate independently, and some market analysts project a valuation above $1.5 billion by end-2026 as the mobile product scales.
5. Mirage (Captions)
- Total Funding: $175M+
- Latest Round: $75M growth financing (March 24, 2026)
- Lead Investor: General Catalyst (Customer Value Fund)
- Founded: 2020
- HQ: New York, NY
- Key Investors: General Catalyst, Index Ventures, Andreessen Horowitz, Abstract Ventures
Mirage is the AI lab behind Captions, one of the most widely used AI video editing apps globally with over 20 million users and 200 million-plus videos created on the platform. The company was known as Captions until a rebrand in September 2025 that signaled a shift from consumer app to AI research lab. CEO Gaurav Misra announced the $75 million growth financing from General Catalyst’s Customer Value Fund on March 24, 2026, bringing total funding above $175 million.
What makes Mirage notable as a business is its revenue mix. Only 25% of revenue comes from the United States, making it unusually internationally diversified for an AI video startup. The platform recorded over 3.2 million downloads in the twelve months prior to the March 2026 raise and generated $28.4 million in in-app revenue. Enterprise customers include HubSpot and CoreWeave. The company switched to a freemium model in January 2025 to compete directly with ByteDance’s CapCut and Meta’s Edits.
General Catalyst’s Pranav Singhvi called Mirage “clearly ahead of the pack” on unit economics. The capital is being deployed primarily in Asia, where demand is reportedly the strongest outside North America, and toward proprietary model development for pacing, framing, and attention dynamics in short-form video.
6. HeyGen
- Total Funding: $74M
- Latest Valuation: $500M+
- Latest Round: $60M Series A (June 2024)
- Lead Investor: Benchmark
- Founded: 2020
- HQ: Los Angeles, CA
- Key Investors: Benchmark, Conviction Partners
HeyGen is one of the most capital-efficient AI video companies in the market. The Los Angeles-based startup raised $60 million in its Series A in June 2024 led by Benchmark at a $500 million valuation, with total funding of approximately $74 million. What the funding numbers understate is the revenue trajectory: from $1 million ARR in early 2023 to an estimated $95 to 100 million ARR by late 2025, and Forbes reporting nearly 200,000 paying customers by end of that year. The company turned profitable as early as Q2 2023.
HeyGen builds AI avatar video for marketing, sales, and communications, competing primarily with Synthesia. The two platforms serve different core segments: HeyGen skews toward individual creators and mid-market teams with a lower entry price, while Synthesia dominates the Fortune 100 enterprise tier. HeyGen’s G2 rating of 4.8 and its recognition as G2’s fastest-growing product in 2025 reflect a customer base that converts well on organic demand.
The company launched API plans in May 2025, adding a developer and B2B2C distribution layer. Customers include OpenAI, HubSpot, and Ogilvy. A new round at a significantly higher valuation is widely expected given the ARR trajectory, though no formal announcement has been made as of this writing.
7. Reactor
- Total Funding: $59M
- Latest Round: $59M Series A (May 28, 2026)
- Lead Investor: Lightspeed Venture Partners
- Founded: 2025
- HQ: San Francisco, CA
- Key Investors: Lightspeed Venture Partners, WndrCo (Jeffrey Katzenberg), Amplify Partners, Sky9 Capital, FPV Ventures
Reactor is the most recent entrant in this list and the one most worth watching as a structural bet on where AI video goes next. The San Francisco startup emerged from stealth on May 28, 2026, with $59 million in funding led by Lightspeed Venture Partners, with participation from WndrCo (Jeffrey Katzenberg’s holding company), Amplify Partners, Sky9 Capital, and FPV Ventures. AWS is the company’s preferred cloud provider.
Reactor is not a video generation tool. It is the infrastructure layer between AI model labs and developers, enabling real-time generative video with sub-50ms frame latency, a specification that makes interactive applications practical for the first time. Conventional text-to-video systems take up to 10 minutes to produce 10 seconds of output. Reactor’s architecture is designed for near-zero time-to-first-frame, unlocking use cases in gaming, simulation, robotics, spatial computing, and interactive media that batch-generation models cannot address.
The company was founded in August 2025 by Alberto Taiuti (CEO) and Bryce Schmidtchen (CTO), both former technical leads on Apple Vision Pro. The 16-person team includes engineers and researchers from Apple, Netflix, Meta, Google, Adobe, Replicate, and Microsoft. Katzenberg joined as a board observer after describing Reactor as “a bridge between the model world and real-world applications.”
The timing is significant. Reactor’s stealth exit comes exactly as the AI video market is bifurcating between content generation and real-time interaction. The infrastructure layer may be as strategically important as the models running on top of it.
What the Funding Data Actually Says
Three patterns stand out when you look at this market as a whole.
The market has split into two businesses. Frontier video generation, Luma, Runway, and Pika, is capital-heavy and model-centric. It requires hundreds of millions in compute investment and competes on generation quality. Workflow AI video, Synthesia, HeyGen, and Mirage, behaves more like SaaS: subscription revenue, enterprise contracts, and ARR that compounds without constant model reinvestment. The funding multiples and the business dynamics are entirely different.
Infrastructure is attracting its own capital wave. Reactor’s $59 million raise is not an outlier. It is part of the same pattern visible in the broader AI funding market: capital flowing below the model layer into the plumbing that makes models commercially deployable. Real-time video inference is the video equivalent of what Groq and DeepInfra represent in the text inference market.
Sora’s failure clarified the cost structure for everyone. OpenAI’s $15 million per day burn rate against $2.1 million in total lifetime revenue set a benchmark for what unsustainable compute economics look like in this space. Every other video AI company now operates with that data point visible. It is one reason Runway is pivoting toward world models with potentially wider commercial applications, and why workflow-oriented companies like HeyGen and Synthesia, which run far lighter inference workloads per customer, are scaling revenue faster than their generative counterparts.
Global AI video funding in 2025 totaled $3.08 billion, up 94.6% from 2024. The 2026 data is still accumulating, but Runway’s $315 million, Synthesia’s $200 million, and Mirage’s $75 million in the first five months alone suggest the category is on pace to exceed last year’s total.