Demi Guo
Last reviewed
Sources
11 citations
Review status
Source-backed
Revision
v1 · 1,560 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Sources
11 citations
Review status
Source-backed
Revision
v1 · 1,560 words
Add missing citations, update stale details, or suggest a clearer explanation.
Demi Guo (born around 1998) is a Chinese-born American technology entrepreneur and former artificial intelligence researcher who is the co-founder and chief executive officer of Pika, a company that builds text-to-video tools using generative AI. She founded the firm, also called Pika Labs, in April 2023 with Chenlin Meng after both women left a doctoral program at Stanford University. Pika drew wide attention in late 2023 with a web product that let people generate and edit short videos from text prompts, still images, or existing clips, and by mid-2024 it had raised roughly 135 million dollars at a valuation of about 470 million dollars.[1][2][3]
Before starting Pika, Guo studied mathematics and computer science at Harvard University and worked as a researcher at Meta AI, where she contributed to published natural language processing and protein-modeling research. A competitive programmer as a teenager, she won a silver medal at the International Olympiad in Informatics in 2015. During 2025 she steered Pika away from a professional creator tool and toward a consumer social-video app aimed at younger users, positioning the company against far larger rivals such as OpenAI and Meta.[4][5][6]
Guo was born in China around 1998 and moved to the United States for university; press accounts describe her as having grown up in Hangzhou.[4] She enrolled at Harvard University, where she completed a bachelor's degree in mathematics and a master's degree in computer science. During her undergraduate years she took internships at large technology companies, reported to include Microsoft and Google, and she joined the artificial intelligence research division of Meta, then operating as Facebook AI Research, while still early in her studies. Several profiles describe her as one of the youngest researchers in that lab at the time, a claim that originates with the company and its founders rather than independent verification.[4][6]
Guo's research output gave her academic credibility unusual for a first-time founder. She was among the authors of a widely cited 2021 paper in the Proceedings of the National Academy of Sciences on large-scale protein language models, part of the Evolutionary Scale Modeling (ESM) line of work, and she co-authored natural language processing research on efficient fine-tuning of large models. Her publications also touched on 3D representation learning, reflecting the mix of language and graphics that would later shape Pika's video work.[6][10]
As a high-school and college-age competitor, Guo excelled in programming and mathematics contests. Besides her 2015 IOI silver medal, profiles credit her with first-place finishes in the United States of America Computing Olympiad open division and a gold medal at the Math Prize for Girls.[6] She went on to begin a PhD in computer science at Stanford University's AI Lab, working at the intersection of natural language processing and graphics. In April 2023 she left the program, without completing the doctorate, to build Pika full time.[4][1]
Guo founded Pika in April 2023 together with Chenlin Meng, a friend and fellow Stanford AI PhD student who became the company's co-founder and chief technology officer. Meng is an established researcher in diffusion models, the generative technique behind most modern image and video synthesis; she was advised at Stanford by Stefano Ermon and co-authored influential papers including Denoising Diffusion Implicit Models. The pair has said they started the company out of frustration with existing video tools, which they found powerful for researchers but difficult for ordinary creators to use.[3][7]
The company's stated mission is to let anyone with an idea make professional-looking video, lowering the barrier from specialist software to a few words of description. Guo has framed this as a bet on self-expression rather than industrial film production, arguing that most people will never try to make a movie with AI but enjoy making short, personal videos. By the time of its 2024 fundraising the startup remained small, with a team of roughly a dozen people that included researchers drawn from Google, Meta, and Uber.[3][5]
Pika competes in a crowded market for AI video generation that includes Runway, Luma AI, Chinese entrants such as Kling, and, from late 2024, OpenAI's Sora. Guo's early conviction that AI video would become a mass-market category, formed before the field's 2024 boom, has been a recurring theme in coverage of the company.[5][6]
Pika introduced its first widely used product, Pika 1.0, alongside its November 2023 funding announcement. The web platform generated and edited short clips from text, images, or video, and offered editing features such as extending a clip's length, restyling footage between live action and animation, and changing a video's aspect ratio. A free preview spread quickly on social media and through a large Discord community.[1][2]
The product then iterated rapidly. Pika 1.5, released in October 2024, improved video quality and introduced Pikaffects, a set of playful one-tap effects that transform a subject by, for example, crushing, melting, inflating, or exploding it. In December 2024, days after OpenAI opened Sora to the public, the company shipped Pika 2.0, whose headline feature was Scene Ingredients: users upload their own characters, objects, and backgrounds and combine them with a prompt for more precise control over a generated scene. Later releases extended the model line through Pika 2.5 and added features such as Pikaformance, which syncs facial expressions to uploaded audio in clips of up to 30 seconds at 720p resolution.[7][8]
Through 2025 Guo repositioned Pika as a consumer social product rather than a creator workbench. The company launched an iPhone app in July 2025 and, that autumn, a TikTok-style short-video app built around quick, expressive creation for Generation Z and Generation Alpha users. A flagship feature called Predictive Video, introduced in October 2025, lets someone upload a selfie and issue a simple instruction such as "make me a rock star," with the system inferring script, music, choreography, lighting, and camera work. A related product, pika.me, creates persistent AI avatars, or "AI Selves." Guo has described the strategy as competing on emotion and play rather than the photorealistic, cinematic output emphasized by Sora and Meta's video efforts.[5][7]
Pika raised money quickly for a company of its size. Before its public launch it took in roughly 20 million dollars from angel investors and seed funds, with backers reported to include Quora founder Adam D'Angelo, former GitHub chief executive Nat Friedman, and Giphy co-founder Alex Chung. In November 2023 it disclosed a 35 million dollar Series A led by Lightspeed Venture Partners, bringing the total raised to about 55 million dollars.[1][2][11]
In June 2024 Pika announced an 80 million dollar Series B led by Spark Capital, with participation from Greycroft, Lightspeed, and actor and investor Jared Leto. The round roughly doubled the company's valuation to about 470 million dollars and lifted total funding to around 135 million dollars. As of 2026 no further priced round had been publicly reported, and the 470 million dollar figure remained the company's last disclosed valuation.[3][5][7]
| Round | Date | Amount | Lead investor | Reported valuation |
|---|---|---|---|---|
| Seed | 2023 | about $20 million | Angel and seed investors | Not disclosed |
| Series A | November 2023 | $35 million | Lightspeed Venture Partners | Not disclosed |
| Series B | June 2024 | $80 million | Spark Capital | about $470 million |
User growth tracked the product's momentum. In 2024 Pika reported a community of more than 500,000 users generating millions of videos each week. By 2025, after the consumer pivot, the company said it had about 16.4 million registered users across its apps, with roughly 1.4 million average monthly active users in the first half of the year; reporting indicated that fewer than a quarter of a million were paying subscribers, underscoring the challenge of converting a viral audience into revenue.[5][7]
As of June 2026, Guo remained co-founder and chief executive of Pika, one of the more closely watched young founders in generative video. Her path, from olympiad medalist and Meta researcher to a Stanford dropout running a venture-backed startup, has made her a frequent example in coverage of women building consumer AI companies, even as Pika faces intense competition from far better-funded rivals.[4][5][6]