Business
Pika, a new TikTok-like AI app, makes playful, creative short videos from just a few words
Welcome to Eye on AI, with AI reporter Sharon Goldman. In this edition…a TikTok-like AI video app that is built for Gen Z…Challenges renting San Francisco apartments during the AI boom…TSMC profit surges 39% to hit yet another record on AI chip demand…Spotify partners with Sony, Universal, Warner and more to develop AI music products.
Demi Guo, the 26-year-old founder of AI video company Pika, holds degrees in mathematics and computer science. But she’s also always had a creative streak, including writing poetry, and she loves showing off on social media with image and video tools.
That’s why she dropped out of her Stanford Ph.D program and co-founded Pika in April 2023 with aspiring animator and fellow Ph.D student Chenlin Meng. Just as AI video companies like Runway were starting to get hot, the duo, after experimenting with making long-form generative AI films, decided to create an easier-to-use AI video generator—sparking interest from top Silicon Valley investors like Nat Friedman.
Now, has Pika has raised about $135 million at a $470 million valuation and boasts 14.5 million users across several creative apps. It recently launched a new TikTok-like AI video app, also called Pika, that is already trending, and last week it debuted a new feature in that app called Predictive Video. Unlike other video tools which require long prompts to generate good results, Predictive Video allows users to upload a selfie and say something simple, like “make me a rock star,” or “I’m giving a TED Talk,” or “make me sing in Japanese”; the tool then infers your intent, delivering a complete video with a script, music and dance moves, background, lighting, and camera angles and visual effects.
Instead of just producing a standalone video clip, Pika can anticipate motion and interaction, producing a sort of mixed reality by allowing users to weave themselves or real-world elements into AI-generated scenes—something that traditionally would take significant animation skills and production time.
All of it is tailor-made for Gen Z and Gen Alpha, cohorts which have grown up on short-form video and uses online platforms to express themselves and share their thoughts, Guo explained. “Most nonprofessionals will never try to create a film using generative AI, but lots of people like to make short videos,” she said. “It’s really about self-expression.”
That, she insists, is the opposite of what has become a catch-all tool for AI video: Slop. There may be artificial content at scale, but Pika is meant to, ironically, use generative AI to help people get more real, she says.
“We really believe it is not meaningless content,” she emphasized. “It’s about self-expression, the personality is actually real behind it. So a person who is funny will post funny videos, or a person that is very egocentric will post egocentric videos.” Some people like to use one visual effect, or create a video that captures a vibe or a mood, or simply use the tool to do what they wish they could do in real life, she added.
“It helps you achieve that in some sense,” she said. “Maybe you are a really bad singer, but in AI, you can sing very well, like a pop star.”
Pika is stepping into a competitive landscape dominated by giants like OpenAI’s Sora and Meta’s new Vibes, but it’s carving out a very different niche. Sora is focused on cinematic, photorealistic text-to-video generation, and Vibes is about bringing AI-generated video feeds into Meta’s massive ecosystem. But Pika positions itself closer to the ground-level habits of Gen Z and Gen Alpha: It’s less about polished productions at scale, and more about emotion, creativity and interaction for everyday storytelling.
With that, here’s more AI news.
Sharon Goldman
sharon.goldman@fortune.com
@sharongoldman
FORTUNE ON AI
Don’t fear the AI bubble, it’s about to unlock an $8 trillion opportunity according to Goldman Sachs – by Jim Edwards
BlackRock’s $40 billion deal highlights the unstoppable AI data center gold rush, as CEO Larry Fink pushes back on AI bubble fears – by Sharon Goldman
Want to build your own chatbot for $100? A glimpse into AI’s small, cheap, DIY future – by Sharon Goldman
Amazon is planning a new wave of layoffs, sources say – by Jason Del Rey
AI IN THE NEWS
Challenges renting San Francisco apartments during the AI boom. According to the New York Times, a surge of AI start-ups is reshaping San Francisco’s housing market, driving up demand and prices and intensifying competition for apartments. After his company Cluely raised $5.3 million, 22-year-old founder Roy Lee leased eight luxury apartments just steps from the office to foster a “frat house” work culture—emblematic of how tech firms are reshaping city living. Fueled by growth from companies like OpenAI and Anthropic, San Francisco rents have risen 6% in the past year, more than double New York City’s rate, with neighborhoods near A.I. hubs like Mission Bay seeing 13% spikes. The scramble for housing has led renters to tour dozens of units, face same-day rejections, and even bring cash to viewings, raising new concerns about affordability and displacement in one of America’s most expensive cities.
TSMC profit surges 39% to beat estimates and hit yet another record on AI chip demand. CNBC reported that Taiwan Semiconductor Manufacturing Co. (TSMC) posted record third-quarter results as surging demand for artificial intelligence chips fueled a 39.1% jump in profit from a year earlier, beating analyst estimates. Revenue climbed 30.3% to NT$989.92 billion ($33.1 billion), while net income rose to NT$452.3 billion. CEO C.C. Wei said growing consumer adoption of AI models is strengthening the company’s conviction in the “AI megatrend,” prompting TSMC to raise its 2025 revenue growth forecast to the mid-30% range and boost planned capacity investments to at least $40 billion. High-performance computing, including AI and 5G chips, accounted for 57% of revenue, and advanced 7-nanometer-and-smaller chips made up 74% of wafer sales, underscoring TSMC’s central role in supplying cutting-edge processors to clients like Nvidia and Apple.
Spotify partners with Sony, Universal, Warner and more to develop AI music products. Variety reported that Spotify is teaming up with major music industry players—including Sony, Universal, Warner, Merlin, and Believe—to develop “responsible” AI products aimed at empowering artists and songwriters, marking one of the first large-scale collaborations between streaming and music rights holders on AI. As part of a broader push to invest heavily in AI research and product development, Spotify plans to build a generative AI lab and product team focused on four key goals: creating tools through up-front agreements with labels and publishers; allowing artists and rights holders to opt in to generative music technologies; opening new revenue streams; and strengthening artist-fan connections. The initiative reflects a growing effort across the music industry to shape how AI is integrated into creative and commercial practices.
AI CALENDAR
Oct. 21-22: TedAI San Francisco.
Nov. 10-13: Web Summit, Lisbon.
Nov. 26-27: World AI Congress, London.
Dec. 2-7: NeurIPS, San Diego
Dec. 8-9: Fortune Brainstorm AI San Francisco. Apply to attend here.
EYE ON AI NUMBERS
26%
That’s how many Americans are familiar with AI mental health tools, according to a new YouGov survey that highlights just how far such tools have to go to win over the public. A mere 11% say they’d consider using one, and trust is even lower—just 8% say they trust AI in this space, while 40% explicitly don’t. The top reasons for skepticism: concerns about lack of human nuance (53%), harmful or inaccurate advice (50%), and data privacy (49%). And despite the flood of startups and investment in AI therapy apps, only 4% of Americans have ever tried one.
