Charts
DataOn-chain
VIP
Market Cap
API
Rankings
CoinOSNew
CoinClaw🦞
Language
  • 简体中文
  • 繁体中文
  • English
Leader in global market data applications, committed to providing valuable information more efficiently.

Features

  • Real-time Data
  • Special Features
  • AI Grid

Services

  • News
  • Open Data(API)
  • Institutional Services

Downloads

  • Desktop
  • Android
  • iOS

Contact Us

  • Chat Room
  • Business Email
  • Official Email
  • Official Verification

Join Community

  • Telegram
  • Twitter
  • Discord

© Copyright 2013-2026. All rights reserved.

简体繁體English
|Legacy

AIGC's "King" Pika is born, with a valuation of 200 million dollars in six months.

CN
巴比特
Follow
2 years ago
AI summarizes in 5 seconds.

Chinese team, valued at over a billion, co-invested by OpenAI, the video AI generation tool Pika1.0 was stacked with Buff as soon as it was born, and it became popular all over the Internet in just a few days. The community has over 550,000 users, and it has been praised by netizens as the best video generation tool at present, becoming the strongest competitor to Gen-2 created by the same type of company Runway.

The company behind Pika1.0, Pika Labs, was established in April of this year, with the main focus of "making everyone a creative video director and producer." Currently, Pika 1.0 can produce various types of videos such as 3D animation, cartoons, or movies, and also supports canvas extension, local modification, and video length extension editing functions. According to netizens' tests, compared to Gen-2, Pika1.0 excels in generating movie scenes.

In just a few months, Pika Labs has already raised $55 million in financing, with a valuation of over $2 billion. The company's list of investors is full of familiar faces in the field of artificial intelligence, including OpenAI's co-founder Andrej Karpathy and Hugging Face's co-founder and CEO Clem Delangue.

It is also worth mentioning that Pika Labs is a Chinese team of only 4 people. The founder, Wenjing Guo, and co-founder and CTO Chenlin Meng were doctoral students at Stanford University's AI Lab; the third founder, Karlin Chen, holds a master's degree in ML&CV from CMU and was an engineer at SenseTime; the fourth employee, Matan Cohen Grumi, is the creative director and also a television commercial director.

Interestingly, the founder Wenjing Guo and the team created Pika1.0 because Gen-2 was not user-friendly.

Pika1.0 Excels in Making Movies

In addition to the conventional AIGC functions of text-to-video and image-to-video generation, Pika 1.0 not only allows video style transformation but also can directly complete content editing, change aspect ratios, and other actions. These functions allow Pika 1.0 to truly shift the AI's role in videos from "generation" to "production."

Pika's official demonstration video shows that by inputting a piece of text and providing requirements for characters, scenes, and styles, this tool can generate the video frame you need. For example, by outputting "Elon Musk in a spacesuit, 3D animation," an animated version of Musk as an astronaut appears, with a rocket that has already been launched in the background. The character's appearance, actions, background, and other elements are all very vivid.

In addition to text-to-video generation, Pika 1.0 also supports image-to-video generation. By inputting an image and descriptive words, the image will come to life according to your textual description.

If you want to change the style of an existing video, simply input style prompts in the text box, and it can handle styles like anime and pixel art, easily merging videos of all styles to create a "universal style of the universe" video.

Pika 1.0 can also perform video editing: by selecting the editing area and inputting descriptive words, it can modify specific areas based on the text prompts. For example, by drawing a box around a model's clothes with the mouse, you can change the style and color of the clothes, or even give a gorilla a pair of sunglasses with just one click.

Furthermore, Pika 1.0 can easily modify video dimensions, a technology that has already been implemented in the AI image generation tool Midjourney, and Pika is the first to apply this feature to video production.

Netizens have compared Gen-2 and Pika1.0 in generating the same wedding scene, and the conclusion is that Pika1.0 is better at using movie scenes to enhance the richness of the video, while Gen-2, although producing better video quality, has obvious flaws in handling the movement of characters.

After watching Pika's promotional video, the CEO of Hugging Face exclaimed, "In 2024, a mainstream media field will be filled with AI-generated content."

Currently, Pika1.0 supports free use for users, but requires application for access. Some netizens vividly described the current popularity of Pika1.0 as:

In short, the Pika1.0, which supports free trial, is currently in a state of high demand for access.

550,000 Users Produce Millions of Videos Every Week

Although Pika1.0 is better at making movies, it is not designed for professional movie production. "What we developed is not a movie production tool, but a product for everyday consumers," explained Pika CEO Wenjing Guo. "Although we have creativity, we are not professionals. If there is a tool like Pika, we are likely to shine at the AI Film Festival."

Wenjing Guo's story with the AI Film Festival dates back to last winter. At that time, the startup company Runway, which was already valued at $500 million, held the first AI Film Festival with a prize of up to $10,000. Wenjing Guo and a few of her Stanford classmates planned to use the winter break to make a movie using generative AI and were confident of winning the prize.

However, because the team did not have a background in film, even with Runway's AIGC tool Gen-2, the video they produced was not ideal, and they ultimately missed out on the prize, which left Wenjing Guo very frustrated.

This regret was the opportunity for Pika's birth: since the tool was not handy, they decided to create their own. In April of this year, Wenjing Guo and her classmate Chenlin Meng dropped out of Stanford and founded Pika.

Wenjing Guo (left) and Chenlin Meng

The two "iron ladies" got down to business quickly and, three months later, Pika launched its Discord server, similar to the usage of Midjourney, where users need to input text or images in the chat box to generate short videos and share them with others in the community.

Soon, tens of thousands of Pika1.0 users gathered on Discord, and now there are 550,000 people, with rapid growth, producing millions of new videos every week.

This small startup of only 4 people also quickly completed three rounds of financing, totaling $55 million. The first two rounds were led by Nat Friedman, the former CEO of GitHub, and the most recent $35 million Series A financing round came from Lightspeed Venture Partners. According to Forbes, Pika Labs is currently valued between $200 million and $300 million.

Pika's investors occupy a significant position in the Silicon Valley AI field, including OpenAI's co-founder Andrej Karpathy, board member Adam D'Angelo, Hugging Face's co-founder and CEO Clem Delangue, and Perplexity CEO Aravind Srinivas.

Michael Mignano, a partner at Lightspeed, invested in Pika in September of this year and highly praised the efficiency of this small team: "For a startup, the biggest advantage is speed, and this team's efficiency is the fastest I have seen."

Nat Friedman, the former CEO of GitHub, as one of the investors, also marveled at Pika's efficiency. It is said that Friedman suggested adding a feature to embed text into videos one afternoon, and by 3 o'clock in the morning, he was informed that the feature had already been developed.

Currently, Pika has not disclosed many technical details about the operation of the underlying models, but Pika's efficiency and product are already enough to excite venture capitalists. However, standing in the spotlight, Pika still has to face more competition.

Just last week, Runway also launched a new feature, motion slider, which can turn a picture into a dynamic video with just one stroke. In addition to targeting consumers, Runway has also partnered with film production companies.

Furthermore, the most popular AI image generation tool, Midjourney, is also developing video features.

Stability.ai has already released their Stable Video Diffusion video model, which allows users to adjust various parameters such as iteration steps and redraw amplitudes to help creators control the process of generating images, including features such as style, posture, and lines. SD also supports 3D synthesis of objects.

The video AIGC track is flourishing, with many tools available, but it depends on whether there is enough creativity to use them.

免责声明:本文章仅代表作者个人观点,不代表本平台的立场和观点。本文章仅供信息分享,不构成对任何人的任何投资建议。用户与作者之间的任何争议,与本平台无关。如网页中刊载的文章或图片涉及侵权,请提供相关的权利证明和身份证明发送邮件到support@aicoin.com,本平台相关工作人员将会进行核查。

星球发贴瓜分10万U
广告
|
|
APP
Windows
Mac
Share To

X

Telegram

Facebook

Reddit

CopyLink

|
|
APP
Windows
Mac
Share To

X

Telegram

Facebook

Reddit

CopyLink

Selected Articles by 巴比特

1 year ago
Baidu AI, needs to make money through Killer App.
1 year ago
Global AICoin Music Concert, the first time hearing the voice of China
2 years ago
These five women are changing the AI industry
View More

Table of Contents

|
|
APP
Windows
Mac
Share To

X

Telegram

Facebook

Reddit

CopyLink

Related Articles

avatar
avatar深潮TechFlow
4 minutes ago
Bitcoin Toll Station of the Strait of Hormuz
avatar
avatar深潮TechFlow
11 minutes ago
USDD officially launches WBTC Vault to diversify collateral assets, enhancing security and stability.
avatar
avatar深潮TechFlow
21 minutes ago
IOSG | When Fintech Merges with Crypto Infrastructure: The Next Decade of Digital Finance
avatar
avatarOdaily星球日报
25 minutes ago
24H Popular Cryptocurrencies and News | Iran Plans to Charge Oil Tanker Pass Fees in Bitcoin During Ceasefire; More Officials Mention Rate Hike Possibility at Federal Reserve Meeting (April 9)
avatar
avatarTechub News
27 minutes ago
CZ's autobiography "Life at Binance" excerpts: countless exciting details, a large number of early photographs.
APP
Windows
Mac

X

Telegram

Facebook

Reddit

CopyLink