This report is a deep dive into the world of Gen-AI—and the first comprehensive market map available to everybody. We provide an overview of over 160 platforms in the space and their investors, as well as insights from leading thought leaders on the potential of this technology. This hands readers a unique opportunity to gain a comprehensive understanding of the generative AI market and the potential for new players to challenge established players like Google.
“Generative AI is a foundational technology, and as always with these new platforms, the opportunities that it opens are ample—we passed the stage of "if" and we are at the stage of "when" and "how." We are seeing the infrastructure layer maturing and democratizing as LLMs get open sourced, which accelerates the application layer.’’ —Irina Elena Haivas, Investor and Partner at Atomico
Please note: The information provided in this piece is based on Antler’s day zero investment approach and the support we provide to founders around the world. The platforms featured in our industry mapping are sourced from Crunchbase. It is worth noting that some of these platforms may intersect both AI and Gen-AI. If you believe your platform should be included in our future mappings, please reach out to us at Ollie.Forsyth@antler.co.
What is Gen-AI?
Imagine a world where instead of spending days writing a blog post, a week creating a presentation, or several months on an academic paper, you can use generative assistant tools to complete your projects in minutes. These tools not only help us with our projects, but also support us in making better decisions.
Here is an example of how powerful Gen-AI platforms may become: For those familiar with our reports on The Creator Economy, imagine a world where creators can upload their content into any language and have their own voices used as the voiceover, instead of relying on robots or local translators. This is a brave new world where we have access to powerful tools that can save us countless hours and enhance our work.
"We’re at an inflection point in generative AI, for two reasons: computers can create better than ever, and it’s never been easier for people to interact with them.’’ —Molly Welch, Investor at Radical Ventures.
“At Media Monks, we believe that Generative AI will have a significant impact on our industry, though it is difficult to imagine the real scope of this amazing technology. We have been researching Generative AI for about five years and the rate of innovation has become exponential. Advances in the technology are occurring within our production timelines, which range from 1–6 months. What this means is that tools we use at the beginning of a project are already obsolete by the time we go live.” —Samuel Snider Held, Creative AI Designer & Engineer at Media Monks.
AI vs Generative AI
Artificial Intelligence (AI) is a broad term that refers to any technology that is capable of intelligent behavior. This can include a wide range of technologies, from simple algorithms that can sort data, to more advanced systems that can mimic human-like thought processes.
Generative AI (Gen-AI), on the other hand, is a specific type of AI that is focused on generating new content, such as text, images, or music. These systems are trained on large datasets and use machine learning algorithms to generate new content that is similar to the training data. This can be useful in a variety of applications, such as creating art, music, or even generating text for chatbots.
In essence, AI is a broad term that encompasses many different technologies, while generative AI is a specific type of AI that focuses on creating new content.
The vast opportunity unfolding
It is likely that Gen-AI will have a significant impact on the creative industries in the future. While some creatives may be replaced by Gen-AI systems, others may find new opportunities to work with these systems or to create content that is enabled by Gen-AI. In many cases, it may actually enhance the work of creatives by enabling them to create more personalized or unique content, or to generate new ideas and concepts that may not have been possible without the use of AI.
One potential benefit of Gen-AI for creatives is that it can enable them to create content more quickly and efficiently. For example, a writer may be able to use a Gen-AI system to generate rough drafts of articles or stories, which they can then edit and refine. This can save time and allow creatives to focus on the most important aspects of their work.
"Generative AI is a huge wave that’s going to create unavoidable ripples across almost all industries and for the vast majority of them, we think it’ll be incredibly value-adding. We see the biggest opportunities as platform plays built on top of the underlying models, where UX, accessibility, and embeddedness will be key differentiators in this race. All of this needs to be powered by a killer go-to-market strategy and above all, speed! The next half-year will be pivotal.’’ —Stephanie Chan, Investor at Samaipata Ventures.
The impact of Gen-AI
This technology can have many different impacts depending on how it is used. For example, Gen-AI can be used to create new content, such as music or images, which can be used for a variety of purposes such as providing the creatives with more flexibility and imagination. It can also be used to improve machine learning algorithms by generating new training data. Overall, the impact of Gen-AI is sure to be significant, as it has the potential to enable the creation of new and useful content and to improve the performance of machine learning systems.
“We are heading for a time when artificial intelligence is widely available. But being widely available and actually usable to achieve business outcomes are two very different things.” —Dave Rogenmoser, CEO & Co-founder of Jasper.
How do the training models work in practice?
Gen-AI training models work by learning from a large dataset of examples and using that knowledge to generate new data that is similar to the examples in the training dataset. This is typically done using a type of machine learning algorithm known as a generative model. There are many different types of generative models, each of which uses a different approach to generating new data. Some common types of generative models include generative adversarial networks (GANs), variational autoencoders (VAEs), and autoregressive models.
For instance, a generative model trained on a dataset of images of faces might learn the general structure and appearance of faces then use that knowledge to generate new, previously unseen faces that look realistic and plausible.
Generative models are used in a variety of applications, including image generation, natural language processing, and music generation. They are particularly useful for tasks where it is difficult or expensive to generate new data manually, such as in the case of creating new designs for products or generating realistic-sounding speech.
“These new foundational models as well as applications built on top accelerate the pace of many industries: generating creative content for gaming and social media companies, automating manual processes within enterprises, helping scale operations previously unimaginable such as movie, music, and comics production—the possibilities are endless.’’ —Manjot Pahwa, Investor at Lightspeed Venture Partners
How are language models created?
There are several ways to create a language model, but the most common method involves using a machine learning algorithm to train the model on a large dataset of existing text. This process typically involves the following steps:
- Collect a large dataset of existing text. This dataset should be representative of the language or style of text that you want your model to be able to generate.
- Preprocess the text data to clean and prepare it for training. This typically involves tokenizing the text into individual words or phrases, and converting all of the words to lower case.
- Train a machine learning algorithm on the preprocessed text data. This can be done using a variety of algorithms, including recurrent neural networks (RNNs) and long short-term memory (LSTM) networks.
- Fine-tune the trained model by adjusting the model's parameters and hyperparameters, and by using additional training data if necessary.
- Test the model by generating sample text using the trained model and evaluating the results. This can be done by comparing the generated text to the original training data, or by using other metrics such as perplexity or BLEU scores.
- Refine the model by repeating steps 4 and 5 until the generated text is of high quality and matches the desired language or style.
“It is important to note that creating a language model requires significant computational resources and expertise in machine learning—although the space is still early, platforms are spending millions of dollars on fine tuning their products and services.
The current challenge for founders in the Generative AI category is to build not just a product, but also a defensible business model with the capacity to endure. Any competent developer can wrap an application skin around these underlying generative engines. The solution is to incorporate sustainable competitive differentiation through strategies like embedding network effects, raised switching costs, ingrained product partnerships, etc.’’ —David Beisel, Partner at NextView Ventures.
Why does Gen-AI exist?
Gen-AI exists because it has the potential to solve many important problems and unlock the door to myriad new opportunities in a wide range of fields. Some of the key reasons why Gen-AI is a growing field of research and development include:
- Gen-AI can create new content. One of the key benefits of Gen-AI is its ability to generate new content, such as text, images, or music. This can be used to create new art, music, and other forms of creative expression, and to generate data for training machine learning models.
- Gen-AI can improve efficiency and productivity. By automating the generation of content, Gen-AI can help save time and reduce the need for manual labor. This can improve efficiency and productivity in a variety of fields, from journalism and content creation to data annotation and analysis.
- Gen-AI can improve the quality of generated content. With advances in machine learning and natural language processing, Gen-AI is becoming increasingly sophisticated and capable of generating high-quality content that is difficult for humans to distinguish from real content.
- Gen-AI can enable new applications and uses. The ability of Gen-AI to create new content opens up many possibilities for new applications and uses. For example, it can be used to create personalized experiences, such as personalized news articles or personalized music recommendations.
“This isn't as widely known. My opinion is that generative AI models are magical now because they've been able to take in people's inputs through language. And because they are able to represent so many different concepts—and combine them—they can make beautiful, wild, and creative results. It's exciting, thrilling, and perhaps a little scary. For creatives, this means finding inspiration with a muse, creating prototypes faster, and refining pieces with the combined skill of the model (Photoshop++).’’ —Sharon Zhou.
Looking into the future—Gen-AI revenue models
There are several potential revenue models for companies that use Gen-AI technology. Some possible revenue streams include:
- Licensing the technology to other companies or organizations that can use it to improve their products or services.
- Selling the outputs of the AI system, such as generated images, videos, or text, to customers who can use them for various purposes.
- Providing access to the AI system as a subscription service, where customers can use it to generate their own outputs
- Using the AI system to improve the efficiency or effectiveness of a company's existing products or services, and then charging customers for those enhanced offerings.
- Creating new products or services that leverage the capabilities of the AI system, and selling those directly to customers.
Why now?
There are several reasons why now is the time for Gen-AI. First, advances in machine learning and natural language processing have made it possible for AI systems to generate high-quality, human-like content. Second, the growing demand for personalized and unique content, such as in the fields of art, marketing, and entertainment, has increased the need for Gen-AI platforms. Third, the availability of large amounts of data and powerful computational resources has made it possible to train and deploy these types of models at scale.
“There has been a promise that AI is going to change the world and we’ve been waiting for it since 2012. In the past two or three years, something has finally changed. While recent excitement around generative AI has been text-to-image, I believe AI-powered text generation will prove to be far more transformative. And now, with increased access to cutting-edge language models, we are seeing this technology proliferate into everyday products—completely changing the way companies do business and reimagining how humans experience technology." —Aidan Gomez, co-founder & CEO at Cohere.
Description of Gen-AI landscape categories:
- Text: Summarizing or automating content.
- Images: Generating images.
- Audio: Summarizing, generating or converting text in audio.
- Video: Generating or editing videos.
- Code: Generating code.
- Chatbots: Automating customer service and more.
- ML platforms: Applications / ML platforms.
- Search: AI-powered insights.
- Gaming: Gen-AI gaming studios or applications.
- Data: Designing, collecting, or summarizing data.
The Gen-AI fundraising landscape
With a number of investors focused on the Gen-AI space, we have shortlisted the most active ones:
The Gen-AI unicorn landscape
Although the sector is still emerging, a few unicorns are already emerging. Two unicorns were produced in 2019, one in 2020, and four in 2022 thus far.
Trends:
How is Gen-AI being used for arts and music?
Gen-AI is being used in art and music in a few different ways. One common application is using generative models to create new art and music, either by generating completely new works from scratch or by using existing works as a starting point and adding new elements to them. For example, a generative model might be trained on a large dataset of paintings and then be used to generate new paintings that are similar to the ones in the dataset, but are also unique and original.
How is Gen-AI being used for gaming?
Gen-AI is being used in gaming in a number of ways, including to create new levels or maps, to generate new dialogue or story lines, and to create new virtual environments. For example, a game might use a Gen-AI model to create a new, unique level for a player to explore each time they play, or to generate new dialogue options for non-player characters based on the player's actions. Additionally, Gen-AI can be used to create new, realistic virtual environments for players to explore, such as cities, forests, or planets. Overall, it can be used to add a level of dynamism and variety to gaming experiences, making them more engaging and immersive for players.
‘“Generally, the short-term innovation areas will be extremely positive. Games and online 3D experiences have been notoriously hard to build—Generative AI will completely upend that by making it exponentially easier to create game assets. The potential downsides, or rather consequences, of applying Generative AI in gaming is more existential. While single dimension applications like AI-generated copywriting or image creation are merely amplifiers of existing tasks we perform and still allows us to control the application of the output (i.e., we can decide to to accept/reject a piece of copy and decide where to use the copy), our interactions with AI in gaming will be much more multidimensional. Over time, AI (whether it’s environmental, behavioral, or NPC characters) will evolve and adapt to human heeds and likewise, humans will get used to socializing and regularly interacting in these AI-generated realms.’’ —Annie Zhang at Roblox.
How will Gen-AI impact the creator economy?
With The Creator Economy already a $100 billion dollar industry poised for continuous disruption, Gen-AI is likely to have a significant impact on creatives—especially those creating music, art, or writing. However, it does present the opportunity for creators to be global from day one, allowing their content to be turned into any language using the creators voice or turning their creativity into more engaging content.
"Generative AI will turn creators into super-heroes and will augment areas where they aren't as strong. Think of it more as a creator co-pilot, rather than a creator replacement.” —Jim Louderback, Author of Inside The Creator Economy.
For the creator economy to succeed, platforms will need to adapt to the creators’ personalities so the creators have some form of connection with their fans when the content may have been mostly supported with AI platforms.
‘’I'd argue that the human element is essential for art to have value. When AI-generated art is created by algorithms and machines, rather than by individuals with their own experiences, emotions, and perspectives, it can be seen as lacking the authenticity and humanity that are often seen as essential to great art. This can make it difficult for some viewers to connect with AI-generated art on an emotional level, which can reduce its impact and significance.’’ —Ivona Tau, creator.
However when we ask a creator what impact Gen-AI will have on them, one creator said:
“Not much. That said, I'm watching what's happening with great interest. I'm truly inspired by the results other people are getting with the help of generative models. You often hear artists call AI image models as ‘tools,’ but AI is so much more than a tool. It's a creative partner, a synthetic genie, or an inspirational ally.’’ —James Gurney, artist.
What does the future hold for the space and what challenges might it face?
There are many challenges that lie ahead for Gen-AI, including improving the quality and diversity of the outputs produced by these models, increasing the speed at which they can generate outputs, and making them more robust and reliable. Another major challenge is to develop generative Gen-AI models that are better able to understand and incorporate the underlying structure and context of the data they are working with, in order to produce more accurate and coherent outputs. Additionally, there are also ongoing concerns about the ethical and societal implications of generative AI, and how to ensure that these technologies are used in a responsible and beneficial way.
Let’s take a closer look at a number of these concerns:
Copyright. As of today it's challenging to see how these platforms identify the original source of truth or where artwork came from - the models are trained by hundreds of millions of data points. Creators are concerned about how these platforms will be able to mitigate copyright infringement of the creators’ work. As we saw with a recent case—tweeted by Lauryn Ipsum—there are images being used in the Lensa app that have backgrounds of the original artist’s signature.
“One of the most pressing issues in generative AI right now is system trustworthiness. Large language models like OpenAI’s ChatGPT are prone to sharing incorrect or false responses. In image generation, where systems have been trained on large volumes of imagery, there are copyright and intellectual property questions around system outputs, making enterprise users uncertain about integrating them into products or workflows.’’ —Molly Welch, investor at Radical Ventures.
Students writing their dissertations. As these platforms become smarter, young savvy students will adopt them in their daily lives. How will this impact their academic work and how will their professors be able to identify if this is truly their work? Gen-AI will have a huge impact on the education space that remains to be seen.
“The opportunities for students to use chatGPT to supplement their learning is endless, assuming that the ChatGPT model continues to improve. Students can use it to generate content for quizzes and flashcards to help them study, optimize existing code, or even write summaries for study guides. The key word here is supplement. Students should use ChatGPT in addition to their own original work they're already putting in. ChatGPT can be problematic when students use the content as a replacement for their work or even submit ChatGPT content as their own original thought. University administration and students need to work together to build policy to clearly state what is acceptable in this new world. I took an open-book exam last week that explicitly prohibited the use of ChatGPT or any other AI support.” —Cherie Lou, creator and student at Stanford University.
Disinformation vs misinformation. Although these systems are insanely smart, they will inevitably provide misinformation at times. For example, in a recent Channel 4 interview in the UK, the host was asking the Open AI about his career path, and the chat-bot assistant gave inaccurate information. As the training models become more adaptive and learn more about us, in time there will be fewer mistakes in the algorithms.
Drawbacks of Gen-AI include:
- The risk of bias in the generated data, if the training data is not diverse or representative enough.
- Concerns about the potential for generative AI to replace human labor in certain industries, leading to job loss.
- The potential for Gen-AI to be used for malicious purposes, such as creating fake news or impersonating individuals.
It’s possible Gen-AI will replace millions of jobs from designers to producers to artists; however, creatives will always exist in some aspect.
Gen-AI will impact the metaverse—exactly how remains to be seen.
It is difficult to predict exactly how generative AI will impact the metaverse, as the latter is still a largely theoretical concept and there is no consensus on what it will look like or how it will function. However, Gen-AI will play a significant role in its creation and development, as it will allow for the automatic generation of content and experiences within the virtual world. This could potentially lead to a more immersive and dynamic metaverse, with a virtually limitless supply of new and unique experiences for users to enjoy. It is also possible that Gen-AI could be used to automate various tasks within the metaverse, such as managing virtual economies and ensuring that the virtual world remains stable and functional. Overall, the impact of Gen-AI on the metaverse is likely to be significant and wide-ranging.
“There will be business opportunities in different layers of the AI stack, and we are already seeing some business models emerging. Obviously it's very expensive and complex to produce foundation models like GPT-3, and the few companies that can do it will be paid handsomely. But there are countless opportunities to develop more specialized models and to bundle general capabilities into something that a particular target market needs. This is the equivalent of vertical SaaS, applied to AI. We are probably going to see a lot of AI-enabled SaaS plays that provide a holistic solution with great UX for a particular market.Further down in the stack, providing the right kind of training data, enabling ML engineers to build specialized models quickly and assuring the robustness of models are all very viable businesses.’’ —Andreas Goeldi, Partner at BTOV Ventures.
Let’s shape the future together
Get ready for a technology shift that will revolutionize the future of work! We are on the brink of a new era in which thousands of jobs will be transformed and new ones created. These cutting-edge Gen-AI platforms will undoubtedly support and enhance our daily lives, but it will take time for us to fully adapt to them.
“This unprecedented level of human-machine collaboration is in full swing and the game is now open to whoever will take the lead in fully integrating the generative AI method, regardless of the industry you are in.’’ —Gabrielle Chou, Associate Professor at New York University, Shanghai.
If you're building the next generation of Gen-AI platforms, we would love to meet you and learn more about your work. Let's shape the future together!
Apply to an Antler residency in 20+ cities across six continents.
Read the Antler India team's new content series, "AI: Humanity's Miracle Machine."
