Generative AI refers to the use of advanced machine-learning techniques to create original content or output. The generative AI startup ecosystem took off in 2020, following the introduction of GPT-3—Open AI’s 175 billion-parameter large language model (LLM). Subsequently, since September 2022, ChatGPT has taken the world by storm, gaining 100 million users in just two months and becoming the fastest platform to do so.
At present, generative AI startups engage in the space in four broad ways: 1) By leveraging closed-source FMs from creators, including OpenAI (GPT-n series, DALL-E series, ChatGPT, Whisper), Google (Vertex AI, PaLM API), AI21 Labs (Jurassic-series), and Anthropic (Claude); 2) By developing in-house, proprietary AI models; 3) By deploying a combination of third-party models with in-house algorithms; and 4) By harnessing open-source models like EleutherAI’s GPT-J and GPT-Neo and others available through platforms like HuggingFace.
Generative AI has found widespread applications across various commercial domains. It is widely used to generate personalized marketing content and to enhance entertainment experiences, such as creating game characters, artwork, and music. It also powers natural language conversational chatbots, aids in product design, automates workflow, and facilitates personalized learning experiences, among other capabilities.
Generative AI solutions are increasingly adopted across various industries, particularly the information technology and industrial sectors. This specifically includes the software subsegment under information technology and the professional services subsegment under industrials.
The most common use case is for developing content, particularly for marketing and sales purposes like ad creation, customer training, and customer assistance. Other prominent use cases include conversational content creation like chatbots to assist employees and customers.
We have identified key Generative AI use cases below:
The generative AI applications space is currently dominated by startups, particularly in segments such as 1) marketing content and 2) design, publishing, and digital assets. The startups in these two segments alone account for over half of the identified players in the industry.
However, about two-thirds of these companies are in pre-seed or seed stages, suggesting that product development is still in its infancy. Early- and growth-stage companies account for the smaller share of the space, with companies like OpenAI, AI21 Labs, and Anthropic standing out for their in-house foundation models and chatbots. Among startups that have successfully commercialized their products, the majority appear focused on marketing content creation. Key examples include Persado, Jasper, and Synthesia.
Microsoft, Alphabet, NVIDIA, Amazon, and Meta are among notable incumbents that have entered the market with in-house generative AI tools and investments.
Big Tech incumbents are primarily active in developing the foundation models underlying generative AI applications. However, these players have also developed products leveraging their own in-house generative AI. Alphabet's “Bard” and Microsoft's “Bing” are cases in point, serving as direct competitors in the conversational AI content segment. Other examples include NVIDIA’s text-to-3D tools, “Magic3D” and “ATT3D,” Amazon’s coding companion, “CodeWhisperer,” and Meta’s testing playground for advertisers, the “AI Sandbox.”
In addition to in-house solutions, these industry leaders have actively sought to expand their presence in the generative AI space by investing in or partnering with other generative AI startups and related companies. Microsoft stands out in this regard, with its significant investments and partnerships with OpenAI since 2019.