The state of generative AI: 5 charts on growth, adoption and development.

From headlines to boardrooms, AI is getting a lot of attention. 

The remarkable advancements in generative AI models, combined with the availability of user-friendly and openly accessible platforms like ChatGPT, Midjourney, and Bard, have sparked a surge in both public interest and corporate enthusiasm. 

Behind the buzz, a thriving ecosystem of technology startups is emerging, producing products and services applying foundational model progress to real-world challenges.

In this article, we take a look at the funding, adoption and development trends behind this new technology.

1. Generative AI adoption has outstripped every tech product to date.

The user adoption rate for generative AI is rewriting the tech industry playbook, outstripping all records to date. ChatGPT amassed 1 million users in just 5 days and 100m active users in just 2 months according to research from UBS. Put into context, it took TiKTok and Instagram 2 ½ years to reach the same milestone according to data from Sensor Tower.

2. A surge in generative AI startup investment.

Foundational generative AI businesses are exciting but very capital-intensive. We’ve seen a capital influx, with a number of big rounds, at high valuations as investors run to get into the space: Stability AI raised a $100 million round at a $1 billion post-money valuation in October 2022; Microsoft invested $10 billion into OpenAI in January, Anthropic raised a $300 million at a $5 billion post-money valuation and in June, France’s Mistral AI raising a $113M seed round at a $260m valuation. 

3. Capital concentration in foundational model production.

The first phase of generative AI development has naturally started with the development of foundational models. As the sector evolves, we expect to see downstream opportunities (including application layer and middleware) attract increasing investment. 

These foundational models have raised big rounds at high valuations, taking the lions share of the capital. Why? Two words: training costs. Training large language models (LLMs) is very capital intensive (Open AI’s losses doubled last year to $540m as it developed ChatGPT). Founders and investors within the space are betting on adoption in place of revenue growth, and are banking on monetisation to come as the value-chain in front of these foundational players evolve.

The second phase of downstream applications are less capital-intensive and closer to the market (meaning that they have a faster route to monetisation). We see gathering momentum in these businesses as enterprises race to build and adopt generative AI tools (in late June, for instance, Typeface a company building generative AI for brands raised $100 million at a $1 billion valuation, led by Salesforce Ventures, just months after a $65 million round in February.)  These businesses hold long-term promise as foundational models improve, costs fall and founders find new, meaningful solutions for enterprise applications. At Forward, we approach the market optimistically, but pragmatically. We’re mindful of hype and look for teams focused on deep customer problems, building solutions based upon an intimate sectorial understanding that will enable them to build competitive moats less dependent on foundational model technology alone.

4. Rapid emergence of a generative AI ecosystem.

Foundational models, though nascent, are already inspiring a flurry of activity downstream in the UK startup ecosystem. GitHub’s star history provides a window into developer interest and adoption. This chart illustrates both the fast growth in adoption of these platforms and the increasing interest in new models from Stability and Nomic.

At Forward, we’ve seen the startup community rapidly address the technology, with a slew of new, exciting ‘application layer’ businesses emerging both from new startups in our deal flow and within our portfolio companies. Applications range across text, video, image and code (see Sequoia Capital’s generative AI Landscape for a breakdown). Many of the most exciting focus on solving problems for businesses, creating new possibilities, better user experiences and driving operational costs down. These businesses can monetise solutions faster than their foundational counterparts, and we’ve already seen commercially successful products based on generative transformer and score-based diffusion models (like Robin AI) yield results. 

For a more in-depth look at deal flow, check 550+ startups in NFX’s open-source generative tech market map.

5. A young, market with big potential. 

Out of more than 250+ generative AI companies profiled by CBInsights, a third remain bootstrapped, while a further half are still at the Series A or pre-A stage. As the sector matures, we expect a larger share of funding to be directed to downstream, market-ready applications with clearer routes to monetisation.

Separating signal from noise.

As you may have gathered from this article, the generative AI market is hot right now. There’s a lot of interest from enterprises and entrepreneurs who see opportunities to leverage the value, and investors who see the potential upside in the technology. But, picking stars in a galaxy of opportunities is tough. 

At Forward, our investment approach has always been to focus on the fundamentals: customer-focused businesses solving acute problems that leverage an intimate understanding of a sector to deliver value and build competitive advantage.

Founder? Learn more about our funding here, or apply to join our Office Hours here. ‍

Investor? Learn more about our investment strategy and how we can help you navigate the generative AI landscape by subscribing to the Forward Thinking Bulletin, our monthly newsletter.