It’s a bit crazy to think that even a year ago, the term ‘generative AI’ was not exactly on everyone’s lips. ChatGPT, and to a large extent LLMs, burst into the scene only in November 2022 (after many years of research). Large language models (LLMs) have not only transformed the activities performed by computers but have also fundamentally altered the core nature of human-computer engagement.
Consider what followed since November:
- New foundational models, more efficient ways of training models etc
- An explosion of open source models, the rise of HuggingFace
- Breakthrough research across voice, video, text, multi modality, etc
- Billions of dollars invested in thousands of startups at the application layer of generative AI.
To put things in context, let’s start with the basics:
2023 witnessed an impressive influx of $15 billion invested in generative AI startups, significantly surpassing last year’s $4 billion. It’s essential to note that this figure is influenced by several massive transactions, including the $10 billion round of funding received by OpenAI.
In the US alone, the share of venture capital dedicated to AI doubled in 2023 alone
There have been a few other ‘mega rounds’, mostly into foundational models, with some exceptions (such as Character AI)
If you take into account how that capital has been distributed over the past year, the majority went to what you’d refer to as the ‘picks and shovels’ of generative AI – new LLMs (to rival GPT), GPUs (to rival Nvidia), data and training, etc
As opposed to other major platform shifts, where perhaps the incumbents were late to lean in (Microsoft missed mobile, Google missed social, etc), when it comes to generative AI, while many were hit by surprise to the success of ChatGPT, they were quick to respond and lean in massively.
There’s a question on whether all this GPU hogging, and building capacity is not just a bubble waiting to pop, similar to the Telecom crash in the early 2000’s. As David Sacks from All in put it:
On aggregate, a lot of funding was also deployed to startups in the application layer – investing in founders leveraging generative AI to either solve narrow tasks, or bring automation to tasks to reduce costs, increase accuracy, speed etc. As a data point, 60% of the 212 startups in the Summer 2023 Ycombinator batch that has just graduated identified themselves as AI startups.
Gartner recently confirmed that we’re at the peak hype of generative AI, and that perhaps because of it, many investors have been waiting by the sidelines. Why should they risk investing in a startup that could get commoditised by either OpenAI, Google, Microsoft or open source technology? And even if they find a fantastic team working on a valuable problem, would it be redundant if we reach AGI in the next decade?
But rather than dismiss anything that is not ‘picks and shovels’ in generative AI, few opinionated funds and investors (ourselves at Remagine Ventures included) are saying – “don’t be so quick to dismiss this”. I recently wrote about opportunities for Israeli startups in generative AI as well as why gaming is ripe to be disrupted.
Seth Rosenberg at Graylock published ‘Product Led AI’ – pointing to opportunities for founders building AI-first companies. For example, he sees great potential ‘for co-pilots are “branded” sales people, like wealth managers, insurance brokers, and mortgage brokers’ and re-defining the product surface area, for example: how would AI change customer service or productivity?
Here’s an interesting excerpt on finding value despite the noise:
There’s lots of noise in AI. From true techno optimists who envision AI as the great amplifier of humans, to pessimists who see every app as just a thin layer on top of OpenAI, to the optimists-turned-pessimists who believe AI will automate all jobs (and take over humanity).
Undoubtedly, there will be detractors who believe many products are simply features on top of foundational models. But builders who see AI as the driving force behind product development and GTM strategy will actually create new markets and experiences that never existed before.
By combining expertise in products and domains with a fundamental understanding of human behavior and AI, these builders will bring defensible, valuable AI-first products to life.
Seth Rosenberg, Greylock
Christoph Janz, co-founder and managing partner of Point Nine Ventures, pointed to the potential of Generative AI to accelerate the adoption of vertical SaaS. As a seasoned SaaS investor (and the person behind the SaaS napkin, one of my favourite SaaS industry benchmarks), he rightfully points out that the new crop of SaaS companies will be built with AI in mind, and leveraging AI, startups can create ‘magical user experiences’, that will accelerate the ‘time to wow’ and unlock a lot of automation for SMBs, and not just enterprise clients. Index Ventures also pointed out the potential of generative AI to disrupt vertical SaaS.
Investors, innovators, and visionaries should look deeper and recognise generative AI as a powerhouse, not just a tool in the shed. While it’s not an easy time or area to deploy into, I believe that a lot of large companies are going to be created with an AI-first approach, and with less resources required than their pre-AI predecessors.