Disrupting Digital Media
As foundation models have grown larger, the expense associated with training them has risen even faster. ChatGPT-4 alone cost OpenAI approximately $100M to train. Because of this, we expect to see increased dominance at the top of this market. Nvidia is increasingly filling a similar role as Intel did in the 1990s as it provides the bulk of the foundational hardware in the space. A couple of market leaders are likely to dominate the model portion of the market which is very similar to an operating system akin to Windows which allowed Microsoft such control over the computer market in a prior era. We expect significant M&A activity over the next few years as these dominant players consolidate their grip and invest in technological add-ons to maintain the edge of their offerings, but the aspect of this market which will be more relevant to Digital Media is the application layer. While the foundation models, as with an operating system, will define what can be done and how it can be done, businesses and the public will largely interact with these models' specific applications which will offer narrower tailored functionality. As Windows allowed for the development of Photoshop and the iPhone allowed for the development of Uber, there are likely to be many novel and yet unthought-of applications for this revolutionary technology. We expect that companies like OpenAI will continue to allow their models to be built upon and that significant investment will begin to funnel downstream. VC and M&A activity will increase dramatically in this application layer over the next three years. What we are seeing from our position involved in many transactions in this space is that while these new businesses form, many legacy businesses are facing a mix of opportunities and threats. Some segments are likely to face dramatic and rapid challenges. As we saw with the impact on Chegg, a tutoring and homework help company, when its stock fell by half in May due to a shift by students using ChatGPT-4, the ramifications can be sudden. Still, the potential of this technology to reduce costs in content production is staggering. News outlets, textbook publishers, and even software development agencies are likely to remake their business models in the coming years as they adapt to a technology likely to be as disruptive to their industries as the internet was.
FE International
Page - 21
2024 Digital Media M&A Report
Made with FlippingBook - Online magazine maker