By Rodrigo Madanes, Global Innovation AI Officer, and Jeff Grabow, US Venture Capital Leader, at EY
Artificial intelligence (AI) is a bright spot in the venture capital (VC) landscape. Without this paradigm-shifting technology, investment across the tech-focused VC ecosystem would be contracting. Instead, we are having a much different conversation in terms of tenor and substance. We’re talking about how this moment holds immense promise for countless companies in Silicon Valley and beyond. And those VCs that invest successfully will likely focus on two things: the tech itself and the team leading the startup.
AI represents a foundational technology shift that will change how enterprises work. Like any major shift, the full impact will take time to realize. The current environment feels very much like the mid-1990s when the internet was moving into its commercialization phase.
Unlike the consumer-driven adoption of the internet, enterprises are driving AI adoption. This is an important difference; since large, publicly traded enterprises have profit pressures that affect their ability to invest in new technology, AI adoption could take longer overall than what occurred with the internet. The resulting impact could be that the investment window will stay open longer than we witnessed with the internet cycle.
As a result, startups (and the VC firms backing them) should be asking questions about what the enterprise is really spending money on and then adapting to meet those needs.
Another key factor is the probable consolidation of the AI technology stack. With the internet, the technology stack was built up before there was significant consolidation within it. We will likely see the same phenomenon in AI, as well.
Now, if you’re a tech-focused VC and you haven’t made some form of investment in AI in the past couple of years, you might be inviting some questions. However, when significant technology shifts like this happen, progress often comes in fits and starts as the market foundation establishes itself and its development and evolution take shape. There’s been a massive rush of interest for large language models (LLMs). Some VC investors argue that the LLM train has essentially left the station—with the fierce competition in the market, the costs to fight to win are simply too high.
This shouldn’t make VCs concerned that they are too late for AI. We’re still very much in the early innings of AI and its impact on the market—and likewise, the returns it can deliver for VC investors.
As a result, VC firms are busy identifying how the market will evolve. The market map can be divided up in many ways, with some choosing to view investments through the lens of use cases: consumer, enterprise horizontal, enterprise vertical, and prosumer, highlighting the multimodal nature of GenAI applications. Others still find it more useful to organize it by modality, like AI foundation models, Agentic AI, enterprise tooling, and others.
The consensus is that new entrants in foundation models are unlikely, as the cost of talent, compute, and data required to build the next generation of foundation models is far too high. This space remains highly competitive, but the winners will be among the existing competitors. VCs believe there are opportunities for small language models focused on industries or areas that can help increase accuracy when seeking industry-specific insights.
VC firms are still investing in Agentic AI companies, as these have clear ROIs for enterprises. The field is nascent, allowing relatively small companies to create substantial value. In the enterprise AI tooling space, investments have become increasingly crowded. there has been some crowding of investments. Enterprise tooling has traditionally been an area of good returns and success for VC firms. Some winners are starting to emerge, and companies are aligning themselves to become horizontal or vertical category winners.
In the more research-oriented areas, there is a fair amount of activity. Some areas of multimodal AI have been consistently invested in (image generation, video generation). Others are still incipient (speech models, specialized document readers for presentations, PDFs, or spreadsheets). Spatial and motion AI is relatively early, too, but some robotic demos have captured the imagination of many. The technology is advancing at breakneck speed.
So yes, circumstances and environment matter a great deal. For example, think about how expensive data used to be and how that’s plummeted in recent years and decades. Timing the tech to the right moment and adjusting accordingly is crucial. Again, sometimes being first isn’t best; perhaps it is best to learn from those suffering the scrapes as they cut through the brush.
This is where the second factor – talent – that VC firms are looking at becomes critical. Ultimately, VC isn’t just betting on tech; they are betting on teams. They are looking for people who can adapt and discern signals from noise. This is crucial because an investment thesis can’t rely on tech alone. The history of Silicon Valley and global commerce is littered with examples where the best tech did not win out. Not to date both of us too much, but Betamax was a superior technology to VHS.
That’s in part why investors are looking for founders and teams that can take feedback from a whole host of sources and then plot a refined course forward. Investors don’t necessarily care whether a founder’s idea is perfect; they care about whether the founder can pivot from something that isn’t working to something that very well might. That means maintaining a strategic direction for your business while adapting to unexpected or unplanned tactical changes. Flexibility and ingenuity are key to success (and successful returns).
From a VC’s perspective, not all AI technology is created equal. We are still very much in the early innings of AI’s impact on markets and commerce. The promise of this technology is strong but not preordained for those looking to invest. Knowing which startups to back demands getting the tech and the talent right.
The views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organization or its member firms.
About the Authors
Rodrigo Madanes is Global Innovation AI Officer at EY.
Jeff Grabow is US Venture Capital Leader at EY.
