Generative AI – especially tools like ChatGPT – has captivated the tech world. It’s a transformative technology with endless applications, and naturally, investors are eager to find the next breakout company in the space. But before pouring capital into the latest “ChatGPT-powered” startup, VCs need to step back and ask: What exactly are we investing in?
The reality is that no startup can compete head-on with OpenAI, Google, or Microsoft when it comes to building foundational AI models. The sheer computational power and financial resources required to improve on GPT-4 are out of reach for any early-stage company.
This means most “ChatGPT startups” will fail.
Why? Because their innovation is just a thin layer of application on top of OpenAI’s API – one that could be easily replicated, outpaced, or replaced as the technology evolves.
So, what’s the right investment approach? The key is to look beyond direct competition and instead focus on startups that leverage generative AI to create meaningful differentiation in their product or service.
Where Startups Can Win: AI as an Enabler, Not the Product
Rather than betting on companies trying to build their own ChatGPT competitor, investors should focus on those that integrate generative AI into existing workflows in a way that adds real value. The best use cases aren’t about AI replacing humans but about enhancing efficiency, improving customer experience, or reducing operational costs.
Our portfolio company, Krisp, is a good example here.
Krisp initially gained traction with its AI-powered noise cancellation software for virtual meetings. Recognizing the potential of generative AI, they leveraged ChatGPT to build a virtual meeting assistant that automatically summarizes calls, extracts action items, and generates meeting notes—all while keeping the transcription on-device instead of sending data to the cloud.
This isn’t just a gimmick; it’s a genuine improvement to their product offering that enhances both user experience and data security. This is a great example of how ChatGPT is being used by tech companies to innovate and improve existing products. With the help of ChatGPT, Krisp is setting a new standard for virtual meeting productivity, and I think other companies will likely follow suit.
The Risks: Why Some Industries Should Steer Clear of ChatGPT
While generative AI presents incredible opportunities, it isn’t a universal solution. Some industries simply can’t afford the risks associated with large language models.
For example, in healthcare, finance, and defense, even a 1% error rate can have catastrophic consequences. A misdiagnosis, incorrect financial prediction, or security oversight could be far more costly than the efficiency gains AI provides. Additionally, ChatGPT doesn’t guarantee user anonymity, creating privacy and regulatory challenges that make it unsuitable for industries handling sensitive data.
Investors should be wary of startups promising generative AI solutions in these high-stakes sectors. In most cases, the regulatory hurdles and risk exposure will outweigh any competitive advantage.
Looking Ahead: What is the Future of Generative AI in Business?
If we draw a parallel to the rise of cloud computing, the trajectory of generative AI becomes clearer. Just as AWS, Google Cloud, and Microsoft Azure emerged as dominant players in cloud infrastructure, we will likely see a similar scenario unfold in AI, with OpenAI (via Microsoft) and Google controlling the foundational models.
This means that in the long run, most generative AI startups won’t be infrastructure providers—they’ll be application-layer companies that find novel ways to integrate AI into their products. And while the current pace of AI innovation is rapid, we may see a slowdown in new AI-first startups as competition consolidates around a few major players.
The real investment opportunities will emerge from companies that strategically embed AI into their existing offerings, not those trying to reinvent the wheel. The winners will be the businesses that use AI to enhance efficiency, improve customer engagement, and create new revenue streams—without being entirely dependent on the underlying models themselves.
For VCs, the playbook is clear: Don’t chase the hype. Look for startups that are solving real problems with AI, not just building thin layers on top of someone else’s technology.
