🚨 Growth-stage investing isn’t all spreadsheets and revenue charts. James Luo of CapitalG joins us for VC Wednesdays, and shares why he puts just as much weight on conviction in the founder. 🚨
✒️ What’s the No. 1 signal you look for to invest at the growth stage?
A couple things don’t change no matter what stage. One, the size and scale of the opportunity. We want growing markets with secular tailwinds — multi-tens-of-billions market opportunities. The second is the team. We always look at why the founder is uniquely positioned to solve this problem, but also followership: Who are the people on your leadership team, and where are they coming from? It's an amazing signal if your exec team left their jobs and joined you because they trust you as a leader and want to be part of the journey.
✒️ What trend or sector are you the most excited about, and why?
The macro trend obviously is AI. But people forget that AI starts with hardware. ChatGPT only works because GPUs, voltage regulators, network switches and power transformers are running in the background. We’re facing physical-world constraints now: power, networking, data center efficiency — and that’s why we’re spending more time on hardware and semiconductors. We just announced an investment in Empower Semiconductor, whose approach makes last millimeter power delivery super efficient by collapsing voltage regulation. Networking is also a huge bottleneck — GPU clusters have to talk to each other super fast for both AI training and inference. So we’ve also backed nEye.ai, which is doing optical circuit switching so GPUs can reconfigure on the fly like a big brain.
✒️ On the AI application side, how do you assess differentiation defensibility when so many use the same LLMs and are doing the same thing?
People assume all models are interchangeable but we actually see differences. They are less fungible than you think. Think of it like Coke vs. Pepsi — users develop preferences. We look for content differentiation. For example, AlphaSense is tackling market research and has proprietary financial models, transcripts and expert research — you can’t just Google it. If you pair that with a strong LLM, the results are dramatically better. The other big layer is workflow. AI that integrates deeply into how people work — and saves them time or makes their job better — builds long-term moats. Look at Clay or Abridge. That’s what creates defensibility.
✒️ What is one investment that you regret passing on and why?
We spent a lot of time with Island, the secure enterprise browser founded by Michael Fey. I've known him for many years and always thought the vision was super exciting — the idea that the browser should be the operating system for work. The mistake we made is the classic one VCs make: it just felt really expensive at the time. We thought, let's wait for more proof points before we get conviction at this price. We overcomplicated it.
#VCWednesdays #vc #venturecapital #startups #TechonLinkedIn
