Is AI Creating Groupthink in Venture Capital?

Venture capital is a tough business. It's not enough to be right. For a venture capitalist to win big they have to be both right and non-consensus. 

So when we saw the results for this year's Venture Trends Survey and Event at Decile Group, we saw what could be a red flag for groupthink in the industry. At the very least, it's a warning that there's a little too much consensus around one of the hottest investment categories right now. 

In fact, it's a historic amount of consensus. We have never seen this much consensus since we started conducting this annual survey of our ecosystem, which includes VCs within our program and the broader venture community. 

When asked to name the hottest sector in the upcoming quarter, for the first time ever, over 50% of respondents all pointed in the same direction: artificial intelligence. 

There are two possible conclusions from this. The first is straightforward: People are just very bullish on AI.  It's a once a generation technology that will truly change everything, like the Internet. 

The other possible conclusion is more complex and more concerning: When that many investors agree on the same direction, there is a lack of thought diversity across the sector, and the early signs of groupthink.

Venture capital has always been a non-consensus game of finding the ideas that most people aren't looking at. So it's concerning from an industry standpoint when everyone is increasingly watching the same space, and this same space happens to also be watched by the likes of Microsoft, Google, and X. 

Is AI driving convergence in venture capital thinking?

In VC, convergence is an old concern. Capital follows signals. GPs rush into hot sectors. Startups start to look the same. We've seen this in crypto, in D2C, in clean tech 1.0. But the convergence around AI seems to be a lot stronger than we saw with other hyped trends of the past. 

Here's our thought bubble: Is AI itself driving this? Is AI not just the topic of the convergence but the technology shaping how the capital is thinking? The engine behind the accelerated groupthink? 

As more investors use AI tools to stay up to date, research opportunities, generate insight, and shape strategy, there's a risk that their thinking becomes shaped by the same systems they use to run their firms. These systems are replacing associates. And while you might joke that associates are all churned out by the same business schools, they were at least separate individuals doing their own research, not the same systems. 

If the result on a macro level is uniformity, that's a huge risk for the industry.

At Decile Group, we think about these questions a lot.

Decile Hub, the platform for emerging managers to launch and run their firm, uses AI extensively to help emerging managers set up and scale their VC funds. We're building automation to reduce friction in fund formation, firm building, capital raising, and internal ops; so that managers can do more of what only humans can do: Build conviction and back people.

In many ways, the platform also acts like a skilled analyst in a firm, who conducts high-quality research, creates deal memos, preps LP pitches, and writes the firm's newsletters. 

But unlike a human analyst, this one doesn't have a unique perspective shaped by personal lived experiences or contrarian thinking; unless you build that in deliberately. And that's where our attention turns next.

Designing Against Convergence: How We're Building AI at Decile Group

So how do we embrace the benefits of AI in VC without falling into the traps it might create?

First: our internal LLM is not trained on generic public data. It is based entirely on inputs from our team, sector experts, and direct data from our community. The result is a model that reflects the lived experience of launching and operating a VC firm today, not a regurgitation of industry narratives.

We are also training the AI to generate counterfactuals by default. If a manager asks the LLM to create a deal memo, the AI is being set up to return a plausible alternative view. The goal isn't to confuse the user. It's to push them to interrogate their own thinking. That's something we want to encourage across the board.

Increasingly, we're building features that prompt managers to add their own personal analysis; in their own voice; when using AI to shape outputs. The aim is to nudge them back into the creative, judgment-driven mindset that brought them to venture in the first place.

We're optimistic about what AI can do in the hands of bold, independent thinkers. But it's up to all of us to stay vigilant; and to use this tool to unlock originality, not erase it.

It feels good when people agree with you. But that's not the path of a VC.