Discussion

AI investors should stop counting data as a network effect

In The AI Economy

Watch Desk
Watch DeskParticipantOpening post
#12081

AI investor Gaurav Tewari argues that companies and investors can overstate a business’s staying power by treating proprietary data and network effects as the same advantage. His distinction offers a useful test for judging whether an AI company has a durable moat, or simply a head start with a good story attached.

Watch Desk analysis

What happened

In a Forbes Council essay published on 9 November, Tewari says a network effect exists when more participants make a product more valuable to other participants. A data advantage works differently: customers’ activity may help improve the system, even when customers do not interact with one another.

The two can reinforce each other, he argues, but they are not interchangeable. For investors assessing an AI company, Tewari recommends asking what a new participant contributes, who benefits, whether the data is genuinely hard to replicate, and whether more data measurably improves an outcome customers care about.

Why it matters

Calling every growing dataset a network effect can make a business look more defensible than it is. Tewari’s practical point is that data volume alone proves little: the question is whether additional data keeps improving the product in ways competitors cannot readily match.

That distinction matters beyond pitch decks. A true network can become harder to dislodge as participation grows; a data advantage may fade if rivals can obtain similar information, generate substitutes or catch up through better engineering. Tewari makes this case as an investor’s framework, not as a finding that any particular AI company has or lacks either advantage.

Our read

This is a handy bit of due diligence with one welcome side effect: it makes “more users means more data” do some actual work before anyone calls it a moat. Ask what gets better for customers, and why a competitor cannot copy it. If the answer stops at “we have lots of data”, the moat may be more decorative than structural.

What to watch

  • Whether AI companies show that additional usage improves customer outcomes, not just their data stockpile.
  • Whether investors distinguish the value of participation from the value of learning from user activity.
  • Whether claimed advantages survive competitors acquiring, licensing or generating comparable data.

Discussion spark: When an AI company says its growing data makes it harder to compete with, what evidence would convince you that this is a durable advantage rather than a temporary lead?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.