Why investors find it so hard to invest in AI startups?

Many fellow AI founders have complained that investors are becoming stingy, and many investor friends are complaining it is becoming harder to invest in AI startups.

I think it is quite normal – neither side is having an easy job, and mutual understanding is probably the right approach.

The truth is, we have long passed the hype curve of AI being the magic cure of everything. After years of hard work and probably billions of dollars invested (by both tech giants and VCs), we have seen very few AI successes.

Too many rosy visions have run into dead-ends.

Market potential, promising technologies, leading researchers from top academic institutions, strong relationships with potential clients, backing from key tech giants etc. etc. – NONE has worked to be a functional judging criterion to make a portfolio of AI startups succeed.

Imagine you being an AI-focused investor, or an-AI founder, it has been a painful journey of which many still could not see the light at the end.

Some investors, who have mandate to invest in AI, have resorted to betting on hardware. Baidu has made progress in self-driving cars? No problem, let’s invest in Nvidia which provides the GPUs.

For the rest of us, we are still marching ahead. Oftentimes, 95% accuracy will make an AI fail, but at 99% accuracy adoption will take off. We have seen that in facial recognition, and now speech recognition.

NLP and others – there is a still a road ahead to find use cases that a product can be scaled.

Hope it will not take too long.