Unai, VR, and the Hardware Lottery

Today: Why VR is hard & Max Perumal’s vision; Trump’s tax returns; corporate VC.

The Agenda 👇

  • All about our portfolio company Unai

  • Reflecting on strategy in the virtual reality space

  • Trump’s taxes and loss carryforwards

  • More fresh news about CVC

I admit I’m not passionate about VR. I used to love gaming and immersive experiences, but that was concentrated in a few years of role-playing games with a group of high school friends in the 1980s (we played Warhammer & Call of Cthulhu). Then I veered into other areas of interest: bass guitar, politics, technology, corporate finance 😅 I can’t say I’ve had an actual gaming experience for the last 20 years.

As a result, I was never really interested in fancy VR headsets. To this day, I’ve never even put one on—even though I once visited the Oculus team in Menlo Park! I see the hype (Erlich Bachman: “Nobody understands it, but everybody wants in”), and the potential applications for what Balaji S. Srinivasan calls “cloud communities”, but VR has typically been a vertical I’ve filed under “check again later”.

It doesn’t help that since VR is about hardware, it has a deep-tech twist that only reinforces my skepticism. In short: I agree deep tech is important because it helps software eat the world, but I don’t think it should be the priority in Europe. Here’s what I wrote about that back in June 2019,

If a region (like Europe) doesn’t have tech champions down the stream, most deep tech investments will benefit others. The market for technology is now global, and large companies are scouting assets all around the world. Thus there is no reason to think that a deep tech investment in a given country will necessarily create long-term value in that country. It will only do so if the resulting asset is used by local players to tackle local problems.

There’s one thing, however, that piques my interest, and it’s the big strategic question in VR: How do you foster adoption when you need to build both a hardware set and the experience that set gives access to? Just one of those is already hard enough to build and market; but both of them together? That’s the really tricky part, and nobody has managed to pull it off so far. Here’s what Benedict Evans just wrote in the latest installment of his weekly newsletter (about Facebook’s Oculus product line):

There have always been two VR questions: the devices needed to get much better and cheaper; and we needed something to do with it beyond games, with broader appeal. We're now 3/4 of the way towards a great headset, but we aren't any closer to broader content...

Facebook hopes VR/AR is the next universal platform after smartphones, and it's doing games not because it *cares* about games but as a way to bootstrap adoption and development on the way. It's not yet clear that will work.

I started thinking about it all again recently when Maxim Perumal, one of our portfolio entrepreneurs, headed toward launch with his VR startup, Unai. And even if, like me, you might not be passionate about VR, Max’s personal story should be enough to pique your interest:

  • Back in high school at age 16, he reverse engineered Oculus’s headset and built his own for a total cost of just $100. Along with two friends, he saw the potential to build a startup based on that and wrote a message to my cofounder Oussama Ammar to ask for advice and support.

  • Oussama met Max, immediately saw his genius, and gave the following advice: it’s too early to launch a startup. Should it somehow succeed, Max would have the miserable life of a child prodigy. Should the startup fail, he risked being turned off entrepreneurship for life.

  • Instead, Oussama suggested open sourcing the headset (see here) and using the subsequent years to meet as many people as possible, learning about as many topics as possible. And with Oussama’s ongoing help, that’s what Max did, as explained in this article in The Next Web 👀

Now 3 years after first meeting with us, Max thinks he finally has everything needed to launch his startup. In short, he’s building a new headset with one application: connecting with people you know within a virtual world—or, as he writes in the slidedeck,

The first version of the headset is an immersive communication device only. In this virtual environment, people will be able to chat & chill with their friends or with strangers—to build meaningful social relationships.

  • In this way, Max is facing the hardware/software problem in a very focused way, making the two interact perfectly together without worrying about being a platform for other developers, or needing to defeat other headset makers, etc.

The old discussion about the hardware/software bundle was well illustrated by the series Halt and Catch Fire. In the first season, Joe MacMillan wants to build a computer, which comes with such high sunk and fixed costs that he has no choice but to literally hijack an old industrial company (Cardiff Electric) from its owner and repurpose its entire value chain (and a brilliant engineer named Gordon Clark). Sadly, it mostly ends in failure.

  • In the second season, the narrative is somewhat refocused on the character of young software developer Cameron Howe, who decides to skip the hardware part and focus on code. Suddenly, everything seems to become easier. Cameron’s team can work at her home rather than in a lab or a factory. The initial costs are much lower since she can reward developers with equity in the new venture rather than a high salary. It’s a great illustration of the radical difference between hardware and software economics, and of the fact that the bulk of innovation in the following years happened in software rather than in the hardware space.

There are precedents, however, of companies that succeeded at handling both hardware and software. One of them, of course, is Apple, whose story has been told many times. Another less widespread example is Atari, which is well documented in this article by Harry McCracken: Atari at 40: Catching Up with Founder Nolan Bushnell.

Indeed, something else happened in the meantime: Apple has proven the validity of the hardware-centric approach to software innovation. As explained by Daniel Eran Dilger in The Apple Era begins as Microsoft, Google shift to a hardware centric model,

It's not common to point out the clearly discernible absurdity of what Google and Microsoft are doing as they desperately seek to transition themselves from broadly licensed platform vendors to mobile device companies. However, it goes without saying why they are trying to do so. The age of broadly licensed platforms is over. The primary vendors of the world's broadly licensed platforms are all shoveling billions at efforts to turn themselves into an Apple, Inc.

This doesn’t mean, however, that everyone can succeed in building hardware. There are many problems:

  • First, it goes without saying, but building hardware is...hard—generally harder than building software. As Danny Crichton wrote in TechCrunch 6 years ago, “Hardware startups are considered “hard,” requiring huge gambles on design, production quality, and supply chain management that software startups simply don’t need.”

  • Second, building both hardware and software exposes you to two categories of technological risk (and that’s without mentioning the risks associated with marketing and selling the product). Ask Rony Abovitz of Magic Leap how it’s been spending $3.5B while being exposed on so many fronts: Magic Leap Tried to Create an Alternate Reality. Its Founder Was Already in One

  • Three, people today just don’t want too many devices; after all, they only have two hands. With most of their attention being taken up by their smartphone, you need to make a very strong case to entice them to consider another device. The bad news? History suggests that it’s close to impossible.

  • Four, the pace of innovation is dictated by software. If you, a hardware startup, try to do more than build a mere operating system and rely on third parties to innovate, you’re doomed to lag behind. Your own pace of building is likely dictated by your hardware’s limitations—and because you can’t keep up with the fast pace of software innovation, your punishment is the curse of obsolescence.

About that, just a few days ago Matthew Clifford of Entrepreneur First quoted a paper by Sarah Hooker of Google Brain about software innovation being dependent on the current state of hardware:

Success within a particular hardware paradigm promotes further investment in that hardware, which “makes it even more costly to stray off the beaten path of research ideas”. We can think of this as “algorithm-hardware co-evolution”.

In other words (echoing Seth Godin): Software is rarely better than the hardware it’s running on.

There have been many solid attempts at building a hardware/software combo in the recent past, but success is not at all guaranteed:

So what about Max’s approach? I see three things that might do the trick:

  • First, having open sourced the technology makes it possible for Max to capture the wisdom of the crowd. It’s been a way to offload a significant fraction of the sunk and fixed costs Joe MacMillan had to handle when designing and building a computer. It’s a paradigm that simply wasn’t possible back in the 1980s. Now it is—and Max made the most of it.

  • Second, renting the device rather than selling it, as is Max’s intention, opens the way for replacing it as soon as it’s obsolete and there’s a better version—rather than having your users stuck with an old device and needing to buy a new one (and risking that only the minority does so, while the majority is just disgruntled with a product they feel hasn’t kept up).

  • Third, Max’s idea is to focus on an application that’s much more mundane than a sophisticated game populated by orks and other monsters. His idea is to market the headset with the most basic value proposition: hanging with your friends (or interesting people that can become friends) and having great interactions.

You might say it’s not much, but think about it. Young people crave socializing these days. The pandemic has only made things worse for them, forcing everyone back to their parents’ houses and having to take precautions when meeting friends.

  • As Benedict Evans wrote, “This should have been a VR moment, and it isn’t”. Then again, maybe it is—maybe we just don’t see it yet. For instance, with the pandemic Zoom has finally imposed the usage of seeing the person you’re interacting with. (The audio quality in phone calls is often terrible, whereas if I see my interlocutor I can read their lips and finally understand them regardless of audio quality.)

So we’re all feeling that we’re getting there, but we haven’t yet had that extra push over the cliff. My only question as I looked at Max’s compelling deck was: what about the price? $100/month to rent the state-of-the-art device might be a price hardcore gamers are willing to pay for immersive experiences. Will it make the cut if it’s about rewarding interactions with other people? What do you think?

🇺🇸 Trump’s Tax Returns

Finally! After four years of wondering what was in those tax returns and why Trump himself seemed so desperate to keep them out of sight, The New York Times got access to the entirety of Donald Trump’s tax history. I’ll let you read it all, but here’s the header:

The Times obtained Donald Trump’s tax information extending over more than two decades, revealing struggling properties, vast write-offs, an audit battle and hundreds of millions in debt coming due.

Donald J. Trump paid $750 in federal income taxes the year he won the presidency. In his first year in the White House, he paid another $750.

He had paid no income taxes at all in 10 of the previous 15 years — largely because he reported losing much more money than he made.

I can’t help sharing an article I wrote in October 2016 about Trump’s unusual relationship with the tax system—pointing out that the use of loss carryforwards to offset profits after years of losses was what every individual was in need of in the Entrepreneurial Age. I can’t say that article ever resonated with a broader audience, but have a look: 

Assessing a loss on your personal tax returns doesn’t necessarily mean that you’re a bad entrepreneur (although it clearly appears that Trump is precisely that). In most cases, such losses are due to initial investment efforts on the part of the partnership; and then it seems quite logical that you shouldn’t pay taxes on money invested to generate future income, no?

🤑 Corporate Venture Capital

Yesterday’s edition was about corporate venture capital. Here are three interesting sources to go further:


If you’ve been forwarded this paid edition of European Straits, you should subscribe so as not to miss the next ones.


From Normandy, France 🇫🇷

Nicolas