The AIVAR (Artificial Intelligence and Virtual/Augmented Reality) market is expected to be worth $3.18 trillion by 2024 [1,2]. Perhaps no company is better positioned to capitalize on the rise of AIVAR than NVIDIA, but like all hardware companies NVIDIA is at risk of being commoditized by low cost-copy cats. NVIDIA has built themselves a nice castle, but if it wants to become the next mega-tech company and avoid the fate Cisco then NVIDIA is going to need a big moat. Because in a war, a castle with a big moat is exponentially more valuable than a castle without a moat.
A moat’s function is to protect the castle, and so it doesn’t have to generate revenue to be valuable. Take Google as an example. Google makes most of its money selling ads (its castle), but Google’s moat is its search engine, a completely free service used by hundreds of millions of people each day and which costs billions to run. To this, Google has added additional moats with Gmail, Waze, Chrome, YouTube, Maps, and Android. The amount of voluntary user generated data that Google collects, puts it in a unique monopolistic position.
Google’s moats have created sticky relationships with hundreds of millions of people who feed it with vast quantities of data about the world — data that would be impossible to acquire algorithmically. This is data can only be acquired by employing the voluntary efforts of hundreds of millions of users who feed Google with about 10 exabytes of new information every day [3] — to put that into perspective, that’s the equivalent to 5 million x 128GB iPhones of storage added each and every day. And while individuals and companies might be able to use access small slices of that data, only Google has the ability to commercially mine it to take advantage of its full potential.
To Illustrate the power of this data-effect, imagine that you could take the top 500 engineers from Google’s Search team to build a new company. If you were starting a hardware company, you could be assured that you could create a formidable competitor. But software is different. These engineers would never be able to make a search engine as good as Google, because Google’s search engine is is not just the algorithm, it is the vast stores of data that those algorithms run on. Data that they’ve gathered for nearly two decades. This data not only informs the algorithms, but new algorithms are written to refine the quality of the data and to generate derivative data, for which new algorithms are built. This is a positive feedback loop, and so no matter how fast you can run, Google will run faster.
Google is an interesting case study because it didn’t start by building a castle, it started by building this impenetrable moat. But in doing so it created a nice protected piece of property which would serve as the foundation for AdWords. Yahoo on the other hand focused on building a castle. It couldn’t have cared less about moats. In fact when Larry Page and Sergey Brin tried approached Yahoo in 1997 to sell them the Google algorithm for $1 million, Yahoo turned them down because Google’s search engine was so good it threatened Yahoos business model.
With NVIDIA’s $50b market cap (up from $14b just 10 months earlier), it’s in a position to go and buy itself a moat. Maybe the cheapest moat NVIDIA could have bought was Yahoo. While Yahoo is a shell of its former self it still has vast quantities of user generated data across multiple domains (these are the best moats). NVIDIA could have then built a special API into special data centers for its AI and AR/VR algorithms and which would only run with NVIDA hardware. Remember, the job of the moat is to protect the castle, so it doesn’t matter much that Yahoo had limited growth, its properties just need to generate enough data to give NVIDA an asset that none of its competitors would have, and which would drive business to NVIDIA over upstarts like AMD or Intel.
So if NVIDIA goes shopping, who could it look at? Well because doesn’t have to make money off its moat (it makes money off its hardware) it could look to turn someone else’s castle into NVIDIA’s moat. Take the satellite company Planet as an example. Planet is generating tremendous amounts of data, but their commercial offering is be quite expensive. However if NVIDIA acquired Planet it could offer these services for free and open up APIs for consumer and enterprise applications, of which the most demanding application which require NVIDIA hardware. By giving the service away for free, it removes friction and provides an opportunity for widespread consumer use, and therefore more data.
So what are some others? And keep in mind that NVIDIA could open up derivative or scrubbed data sets may also be necessary to preserve privacy:
Box: At a $2 billion market cap, the cloud storage provider could provide also NVIDA with a huge enterprise and consumer data opportunity.
Foursquare: With 7 billion check ins and 55 million daily active users. Foursquare has a large dynamic user generated data set created by mobile users. In 2015 Yahoo was rumored to be bidding on foursquare for $900m. Here’s a case where you have a valuable data asset that’s struggling to find a business model. Another great opportunity to turn someone else’s castle into your moat.
Mailchimp: Mailchimp has over 120 million users and and sends 80–100 million messages every day. In their data stream they have vast quantities behavioral data for each email recipient. NVIDIA could even buy one of Mailchimp’s competitors and drastically cut the price while upping the quality of the service.
23andMe: This company has very large data sets on the human genome and human traits/personality which could be very valuable to the biopharma community. While HIPPA privacy rules could cause come complications, there could be ways to scrub or create derivative data.
Centrifuge Systems: A competitor to Palantir, Centrifuge Systems is a big data discovery tool. NVIDIA could create a low-cost version that’s 1/100th the price of Palantir.
Hootsuite/BufferApp: With Twitter’s $12 billion market cap, it’s probably too expensive for NVIDIA, but it could look at one of the tools used to interface with Twitter.
Salesforce Competitor: Salesforce is an expensive beast with tremendous amounts data. NVIDIA could look at buying one of the smaller competitors and offering a nearly free product that’s sufficiently compelling for 80% of the use cases.
These were just a few ideas. NVIDIA has an enormous market opportunity ahead of it, but it’s going to need a moat to stave off competition and that moat is user generated data. With fresh $50b market cap, its in a unique position to do some shopping as a whole bunch of unicorns come to realize that their business is more valuable as someone else’s moat than their own castle.
Disclaimer: I own shares of AMD, Google, and NVIDIA.
[3] http://www.deepwebtech.com/2016/06/lets-take-a-look-at-some-really-big-big-data/
Hacker Noon is how hackers start their afternoons. We’re a part of the @AMIfamily. We are now accepting submissions and happy to discuss advertising &sponsorship opportunities.
To learn more, read our about page, like/message us on Facebook, or simply, tweet/DM @HackerNoon.
If you enjoyed this story, we recommend reading our latest tech stories and trending tech stories. Until next time, don’t take the realities of the world for granted!