Table of contents
Invented by the Chinese more than 2,500 years ago, Go is thought to be the oldest board game ever played. There are vastly more possible Go games than there are subatomic particles in the known universe. In the beautiful documentary AlphaGo, we see the country of Korea captivated by an artificial intelligence (AI) program playing one of the best humans in Go, Lee Sedol, favored to win by a mile.
In March of 2016, human and machine sat down to a Go match – best of five games. In Game 2 – Move 37, the AI chose a higher probability of winning over a smaller margin win, an act of brilliance no human could conceive of. After losing 3 of 5 games, human Lee Sedol chose to keep playing. In Game 4 – Move 78, his unexpected wedge move, “God’s Touch,” was something the AI gave a 1 in 10,000 chance of happening. That move won Lee the game, and showed humans are still equally capable of brilliance.
Today, brilliant humans and the world’s “smartest” artificial intelligence algorithms are both being put to good use in something far more complex than Go – large multinational enterprises.
From IoT to Enterprise AI
The same year the algorithms were wreaking havoc on Korea’s national pride, we wrote about C3 IoT – A Full-Stack IoT Platform for Everyone. At that time, C3 claimed to have the first and only enterprise-scale Internet of Things (IoT) platform in production with implementation across 20 enterprise customers. The company was founded in 2009 by tech legend Tom Siebel, a self-made billionaire with a proven track record of creating value in the tech world. His biggest accomplishment was probably Siebel Systems, a company he founded that went from inception to more than $2 billion in revenues in just six years, having scaled to 8,000 employees in 29 countries. (Siebel Systems merged with Oracle in 2006.)
Founding C3 wasn’t the only big thing that happened to Mr. Siebel in 2009. Perhaps even more impressive is the story of when an elephant kicked the crap out of him in Tanzania. Mr. Siebel had an iPhone in his pocket at the time, and this is what it looked like after a six-ton elephant stomped on it.
The roots of C3 are the pursuit of a goal to make the world a better place. After exiting Siebel Systems, Mr. Siebel went on to pursue philanthropic work in 2007/2008, making an impact on the planet’s energy and climate issues. He then realized what many of the ESG types fail to grasp. Subsidization doesn’t scale. In order for an environmentally friendly solution to change the world, it needs to grasp the coattails of capitalism.
Since its inception in 2009, C3 has raised a total of $228.5 million in disclosed funding. Just days ago, they filed for an IPO to raise even more money.
C3 – A Leader in Enterprise AI
Two words grace the front of C3’s S-1 filing – Enterprise AI. Following those words are graphic portrayals of big numbers. C3’s 4.8 million machine learning models in production are learning from 622 million sensors to make 1.1 billion predictions a day. All the value add generated by their platform has resulted in customers that increasingly spend more and more on the platform resulting in strong quarterly revenue growth.
C3 has focused on establishing key reference clients in each industry they operate in – what they call “lighthouse customers” – who can then be used to convince everyone else to come on board. Evidence of the value they provide can be found in the increased spend once customers begin using the platform.
C3 can deploy enterprise AI applications into production in as little as four weeks. This speed becomes apparent when you look at how quickly the company has moved to deploy their enterprise AI platform across multiple industries, all while working with some of the biggest players in each.
|Oil and Gas||1%||29%|
|Aerospace and Defense||3%||18%|
That pivot away from utilities comprising 67% of revenues in 2018 to 24% in 2020 makes a lot more sense when we dig into the company’s history a bit.
Competitors in Enterprise AI
C3’s S-1 makes a bold statement about the lack of formidable competition in enterprise AI.
Our primary competition is largely do-it-yourself, custom-developed, company-specific AI platforms and applications. We are unaware of any end-to-end Enterprise AI development platforms that are directly competitive with the C3 AI Suite.C3
There was an interesting article published by InformationWeek in 2016 that helps shed some light on this statement. In that piece, Mr. Siebel talked about how C3 was “two to three years ahead of every other company,” noting the success they had in the utilities industry. At that time, Harbor Research claimed, “more than 80% of European smart meters are under management by the C3 IoT platform.” To manage that much information, C3 was achieving 1.5 million transactions a second on their platform. To put that number in perspective, Amazon processes about 426 transactions a second during the busiest of times.
The utilities implementation was the pilot that proved the concept, and C3’s successful move into other industries shows how industry-agnostic the platform is.
Throughout the InformationWeek article, emphasis was placed on the number of devices companies are collecting data from as a measure of their IoT maturity. Names mentioned included General Electric and PTC Inc., two companies we’ve looked at before under the context of enterprise IoT. None of the names mentioned in the article had achieved the level of connectivity that C3 had achieved.
Today, the situation seems to be much the same. In Mr. Siebel’s letter to investors included in the S-1, he makes the following remark:
I believe C3.ai is uniquely qualified to tackle these challenges. But clearly, as an investor, you will need to resolve these questions to your satisfaction.Tom Siebel
When it comes to competitors in this space, there’s the usual big company banter. We also know of at least 7 Industrial IoT Startups Using AI to Monitor Machines. Enterprise IoT startups like Augury are using machine learning algorithms to “listen” to the sounds machines make in order to determine when they might fail, something referred to as predictive analytics. Then there’s emerging industrial IoT startup Samsara, which went from zero to 10,000 enterprise customers in four years, and now claims to have 100 billion connected devices.
Mr. Siebel may be a bit biased, but he’s probably the best judge of whether or not his company has a comparative advantage when it comes to being a leader in enterprise AI. That’s not because the man is a self-made billionaire who gets into grudge matches with elephants. It’s because he makes it a point to surround himself with the world’s most talented people who have helped him confidently arrive at that conclusion.
We attract exceptionally talented, highly educated, experienced, motivated employees.C3
The War for Talent
In a world where some of the largest tech companies hire adventure cartoonists and dare to say meritocracy is a myth, it’s refreshing to see a company that doesn’t pay lip service to “the war for talent” you hear recruiters drone on about during brown bag lunches. C3 makes it a point to talk about their pursuit of hiring only the best talent there is.
We had to open multiple browser windows to confirm what appears to be a problem with C3’s corporate home page. For some reason, it defaults to the hiring page of their website.
Then we realized, that’s intentional. The most important web real estate they have is used to convey messages to future employees. They want everyone to know what a great place to work it is. It harks back to the tech boom of the late 1990s when companies would only hire truly talented tech workers who they then treated like rock stars.
C3 is consistently ranked as one of the best places to work, so unsurprisingly, it’s tough to get hired there. The S-1 filing talks more about selectivity.
We hired 214 new employees in the past year. We received approximately 52,000 applications for those positions.C3
At 0.41%, that makes Harvard’s acceptance rate of 5% look like a cakewalk. C3 is one of those rare places where the recruiters and human resources staff are as competent as the people they’re hiring.
If you’re an investor looking for exposure to the growth of enterprise artificial intelligence, it’s hard to see a situation where you’d avoid investing in the world’s best talent led by an inspirational leader who knows the recipe to succeed in the complex world of enterprise software. Given what we know about the hype surrounding initial public offerings (IPOs) these days, this poses a problem.
To Buy or Not to Buy
For evidence of how hyped these IPOs can become, look no further than Snowflake, a company that Warren Buffet invested in which soared well beyond what anyone could imagine. Volatility is a problem we see with many IPOs, making it difficult to figure out when to buy shares. Waiting until the dust settles doesn’t always result in cheaper shares.
C3 is an attractive addition to our disruptive tech portfolio, especially considering how hard it’s been to find stocks which provide pure-play exposure to the IoT theme. To remove some of the subjectivity out of the equation, we might propose a buying schedule that results in a position being established over four months following the IPO’s debut. Here’s how.
At the moment we’re holding 25 stocks in the Nanalyze Disruptive Tech Portfolio. (This includes one stock we’re presently exiting, and one we’re just getting ready to start accumulating.) That means our target position size for a C3 position would be 3.8% (1/26). For example, if our total portfolio value is $100,000, then we’d be looking at a position size of $3,800.
We can reduce market timing risk by purchasing fixed amounts of stock at fixed intervals over time, something that might look like this:
- Opening day of the IPO – buy $760 worth of shares
- 30-day anniversary of IPO – buy $760 worth of shares
- 60-day anniversary of IPO – buy $760 worth of shares
- 90-day anniversary of IPO – buy $760 worth of shares
- 120-day anniversary of IPO – buy $760 worth of shares
Four months after the IPO, we’d make our final purchase, and be holding our target position size of 3.8%. In this way, we’ve removed all emotion from the buying decision before shares of C3 even begin trading.
Artificial intelligence algorithms are exponentially more powerful and efficient than they were four years ago. A year after Lee Sedol’s infamous loss, AlphaGo Zero debuted, a neural network that knew nothing about the game of Go. After three days of training itself, AlphaGo Zero beat the original AlphaGo – not once, but 100 games to 0 – using a small fraction of the resources.
C3 is doing what we’ve long been waiting for IBM Watson to do – true enterprise AI. They’ve taken inspiration from what we learned when an AI algorithm first beat a human at Go. Only by recruiting and enabling the world’s most talented humans to work with the world’s most powerful AI algorithms will you be able to build a platform that’s capable of transforming entire industries.
Should the IPO go through as planned, C3 will trade under the only ticker that makes sense, “AI.”
Become a premium member and get access to hundreds of premium articles, reports and additional content.
Nanalyze Premium is your comprehensive guide to investing in disruptive technologies. Read by the top investment banks, management consultancies, VCs, and research houses. Trusted by over 100,000 institutional and retail investors. Covering disruptive technologies for nearly two decades.