C3 AI – AN AI PICKS AND SHOVELS PLAY

I’d love to tell you that I’d found the small cap version of Nvidia (NVDA – NASDAQ). But I haven’t. Sadly, there is only one Nvidia (Its stock has TRIPLED this year due to its computer chips being useful in Artificial Intelligence).

In fact, finding any small cap AI play–Artificial Intelligence–is no easy task.

AI is geared to large-cap businesses. The costs are HUGE!.

A single Nvidia H100 GPU costs $40,000. Amazon (AMZN – NASDAQ) and Microsoft (MSFT – NASDAQ) are advertising the ability to tap into as many as 10,000+ GPUs for training deep-learning models.

This is no country for small caps.

But AI is more than just high-end processing. Someone must do the dirty job of managing data.

That is where some small-ish cap names can find a niche.

One name I found is C3.AI (AI – NASDAQ). C3.AI labels themselves as an AI company. It is right in the name. The stock has been getting a lot of attention – from both longs and shorts – for that reason alone. But with over US$700 million cash and no debt, the stock is resilient.

But this isn’t the sexy GPU side of AI. C3.AI is more akin to a picks and shovels play that gets the mine built.

C3.AI offers AI-based enterprise tools to customers on a subscription basis. So they have their C3 AI Application platform, AND they have tools for customer relation management (C3 AI CRM), data science (C3 AI Ex Machina), and network analytics (C3 AI Data Vision).

THE MOST UNDER-APPRECIATED BIT OF AI?
THE DATA GATHERING

However, these tools aren’t what differentiates C3.AI. Where C3.AI has an edge–is processing data.

Investors–even industry–don’t get how big a deal this is. Implementing enterprise AI is more a data processing problem than anything. Compiling a vast range of data, found all over the company and in all sorts of formats, can be a nightmare.

This is exactly what C3.AI specializes in. Their customers include the Department of Defense (DOD), Baker Hughes (BH – NYSE), Dow Chemicals (DOW – NYSE) and Booz Allen (BAH – NYSE).

C3.AI helps companies work through the gathering and processing of data so that it can be used to train models.

HOW BAKER HUGHES USES AI
TO DETECT EQUIPMENT FAILURE

Here’s an example. Baker Hughes was an early adopter of AI. They partner with C3.AI, offering their tools to Baker Hughes customers.

One of those tools called BHC3 Reliability is geared at improving the reliability of industrial equipment.

The AI engine behind BHC3 Reliability can identify equipment/system behavior or system failures before they happen. It can then provide alerts, recommend remedies, and in some cases even perform the corrective action itself.

But before it can do all that, Baker Hughes needs to compile vast amounts of historical data to train the AI on how these processes should work and how they have broken down in the past.

This includes SCADA data, failure data, maintenance logs, climate data, vendor specifications – really, for the AI model to be robust, the more relevant data you can provide it, the better.

That is really where C3.AI comes in. Processing the data, much of it unstructured – datasets can even include email conversations or written reports – and prepping it for the AI model to train on.

By partnering with C3.AI, Baker Hughes customers have been able to identify anomalous operations and receive early warnings of critical issues with advance notice ranging from 10 days to 6 months.

So that is the kind of service C3.AI provides. It really is like digging out the dirt so you can start mining.

SAVING THE US MILITARY BILLIONS OF DOLLARS

What they do with the US Air Force is even more impressive. C3 works with the USAF to help pinpoint when important parts of planes will fail–before they do. They put together 440,000 machine learning algorithms and state-of-the-art NLP analytics, and trained 30 classifiers to calculate the probability of failure on high priority systems.

They have reduced unscheduled maintenance by 40% and increase aircraft mission capability by 6%. This defers capex, and theoretically the USAF needs 6% less planes now. They are reducing maintenance costs BIG TIME. The NPV on their work is in the billions of dollars. No wonder enterprise wide AI could be The Next Big Thing.

C3.Ai offers it as a subscription, a SaaS model. But it’s tough to call them a true SaaS company.

Professional services revenue has ranged from 14% to 18%, higher than a true SaaS company. Their software gross margins are a little lower than SaaS – at 75-80% and there is some indication that true software margins are even lower, around 65%.

I think this indicates that there is more manual heavy lifting than your typical SaaS company. In fact, that is one of the big arguments of the short sellers in the stock (more on that shortly).

C3.AI is effectively a data processing company, something that still requires a human touch.

PRE-RELEASED RESULTS

On May 15th C3.AI pre-released their fiscal Q4 results. They announced a slight beat on revenue and non-GAAP operating loss.

C3.AI said their business was accelerating, anC3.AI said that it was “more active than we have seen since the company’s inception”.

During the quarter C3.AI closed 43 deals, including 19 pilots that were initiated in Q4 FY 23.

They also announced a long list of enterprises converting from pilot to full-fledged paying customers during the quarter: Dow, Alberta Treasury Branches (“ATB”), and Chief Digital and Artificial Intelligence Office (“CDAO”).

Continued product additions and expansions at Shell, Koch Industries, Department of Defense Rapid Sustainment Office (“RSO”), PwC, Ball, ExxonMobil, Con Edison, Missile Defense Agency, Defense Counterintelligence and Security Agency (“DCSA”), Baker Hughes, New York Power Authority, and others.

I am not surprised by the momentum. Virtually every company is at some stage of AI implementation. I bet that most of them do not have the expertise to do it alone.

NOT A CHEAP STOCK

C3.AI has seen a parabolic rise over the last month:

Source: Stockcharts.com

The stock has quadrupled from its lows at the beginning of 2023.

Part of that is momentum from the AI tailwind. But it isn’t all just hype. C3.AI made a 180° change in their outlook earlier this year.

In August of last year, the outlook looked bleak. C3.AI warned of “a significant negative change” to their business, with lengthening sales cycles and economic headwinds.

The bad stock performance in the second half of 2022 was pricing in that weak outlook.

But on their fiscal Q3 call (in March) C3.AI flipped that on its head, saying they were now seeing “improved business optimism” and increased interest in their applications.

This was a $20 stock in August, before they began warning of a slowdown. Then it went to $10. Today it is a $40 stock. Given that the acceptance of AI solutions has improved significantly since then, this doesn’t seem like an outlandish sized move for a small cap NASD stock.

However, that alone doesn’t make C3.AI a cheap stock.

At $39, C3.AI has a market cap of $4.4 billion. They also had $770 million of cash at the end of January.

Earnings and EBITDA are both negative. C3.AI has forecast that they will be profitable on a non-GAAP basis in fiscal 2024 but it will likely be minimal.

That leaves us with a price to sales valuation. Average estimates for revenue for their fiscal 2024 (started in April) are $317 million. That puts the stock at nearly 14x P/S.

C3 AI is expected to grow revenue 19% in fiscal 2024. A 14x P/S multiple for that level of growth is not particularly cheap.

THE SHORTS HAVE A BULLSEYE ON THIS STOCK

Always be wary of shorts. They can influence trading and dig up dirt that the analysts miss.

C3 AI has its share of detractors. First among them is Kerrisdale Capital, which wrote a short report on the stock in March.

Kerrisdale made several (mostly fair) negative points about the company:

1. C3.AI has pivoted twice, first from utility analytics to IoT and now to AI – seemingly trying to capture the flavor of the day.

2. This is not a ChatGPT story – this really has little in common with the large language models that have captivated everyone.

3. C3.AI has seen a huge salesforce ramp with little to show in new customers.

4. Their AI tools are “meh” and many are not truly AI (more like machine learning).

5. The stock is not cheap.

It is hard to disagree with many of these points. The one I’d take most issue with is #3. As I already noted, the preliminary disclosure for fiscal Q4 hinted at growing sales traction.

The rest though, is fair game.

What the report misses is that none of it precludes C3 AI from the opportunity: To guide enterprises through the AI development process.

It’s an opportunity they seem to be capitalizing on.

The tools that C3.AI offers can reach a wide range of industries–inventory Optimization, Customer Churn Management, Predictive Maintenance, and Energy Management, to name just a few.

While we can debate whether these are AI tools or machine learning tools or some hybrid, but the growing customer base says they really are automating workflows.

C3.AI’s founder and CEO, Tom Siebel, is also no flash in the pan. Siebel is a billionaire founder of Siebel Systems. He founded C3.AI in 2009.

WHAT SHOULD INVESTORS DO?

While I see how this story could work, when I look at C3.AI, I am left thinking that the main reason the stock goes higher is because of the AI meme.

I find it hard to justify the valuation on numbers alone. 13x sales is what you pay for high flying SaaS names with stronger growth.

Source: Sentieo

These comps are well entrenched names in their respective vertical. All things being equal, C3.AI should trade at a discount, given the comparable track records.

That said, the stock is heavily shorted.

About 1/3 of the float is already short. That’s bullish–they’ll have to cover if this stock has a bigger run.

But I don’t think that this stock will trade on fundamentals any time soon. This has become a battleground stock in what is starting to become a battleground industry.

Which means that position sizing is key and extreme volatility has to be expected!

QUICK FACTS

Trading Symbols: AI
Share Price Today: $39
Shares Outstanding: 112 million
Market Capitalization: $4,360 million
Net CASH: $770 million
Enterprise Value: $3,590 million

POSITIVES

– Positioned to benefit from enterprise adoption of AI and machine learning
– Business is seeing growing momentum
– Obviously AI is the new hype word and these guys are right in the middle of that

NEGATIVES

– Not a cheap stock
– By some accounts their AI tools are not differentiated
– Target of short-sellers
– Huge run-up in the stock makes it vulnerable to disappointment

EDITORS NOTE–I cover small cap stocks in many sectors. One of my junior biotech picks is now up 40% in a month–and I think it can be one of my biggest wins in 2023! Get my report on this NASD-listed stock before it becomes a 10-bagger–RISK FREE!! Click HERE.