This article was originally published on August 24, 2023, when Shah Gilani was making a very similar pitch for these same three stocks. The current version of the tease is undated, and not quite as long, but he’s still hinting at the same stocks. What follows has not been updated since it was originally published, and my opinions on these small caps haven’t changed, but I have added a brief update at the bottom.

Shah Gilani is selling an upgrade newsletter with a pitch about “stealth stocks” in artificial intelligence (A.I.), and the ads started making the rounds while I was on vacation… so I’m catching up to answer some reader questions. Here’s a little bit of the intro to give you a taste:

“People like Bill Gates and Peter Thiel – the world’s biggest investors – they’re putting millions and billions into serious AI projects.

“They’ve got the inside track on small companies working on sensitive national security campaigns, military maneuvers, and top-secret AI chip technology with geopolitical implications.

“I’ll tell you this – the companies truly revolutionizing AI aren’t babbling to the media or hiring PR firms to churn out press releases.

“They’re silently and diligently putting all their effort into their technology – tech that will create world-class businesses in markets worth tens of billions of dollars… which means these companies aren’t plastered all over the headlines yet.”

Maybe so, though, of course, the risk of failure is dramatically higher with small startups — there’s an inherent tradeoff between looking for the highest possible returns, and looking for the highest probability of positive returns. Home-run hitters usually strike out a lot.

Gilani is using this pitch about “AI stealth stocks” to sell what he says are charter subscriptions in a new “upgrade” newsletter — this time it’s called L.A.U.N.C.H. Investor ($1,497 for the first two years, then renews at $1,497/yr, no refunds)… though it looks to me like this is really more of a “rebrand” of Gilani’s Extreme Profit Hunters than it is a brand-new service.

Here’s some more of the above-the-fold hype:

“AI could be the greatest wealth creator we’ve ever seen. It is truly world changing.

“The economic impact of AI is projected to hit $15.7 trillion by 2030. That means AI will be the biggest capital wave in history and the greatest stock-buying opportunity of our lives.

“The problem is… most people are investing in the wrong companies – the blue chips: Microsoft, Google, NVIDIA, or Amazon – thinking that’s where the biggest AI wealth is falling. That’s not what I’m after… not even close.

“Not when young companies with transformative AI can deliver 10X, 20X – 30X – gains.

“A few years ago, my team told you to invest in this when the stock was around $2, and the folks who listened and held this stock for the long haul have seen a stunning 3,200% gain – enough to turn a mere $3,000 into over $100,000. [that seems to be a transcript error, I assume he’s talking about a past recommendation of AMD with that reference]

“And I think it’s going to keep going. But I also want to tell you about something even bigger – a lot bigger.”

Then Gilani pretty quickly jumps into the teases about his specific A.I. “stealth stocks” …

“AI Breakthrough Stock #1 – The WarWise Military AI with Almost a Billion in Pentagon Contracts

“Three-Year Growth Target: 1,100%

“The first company already has signed contracts equivalent to double their market cap, and it’s all because of a fact you need to understand as soon as possible.

“America’s military is on a secret AI-spending boom of epic proportions. The Pentagon requested $874 million for AI last year. DARPA – the innovation arm of the U.S. military – is expected to spend over $2 billion on AI tech over the next two to three years…

“… the Department of Defense knows the breakthrough AI is coming from new, innovative companies most people have never heard of – including this first AI company I want to tell you about who just vaulted to prime contractor status with their instant battle plan AI technology I call WarWise.”

What’s this “WarWise” product? Apparently some distributed data collection and interpretation system for the battlefield…. here’s how Gilani puts it:

“This kind of AI can make decisions a million times faster than humans can.

“That’s why it’s being called the U.S. military’s triple-threat system… because one: It visualizes, details, and maps out an entire battlefield in ways humans can’t – with millions of data points that can be used to supercharge decision making.

“Two: It uncovers new developments from aerial reconnaissance and makes exceptionally accurate predictions.

“And three: It responds to changing conditions to update battle plans with new tactics. And it does all this in one-billionth of the time MacArthur, Patton, Eisenhower, or the most brilliant military tactician could ever dream of.

“The AI from this little company is so remarkable it could tell our armed forces how and where to eliminate the enemy almost instantaneously. So you can see why an organization like DARPA would spend years trying to develop this.”

Other clues about the company behind it? This is what we get…

“… it already has near $1 billion in Pentagon contracts…

“… less than .04% of Lockheed Martin’s equity valuation….

“… after winning the Advanced Naval Technology Exercise AI Prize Challenge, they won a sole-source multimillion-dollar contract with the U.S. Army.

“They have a five-year contract with the NSA to develop and support strategic priorities and operations. Their AI tools were selected by DARPA for an intersatellite laser communications terminal… and earlier this year, they were awarded a 10-year near billion-dollar contract by the U.S. Air Force.”

And he says it sells “for a few bucks” — and the CEO is experienced in A.I., since she came from IBM and worked on the Watson project… and the company had $255 million in revenue last year. Gilani says she’s also planning on extending these A.I. tools to private industry clients, which is similar to the trajectory Palantir (PLTR) has been on, moving from being a government contractor to also selling data processing tools and services to corporations… and there’s already been some progress on that front, with their “AI for hospitals” being used by hundreds of health care organizations.

So what’s the stock? Well, it’s nowhere near $5 at the moment, as the bloom has come off the rose since their first surge as an AI story stock earlier this year (it was over $5 for a few weeks in February, when investors were chasing any company with “AI” in its name), but here Gilani is teasing BigBear.ai (BBAI).

How do we match the clues? Well, Gilani either had a typo or threw in a red herring, since they actually had $155 million in revenue last year, not $255, but otherwise we’ve got a perfect match — BigBear.ai’s CEO, Mandy Long, did indeed come from IBM, where she worked largely on the Watson AI project, with at least some focus on healthcare. And they did indeed win an AI prize challenge with the Navy about a year and a half ago, it was actually called the “Artificial Intelligence and Networks Advanced Naval Technology Exercise (‘AINetANTX’) Artificial Intelligence (‘AI’) Prize Challenge,” and last year they won a $14.8 million contract to lead an Army contract called Global Force Information Management (a phase 2 contract, they’re the sole prime vendor), and just got a six-month extension on that (for another $8.5 million) a couple months ago when the first nine months ran out.

BigBear went public through a SPAC merger in late 2021, and over the past few months they’ve focused their strategy on targeting industries where they see opportunity to implement advanced machine learning — so they are looking now to build in supply chains/logistics, cybersecurity, and autonomous systems, particularly in the military, in large part through a partnership with L3Harris.. and they do see the growing federal AI spending as a substantial driver that’s growing their addressable market. Here’s how BigBear describes itself:

“A leader in decision dominance for more than 20 years, BigBear.ai operationalizes artificial intelligence and machine learning at scale through its end-to-end data analytics platform. The Company uses its proprietary AI/ML technology to support its customers’ decision-making processes and deliver practical solutions that work in complex, realistic and imperfect data environments. BigBear.ai’s composable AI-powered platform solutions work together as often as they stand alone: Observe (data ingestion and conflation), Orient (composable machine learning at scale), and Dominate (visual anticipatory intelligence and optimization).

“BigBear.ai’s customers, which include the U.S. Intelligence Community, Department of Defense, the U.S. Federal Government, as well as customers in the commercial sector, rely on BigBear.ai’s high value software products and technology to analyze information, identify and manage risk, and support mission critical decision making.”

So yes, they are building AI capabilities for the military and corporate customers, and they do have a meaningful backlog (they say it’s $206 million now, having grown 5% over the past year), and there are some larger long-term contracts which could grow significantly beyond that (Gilani’s tease of “near $1 billion of Pentagon contracts” must be a reference to the 10-year $900 million contract they got with the Air Force back in January, most of which wouldn’t be firm enough now to call it an “order backlog”).

The challenge has been in generating revenue or earnings growth that is as exciting as the “.ai” at the end of their name. They do have solid contracts, they do seem to be growing pretty well and to be building capacity for larger contracts in the future, particularly with the military, but they had $145 million in revenue in 2021 and $155 million in 2022, with analysts forecasting $165 million in 2023, so the top-line growth isn’t setting anyone’s hair on fire… and they’re not profitable. There are only a couple analysts covering this small-cap story, so we should take predictions with a grain of salt, but those analysts don’t see them generating positive cash flow until next year sometime. And the SPAC deal was late enough in the hype cycle that they didn’t get as much cash as they had hoped for back in late 2021 (many of the SPAC holders redeemed for cash instead — in retrospect, a smart move, so instead of starting out in late 2021 with $326 million in cash and a cleaned-up balance sheet, they got about $50 million in cash and bumped their $100 million in borrowing up to $190 million). Right now, they’re down to about $30 million in cash now and they’re still carrying about $190 million in long-term debt (I didn’t look up the debt terms, but they’re probably pretty friendly). That’s not likely to be a crisis, since firms with steady government contracts are generally a good credit risk, but if business doesn’t pick up or become a bit more efficient they could burn through their cash in just 2-3 quarters at this rate, and raising more capital would probably pressure the share price unless they’ve got some good “headline” news by then.

Their last quarter, announced a couple weeks ago, was not dramatic — their results were a hair lower than the estimates, and they continue to offer guidance for the year of “between $155-170 million” in revenue, which means the year might be flat or might have almost 10% revenue growth. Their margins got a little worse, partly because Virgin Orbit, one of their clients, went bankrupt, but they did cut overhead costs over the last year so their “adjusted losses” have been shrinking.

Probably some potential if they can scale their relatively small contracts into larger orders, though that would probably also involve ramping up their cost structure to handle more work, and the balance sheet is a little bit precarious so that might be a challenge. I haven’t seen enough to get me interested at this point, not with the current level of cash burn, the tight balance sheet and the lack of substantial revenue growth, but, well, I guess that’s why it has traded down to a $200 million market cap at $1.38 per share, trading at less than 1.5X sales and less than the value of their order backlog. Building scalability is really tough with contractors, since they’re essentially providing a service and often have a “cost+” mindset to just make sure they break even or make a little money on new contracts, instead of reselling their existing capabilities or products and steadily improving margins, like we typically see with small tech and software stocks, but it’s certainly possible.

And how about number two? This one’s about security scanning…

“AI Breakthrough Stock #2 – the Impenetrable Security AI with Skyrocketing Revenue and Backlog

“Two-Year Growth Target: 1,200%

“Ever wonder why it takes so long to get through security checks at airports around the world?

“Metal detectors and X-ray machines – that’s why. They’re slow. They’re invasive. They’re completely outdated. The fact is… they’re just downright frustrating for the average person to deal with.

“But worst of all… they’re not accurate and cause all kinds of false positives.

“Fortunately, we now have an alternative – an AI alternative that could make you rich. It’s one of the biggest breakthrough AI devices I’ve ever seen.”

So he’s talking up a company with a better security scanning technology than is now in use at airports and other high-security locations:

“I believe this new AI technology is on the cusp of replacing every TSA screener in every airport in America. It’s 1,000 times more advanced at detecting illegal and dangerous weapons.

“And it’s fast. In fact, it’s 10 times faster than ordinary metal detectors. That means no more delays as security guards run you through a machine or wave a wand all over your body or – for crying out loud – have to pat you down.

“This AI device scans everyone with speed, safety, and gives ultimate security.”

And he makes some big-picture promises:

“Once these devices are in airports across America, you can bring in your water… leave your shoes on… and pass through quickly – without worry. For law-abiding citizens, it’s going to make getting through airport security a breeze – not to mention make flying on a plane so much safer.

“And it’s not just airports. This AI is also being used to protect law-abiding American citizens from random gun violence.”

He says specifically that “schools in Atlanta have just started using their technology” and that it’s 10X faster than metal detectors… and that he thinks this AI company is primed to grab a big chunk of the $150 billion in federal money that’s being allocated to improve school safety.

And he mentions some high profile investors:

“… former Florida Governor Jeb Bush, tool conglomerate Stanley Black & Decker, Super Bowl-winning quarterback Peyton Manning, and Bill Gates too.

“My research shows Gates recently purchased 2.5 million shares.”

And provides some financial info that we can feed into the Thinkolator…

“In the fourth quarter of last year, revenue shot up 217% year over year. Sales rocketed 222%.

“Plus, they’re sporting $465 million worth of backlog, and they’re just scratching the surface… a shot at completely overhauling America’s entire airline security – a $20 billion market… and another potential $150 billion from America’s schools.”

And he thinks there’s tons of opportunity in hospitals and sports arenas… and highlight their work with the Cleveland Browns:

“… this company’s AI security technology is already being used at First Energy Stadium – home of the NFL’s Cleveland Browns.

“In their first year as a customer, they shrank their security footprint from 100 metal detectors down to 12 AI-enabled systems. And in their first test, they screened a whopping 18,000 guests in a mere 45 minutes.”

That’s an embarrassment of clues, so we barely had to wait 30 seconds to get a good Thinkolator answer: Here Gilani is teasing Evolv Technologies (EVLV), which is indeed a leader in “AI-based weapons detection security screening.” And, no surprise, is yet another of the stocks that went public through a SPAC merger in mid-2021 (and yes, it received venture funding from Bill Gates and Peyton Manning and Jeb Bush’s firm before that — though none of those folks seem to be major holders of the shares now, it looks like they all registered to sell last year, don’t know if they actually sold out completely).

This is an area where a lot of hype-filled stories have spun up out of nothing in the past, and also quickly evaporated, like Patriot One (now Xtract One (XTRA.TO, XTRAF)) or VSBLTY (VSBY.CX, VSBGF), so my initial response to these kinds of technologies is always somewhat skeptical… but there have also been pretty successful businesses built up in this area in the past, and perhaps Evolv will be the next one, even if they don’t seem particularly close to getting the TSA’s passenger screening business.

Evolv is gradually growing, as some of their test installations become larger contracts with recurring revenue, though they’re not getting any closer to profitability, since their overall costs are growing at least as fast as their revenue in most quarters (though their gross margin has improved this year, so that might mean they’re at least not subsidizing the hardware as much as they did a year ago). There is not much analyst coverage, but the few firms who follow the stock expect that revenue will grow at about 50% per year for the next few years, which might make it possible for them to be close to break even sometime in 2025. They might have enough cash to get there if they can grow their installed base without ramping up the cash burn, since their installations come with a SaaS-like deferred revenue stream that’s promised for future years, but there is some capital intensity here, they must need to spend fairly heavily to install their security systems at new client locations, so more likely they’ll need more capital to grow at some point. Right now, they are valued at almost a billion dollars, roughly 12X trailing revenues, so a lot is riding on getting new contracts up and running with schools and stadiums.

They explain their Evolv Express system here if you’d like to get more background, and their latest investor presentation is here. There are also some July 2026 warrants trading, too, a remnant of the SPAC funding at tickler EVLVW, if you’re looking for a bit more leverage (they look like standard warrants, so they should have an $11.50 strike price and some early redemption rules if the stock gets above $20 for a while — which essentially means you win if the stock goes up 100-300% from here over the next three years, but if it returns less than that or more than that it’s likely that the stock returns will be better than or similar to the warrants… and if the stock never gets near $11 again (it’s around $6 now), the warrants would likely give you a 100% loss).

And one more? Here’s how Gilani teases it…

“AI Breakthrough Stock #3 – the “AI Vision” AI Chipmaker That Reminds Me of a Young Micron

“Three-Year Growth Target: 2,100%

“It’s an American company designing new technology that is solely responsible for making AI see….

“The AI chip war is happening right now. And it’s centered around these miraculous chips because they can process visual data at lightning-fast speeds for AI to interpret.

“We’re talking about trillions of operations every second. Stack that up against today’s top PC processor – which can do about 4 billion a second.

“They’re also the tiniest chips ever created. And at $10,000 a pop, they are going to make this company a fortune. They’re vital in an AI market because AI needs superfast chips that can see in order to process so much data.

“Think of it as the eyes of AI, and companies are scrambling like mad to get these chips to fuel their own AI projects. I’m talking about electric vehicles and autonomous cars – a market forecast to be more than $127 billion by 2027.”

OK, so some kind of sensor chips — maybe actual cameras and light sensors, maybe radar or LiDAR. What else?

“This tiny company is already partnering with Mercedes Benz, Ford, Rivian, Renault, and Toyota. And they’re being used in China – where the EV market is expected to hit $323 billion in sales by 2027….

“They’re also being used for traffic flow analysis to eliminate backups, road condition monitoring to get rid of potholes, and the automotive industry is just the beginning for this company’s AI vision chips.

“They’re also being used in X-ray analysis, cancer detection, self-checkout at grocery stores, automated product assembly, livestock health monitoring… the list goes on and on.”

And they have some kind of patented technology for these “vision chips”…

“… they’ve practically cornered the market with 299 patents. These patents on their AI vision chips could be the most valuable asset they own.”

And Gilani thinks they’ll be a big beneficiary of the CHIPS Act that’s driving more investment in US semiconductor capacity:

“Hundreds of billions of dollars more could pour into U.S. chip manufacturing. This AI vision chip company stands to be a huge benefactor. It’s as sure a bet as selling picks and shovels during the Gold Rush.

“Right now, this little company is dwarfed by Micron, AMD, and NVIDIA. Its market cap is a measly 2% of Intel.”

And he compares them to another giant:

“… they remind me most of Micron Technology. I’m talking about when Micron took the chip industry by storm with their innovative random access memory chip just when the PC revolution was getting started.

“In a little over two years, it jumped to a peak 2,100% gain. That’s 21 times your money. That’s what a stock like this AI vision chip company can do for you.

“It can take over an industry with something brand new, then take over the world in a couple of years. I believe this stock could be one of the top 10 biggest chip companies in the world.”

So hoodat? This is the oft-teased camera chip maker Ambarella (AMBA)… here’s how they describe themselves:

“Ambarella’s products are used in a wide variety of human and computer vision applications, including video security, advanced driver assistance systems (ADAS), electronic mirror, drive recorder, driver/cabin monitoring, autonomous driving and robotics applications. Ambarella’s low-power systems-on-chip (SoCs) offer high-resolution video compression, advanced image processing and powerful deep neural network processing to enable intelligent perception, fusion and central processing systems to extract valuable data from high-resolution video and radar streams.”

All of that makes for a good story, and from what I can tell Ambarella does have some cool technology, and certainly a lot of it is patented (they did have 299 patents issued or applied for as of about a year ago, I assume that number is higher now)… but despite the cool technology and the easy story they can tell (drones, self-driving cars, etc.), they have had a hard time turning that into profits over the years. The last time they reported a real profit was back in 2018, and rising expenses and slack demand for electronic devices this year have made things even worse recently. They certainly could bounce back, as more “internet of things” devices are installed to collect more data and as more autonomous devices rely on image-capture chips to understand the world around them, but the windfalls that Ambarella investors seem to have expected for a decade now haven’t come yet. Whether that’s because of competition from more commoditized imaging chips that are “good enough,” or because there are more advanced providers out there that I don’t know about, they’ve never been able to put together real revenue growth and margin improvement that would tell investors that the story about the quality of and demand for their chips and designs is real enough to turn into actual money. I really don’t know why, but, since we’re talking about the semiconductor business, I suspect it’s competition and pricing pressure from their customers that’s keeping them down.

Here’s what Ambarella’s financials look like over the past five years in visual form… gross margins (orange) have stayed around 60%, which is good for a chipmaker, and admirably steady, the share count has risen at almost exactly the same rate as the revenue, so their revenue per share (green) is up only about 7% in five years, and the operating margins have never gotten to be very good but are recently terrible (losing more than 30% on each sale). The story did heat up in late 2021, as so many did, so the stock soared for a while… but just going from the financials, the 84% move in the stock price over the past five years, which pretty well matches the S&P 500’s 69% total return, is quite a gift from a company that didn’t do much during that time.

The current expectation is that things will pick up for Ambarella — analysts are expecting them to be back to adjusted profitability next year (not counting stock-based compensation), and to grow that meaningfully in the future, so on an adjusted basis they’re trading at about 47X their recent peak adjusted earnings (2022), and the analyst estimate is that they’ll be back to that level of earnings in three or four years. That’s not enough for me to find this appealing, but that pickup from this trough level might be enough growth to get folks excited as we move forward… and they certainly might light a fire if they get much better orders because of the surge in AI investment, even though that doesn’t yet seem to be happening. Ambarella is in attractive businesses, and the story makes sense, it’s just that to my eye they’ve not yet been able to turn that into something that makes the income statement attractive.

So there’s your crop of “stealth AI” stocks pitched by Shah Gilani… anything sound appealing to you? Have other favorites that you think are a better bet? Please let us know with a comment below.

P.S. For those who are curious about the surging excitement of those three small cap “AI story” stocks, which were already well off their highs when Shah started pitching them in August, here’s how they’ve performed since we posted the first version of this article (on 8/24) — that’s the S&P 500 in orange:

Disclosure: of the companies mentioned above, I own shares of Google parent Alphabet, Amazon, and NVIDIA. I will not trade in any covered stock for at least three days after publication, per Stock Gumshoe’s trading rules.