作者 Micah Zimmerman
地点 Alexandria, Virginia
Every artificial intelligence investing conversation these days seems to circle back to the same handful of trillion-dollar names that have already had their monster runs. That's understandable, but it also means the market isn't putting as much focus on smaller companies like Innodata (INOD +9.26%) and BigBear.ai (BBAI +3.40%) as they warrant. And those two just posted some of the clearest evidence yet that AI spending is reaching past the chipmakers and into the businesses that support them.
展开查看正文
Every artificial intelligence investing conversation these days seems to circle back to the same handful of trillion-dollar names that have already had their monster runs. That's understandable, but it also means the market isn't putting as much focus on smaller companies like Innodata (INOD +9.26%) and BigBear.ai (BBAI +3.40%) as they warrant. And those two just posted some of the clearest evidence yet that AI spending is reaching past the chipmakers and into the businesses that support them.
Both of these tickers are volatile, with bottom lines hovering between unprofitable and barely profitable, and they're nowhere near guaranteed to work out. But both also just delivered real, company-reported numbers that argue for investors taking a longer look.
1. Innodata is turning a services business into a platform
Innodata has spent years doing the unglamorous work of preparing and labeling data that large language models learn from, selling that work to big tech customers. In the second quarter, the company reported revenue of $92.1 million, up 58% from a year earlier, with adjusted earnings up 92%.
The more interesting shift is what Innodata is doing with that cash. Outgoing CEO Jack Abuhoff, who transitioned to executive chairman at the end of September as President Rahul Singhal took over as CEO, said the company closed its first engagement on a new self-serve data platform and had 15 more companies evaluating it, alongside talks with two hyperscalers about reselling access to it. That's a meaningful change in the business model: Instead of just renting out people and processes, Innodata is trying to sell software that its customers can use themselves. That type of business tends to carry better margins and stickier revenues.
It's not a finished story. A single customer still provided 37% of its revenue last quarter -- though that was down from 56% the quarter before -- and shares remain more than 45% below their 52-week high near $125 even after that record quarter, suggesting the market wants to see the platform business scale before it pays up for it.
2. BigBear.ai just landed a contract that the stock hasn't priced in yet
BigBear.ai (BBAI +3.40%) sells AI-powered analytics software to the U.S. government, and in April, it won a 10-year contract with the U.S. Air Force worth up to $900 million. Specifically, it's an "indefinite delivery, indefinite quantity" contract, government shorthand for a deal with a pre-approved ceiling that lets the Air Force award task orders to BigBear.ai without running a new competition each time. It doesn't guarantee $900 million in revenue, but it does mean BigBear.ai is now pre-qualified to bid quickly across a decade of Air Force AI work, rather than starting from scratch on every opportunity.
According to CEO Kevin McAleenan, the company's acquisition of generative AI platform Ask Sage last year will broaden its reach across national security customers, and BigBear.ai expanded that platform's capabilities for Department of Defense missions in July. Yet the stock trades near its 52-week low of $2.60, down from a high above $9, even with the backlog growing. BigBear.ai is also still deeply unprofitable, so the government pipeline will have to actually convert into cash before it means much.
What investors should do with this
Neither of these stocks should be picked as core portfolio holdings. Both are speculative, and for them to deliver 10x returns would require years of close to flawless execution from companies that are still small enough to be knocked off course by one lost customer or one delayed contract.
But if I'm putting speculative money into AI stocks at all, I'd rather put it into players like BigBear.ai. and Innodata than into stocks that have already experienced 10x run-ups premised on the assumption that everything will go right for them. When a stock has years of hoped-for future success baked into its share price, even good news can fail to lift it, and any business stumbles get punished hard. Innodata and BigBear.ai are in the opposite situation right now: Their businesses are producing real, verifiable progress, but their stocks are beaten down and don't reflect the good news yet.