Ondas Holdings Inc. ONDS

Key data

The business

Ondas runs two businesses that share little but a ticker. The original one, Ondas Networks, sells software-defined radios, its FullMAX system, that carry data across railroads and other critical infrastructure on a wireless standard the freight railroads adopted. The one the market now cares about is Ondas Autonomous Systems: the Optimus drone-in-a-box, a weatherproof station that launches its own drone to patrol a border, mine, or seaport and lands it to recharge with no pilot on site, plus Iron Drone, which knocks other drones out of the sky. Through 2026 the company bought its way into defense, taking in World View, Mistral, Bird Aerosystems, 4M Defense, and RoboTeam, turning a tiny wireless vendor into an autonomous-and-defense roll-up almost overnight. That is where the revenue came from. The moat is not the technology, which has real competitors. It is regulatory clearance to fly drones beyond the operator's sight and relationships with defense buyers, and neither is proven at scale.

The numbers

The revenue chart looks like a rocket, and almost all of the fuel was bought.

Sales went from $6.3 million to $83.8 million in these reported quarters, with the calculated fourth quarter of 2025 at $30.1 million. Strip out the acquisitions and the organic base is small: the jump is companies purchased with stock and cash, not customers won. The net-income line is noise. The $362.8 million profit in the March quarter was a non-cash gain on warrant liabilities, and the June quarter's $88.4 million loss sits on top of a $162.9 million operating loss. That operating loss, on $83.8 million of sales, is the number that matters, and it is getting worse, not better.

Five straight years of deepening losses tell you this has never been a profitable enterprise. The 2026 half-year swing to a $274 million profit is entirely the warrant mark. The operating business lost money every quarter. Gross margin is sliding too, as the amortization of intellectual property from all those acquisitions runs through cost of goods.

The Q2 10-Q shows the tell: cash fell $368 million in a single quarter, from $1.03 billion, so the comfortable runway that pile implied is already thinner than it looked. What the market is paying for is the leap from $174 million of trailing sales to the $1.3 billion that three analysts expect in 2028. The print that settles it is not net income, which the warrants control, but organic, same-company revenue and the cash line. One shows whether the roll-up sells anything on its own, the other shows how long it can fund the attempt.

Management

Founder and chief executive Eric Brock has built this by acquisition, and he has been a seller of the stock, not a buyer. Insiders sold about $37 million over the past year and bought nothing. Brock alone sold $31.9 million in early June 2026, with plan status not disclosed in the filing, right as the company's cash was draining. Stock-based compensation ran at 58% of revenue over the trailing year, and general-and-administrative expense rose 2006% as the acquired payrolls and deal costs landed. The pay and the dilution reward doing deals, not earning profits, and the deals are funded by the same shares holders own. Nothing here has yet been asked to stand on its own economics.

How it fails or surprises you

One customer is 40% of the business. A single customer, up from 19% a year ago, was 40% of second-quarter revenue, and two others that were half the business a year ago have faded below 10%. If that one program, likely a government contract carried in from an acquisition, pauses or reprices, the revenue line and the 2028 story break at once. Watch the next customer-concentration table and segment revenue.

Cash is draining faster than it earns. Cash fell from $1.03 billion to $658 million in the June quarter while the operating loss ran $163 million. A few more quarters at that pace, or one more acquisition, forces an equity raise, and with the stock at $7.62 and already halved, that raise dilutes existing holders hard. Watch the Q3 cash balance and the share count.

The defense stack lands a real program (right tail). If the acquired defense and counter-drone lines, Bird Aerosystems, 4M Defense, RoboTeam, and Iron Drone, win one large multi-year government contract, the $1.3 billion 2028 estimate proves conservative and the organic-growth doubt raised above is wrong. The market pays nothing for integration working because it never has. Watch for a contract award and organic backlog growth.

Closing thoughts

The payoff depends on two numbers that will print next quarter: the cash balance and whether revenue grows organically or only through more acquisitions. The reported profit is an accounting artifact, so ignore it. The real distribution runs between a defense roll-up that integrates and grows into its billion-dollar promise and a cash-burning collection of purchases whose biggest customer can walk. Near term the left tail is fatter: concentration and a $368 million quarterly cash drop are facts today, while the large contract that would justify the price is a hope. If the top customer pauses and cash forces a dilutive raise, the permanent loss is real. If the defense stack delivers, the upside is large, but its timing is unpaid-for and unknowable. Call the odds tilted to the downside over the next year, judgment not arithmetic.

The bet is still that Ondas turns a pile of just-bought drone and defense companies into one real business selling autonomous drones and counter-drone systems to governments and armies. What breaks it is the 40% customer leaving or the cash running out before the roll-up earns its own keep. The two numbers that tell you first are next quarter's cash balance and organic, same-company revenue. Net income will lie to you either way. If cash keeps falling near $300 million a quarter while organic sales stay flat, the story is already broken, whatever the headline profit says.