Disclosure: the author holds shares of Ondas (ONDS) as of the date of publication, and these holdings may change at any time without notice. Nothing herein is a recommendation to take any action in the security. See the full disclosure at the bottom of the page.
Back in March I published a deep research report on Ondas (NASDAQ: ONDS) - dozens of pages I called "the autonomy factory": a company that acquires capabilities, integrates them into a single body with one "brain," and builds a system-of-systems for the autonomous battlefield. Four months have passed since then - and in Ondas's world, four months are an era. Guidance tripled, the backlog jumped seven-fold, a U.S. company was acquired for $876 million - and the stock? It actually fell. That strange gap is exactly why this review exists: to re-examine the thesis, honestly, on the facts alone - what held up, what changed, and where the real failure points are.
Every figure here comes from primary sources: the SEC filings, company releases, and the investor presentation published on July 20.
What the Company Does Today: Four Platforms, One System
Anyone who met Ondas two years ago knew a small communications company with a drone arm. That is no longer the story. After a wave of 15+ acquisitions, the Ondas of July 2026 is a autonomous defense platform operating in 45+ countries, with 1,700+ employees and 25+ offices - and a core development and manufacturing hub in Israel. It organizes its offering into four integrated platforms:
1. Persistent Intelligence (ISR). Layers of collection - from ground sensors (Insight), through the "drone-in-a-box" (American Robotics' Optimus), tactical UAVs (Talon), Group 4+ long-endurance aircraft (DZYNE's ULTRA and LEAP), up to stratospheric balloons (World View's Stratollite). The goal: continuous intelligence coverage from the ground to the edge of space, without expensive satellites. And a new element that stands out in the July investor deck: in the multi-domain vision diagram (SkyWeaver), Ondas extends its coverage stack one layer higher — the Space domain, with a satellite, above the stratospheric balloons. At this stage this is an architecture/vision layer, not a shipping satellite product — but it signals the ambition to extend continuous intelligence into space itself.
2. Aerial Security and Counter-UAS. The layer where Ondas already proves global demand: Sentrycs (cyber takeover of drones), Iron Drone (kinetic interception), DZYNE's Dronebuster, BIRD's laser system, and the IonStrike interceptor. Its systems protected the Davos forum and the 2026 World Cup this year - and are being deployed at airports in Europe.
3. Precision Strike (Loitering Munitions). The domain that grew from almost nothing in my original thesis into the pillar of the backlog: short-range loitering munitions (BLITZ swarms), medium-range (Skylance) and long-range (Talon DT-300), alongside Rotron, which makes the engines. This is currently the largest slice of the backlog: $229 million out of $457 million.
4. Ground Robotics. Roboteam's RT family, robotic demining (Digger), and INDO's heavy engineering vehicles - including a rapid-deployment container with a command center. This is where the big early-year tender wins sit (~$220 million cumulatively for 4M Defense and INDO).
And above it all - the layer that turns this into a single thesis rather than a collection of companies: the software. SkyWeaver, the platform built together with Palantir on top of AIP, synchronizes all the layers into one operational OS with agentic AI at the edge. In the presentation, Ondas says it in investor language: an outcomes+software pricing model, and the "switching costs" the software creates. This is precisely the "battle of the brains" I described in March against Anduril's Lattice - and the company now says out loud that this is the field it is playing on.
What Happened Since March: The Timeline That Explains Everything
- January: A massive raise of ~$1 billion ($959M net); the deconsolidation of Ondas Networks (more on that below); revenue target of $170-180 million.
- March-April: Closed acquisitions of Mistral (a U.S. prime contractor, a partner in a $982 million U.S. Army IDIQ program for loitering munitions), BIRD, Rotron, INDO and World View.
- May 14, Q1 report: Revenue of $50.1 million - 10x the year-earlier quarter, and 25% above the top of the company's own guidance. A single quarter that matched all of 2025. Guidance raised to $390 million.
- May-June: Orders at a record pace - $30M+ in May, $40M+ in June (led by loitering munitions in Europe and the U.S.); World View selected for a Navy Southern Command maritime intelligence program; Sentrycs integrated into Lockheed Martin's Sanctum platform; the Cyberhawk acquisition announced (critical-infrastructure intelligence - software, data and AI; $95M backlog, closing in Q3).
- July 2: Closed the acquisition of DZYNE Technologies for $875.8 million ($200M cash + ~$676M in stock). Together with World View it formed the Ondas Sentinel division for the U.S. defense market, led by Ryan Hartman (formerly of Insitu and Raytheon). Guidance was raised again - to $525M+.
- July 20-22: A $6.9 million Australian defense order, and the announcement of $70 million in orders over the past month - total announced Q2 orders: $150M+.
The Acquisition Timeline: How "the Factory" Was Built
To understand the Ondas of today, it helps to see the whole acquisition journey on one line - from the foundations to the 2026 blitz:
| When | What was acquired | What it added |
|---|---|---|
| 2021 | American Robotics (US) | Autonomous industrial drone; later the first on the Blue UAS list for fast federal procurement |
| Early 2023 | Airobotics (Israel) | "Drone-in-a-box" - full autonomy without an operator |
| 2024-2025 | Assembling the Israeli platform | Iron Drone (interception), 4M Defense (engineering & demining), Insight Sense (sensors), Apeiro Motion and more |
| Nov 2025 | Sentrycs - ~$225M | Cyber counter-drone (Cyber-over-RF); deployed in 25+ countries |
| Nov 2025 | Roboteam - ~$80M | Military ground robotics (the RT family) |
| Mar 11, 2026 | BIRD Aerosystems | Laser protection for aircraft and helicopters, anti-missile capabilities |
| Mar 16, 2026 | Rotron Aerospace (UK) | Engines for loitering munitions - control of propulsion |
| Mar 17, 2026 | INDO Earth Moving | Heavy engineering vehicles; following a ~$140M win |
| Apr 2026 | Mistral Inc. + World View (US) | Pentagon prime contractor ($982M IDIQ) + stratospheric ISR balloons |
| Jun 18, 2026 | Cyberhawk (announced; closes Q3) | Critical-infrastructure intelligence - software, data and AI; $95M backlog |
| Jul 2, 2026 | DZYNE Technologies - $875.8M | ULTRA/LEAP aircraft, IonStrike, Dronebuster; the basis for the Sentinel division |
15+ acquisitions, and a consistent pattern: each one closes a link in the kill chain I described in March - find, track, decide, engage, assess - or adds a coverage layer (ground, low air, high air, stratosphere). This is not a pile of companies; it is a puzzle with a picture on the box.
Management: Who Steers the Ship
The July presentation shows a leadership bench that expanded with the deals:
- Eric Brock - Founder, Chairman and CEO. An entrepreneur and investor with 30+ years of experience, the architect of the Core + Strategic strategy.
- Neil Laird - CFO. 25+ years in public technology companies.
- Oshri Lugassy - Co-CEO of the OAS division. A military command background and decades in autonomous systems and global business development - the face of the Israeli core.
- Ryan Hartman - President and CEO of the new Ondas Sentinel. 25+ years in aerospace and defense, formerly senior at Insitu (Boeing) and Raytheon, plus U.S. military service - exactly the profile needed to sell to the Pentagon.
Alongside them, the deals also brought key people such as Eyal Banai (Mistral's CEO, who joined with the merger) and directors like David Chinn who strengthened the OAS board. That blend - "the best Israeli minds" alongside U.S. industry veterans - is itself part of the thesis: you cannot sell to the Pentagon without Americans, and you cannot develop fast without Israelis.
The Numbers, Eyes Open
The growth is real. $50.1 million in the quarter versus $4.3 million a year earlier (+1,065%) and versus $30.1 million the prior quarter (+66%). Gross margin climbed to 49.2% (from 35% a year ago) - a sign that in-house production and mix are improving, as I estimated in March under "quality growth."
And the loss is real too. An operating loss of $42.7 million in the quarter, and adjusted EBITDA of negative $10.9 million (versus negative $7.5 million a year earlier). The company itself says: Q2 is expected to be the peak of the losses - expenses running ahead of revenue, ahead of the second-half ramp. One positive milestone was already recorded: the product companies reached positive adjusted EBITDA six months ahead of target, and the target for full OAS profitability was pulled forward from Q3 2027 to Q1 2027 (and the whole company - by Q1 2028).
And now the big asterisk. The bottom line shows "net income" of $361 million - but do not be fooled: it includes a non-cash accounting gain of $389.5 million from revaluing the warrants from the October and January raises, plus $51.5 million from the Networks deconsolidation. This is the same mechanism we saw this week at Intel, but in the opposite direction - a revaluation that swings with the share price without touching the business. Anyone reading Ondas through the bottom line, positively or negatively, is reading it wrong. The right metrics: revenue, gross margin, adjusted EBITDA, and backlog.
The balance sheet is genuinely strong. $1.48 billion in cash and investments, against total debt of just $4.4 million - almost no leverage. But on the liability side sits a warrant liability with a fair value of ~$1.1 billion - a reminder that the big raise was not "free money," but came with conversion rights that will hover over the capital structure for a while.
Testing My March Thesis: What Held Up, and What Didn't
This is the heart of the review, and I am personally committed to examining myself in public.
What held up - and then some:
- "The autonomy factory" works. The central thesis - acquire capabilities, integrate, sell as one system - is translating into numbers: the company calls it the "growth double dip" (acquired companies' revenues accelerate inside the platform beyond their standalone plans), and Sentrycs alone received year-to-date orders exceeding all of its 2025 revenue.
- Backlog as a certainty engine. In March I wrote that the backlog is the "promise" of hitting guidance. It jumped from $68.3 million to $457 million pro forma (7x), and with DZYNE ($111M) and Cyberhawk ($95M at close) - close to $660 million effectively.
- "Hidden growth" materialized. I wrote that the guidance then did not include the March acquisitions and that consolidated revenue might cross $250 million. Reality beat even that: $525M+.
- The operational track record as an edge. Combat-proven systems from Israel are now selling to Davos, the World Cup, Australia and European airports - exactly the marketing edge I described.
What changed or needs updating:
- Ondas Networks is no longer here. In March I devoted a whole chapter to the barriers of the communications division (FullMAX, the rail cycles, the FCC). On January 16 the company executed a capital restructuring that removed Networks from the consolidated financials (with a one-time $51.5M gain). The Ondas of today is effectively a pure defense play - the thesis should be read that way.
- The dilution - the bear case that was right. In March I flagged the option structure of the raise as a central dilution risk. It happened: the share count jumped from ~464 million to ~570 million (+23%), and the DZYNE deal paid ~$676 million in additional stock. And here is the explanation for the strange gap from the opening: market cap barely moved since March (~$4.6B versus $4.7-4.9B) - but it is divided across many more shares. The stock fell from $10.8 to ~$8 not because the business weakened, but because the pie was re-sliced.
- And here is the flip side of that same coin: in March the company traded at a forward sales multiple of ~27 (on $180M guidance). Today, on $525M+, the forward multiple is ~8.7 - a dramatic valuation compression that occurred while the business actually strengthened. That can be read as an opportunity or a warning - depending on your thesis, which we will unpack next.
Anatomy of the Chart: The Price Action, Step by Step
And now the question every holder asks: what exactly happened to the stock? Here is the path, based on daily trading data:
- March (my research period): The stock traded around $10-11.
- May 13, the eve of the report: $8.86 - the market arrived pessimistic.
- May 14, Q1 report day: A jump of 26.5% to $11.21, on enormous volume of ~245 million shares. It touched $12.12 the next day.
- Late May - June 2, the rally: May's order wave (the $30M+ release on 5/29) and World View's Navy selection pushed the stock to a peak of $14.16 on June 2 (close: $13.58).
- June to mid-July, the slide: and here the dissonance began. The stock fell day after day - even on days of excellent news: on the Cyberhawk announcement day (6/18) it closed at $9.27; on the $40M+ order day (6/22) it fell to $8.89; on the DZYNE close (7/2) - $7.41. The low was set on July 17: $6.22 intraday (close $6.53) - a drop of ~56% from the early-June peak.
- July 20-23, the turn: the Australia order (+5.3% on 7/20), then a day of +11.5% on 7/21 on huge volume of ~284 million shares, and the $70 million order announcement that brought the stock back toward $8.
What explains a slide amid a downpour of good news? The supply side. January's raise was done with a single institutional investor, in a structure of 19 million shares + triple-layered warrants (pre-funded warrants for 41.8 million shares, and a warrant for two more shares per unit). Since June, an almost weekly stream of resale prospectuses (424B7) at the SEC registered shares for sale - and in parallel, DZYNE deal shares were added. When supply grows faster than demand, the price falls even as the business improves - and the enormous volumes (60-280 million shares a day, on a ~570 million base) show the stock genuinely changed hands at an unusual pace. It is not pleasant, but it is a market mechanism - not a statement about the business.
My Angle (a personal opinion of Ilan Abramov - not advice, not a recommendation)
Ondas is a company that from the start I said, out of business understanding, is building a system of systems. It acquires businesses out of the understanding that each is a piece of a puzzle - and its aim is that the whole be greater than the sum of its parts. Along the way it keeps adding more and more, and sometimes that hurts the share price, because this is what growth looks like and it can be painful - you have to understand which business you are getting into.
Along the way, all of its moves are justified in my eyes; I understand why they do each thing and the reasons are justified. More than that, I can see how each of its moves complements the others, and it will take time until that happens fully - but I can see that management, made up of the best Israeli minds and others, seems to know what it is doing - steering the ship properly.
But the road is not pleasant. The dilutions can hurt the share price in the short term, and additional forced pressure on the price from the selling of one institution (the same one that provided the over-$1 billion investment this past January) affects us drastically. The price falls from around $14 after the last report to the $6 area - but the business at its core has not changed, and has even improved: more acquisitions that complete the synergistic array (you can see this in the opening of the new division that integrates the DZYNE deal and World View), and signing a deal with the Australian Ministry of Defense - in such a short time.
There is no doubt that in the short term this is complicated - but how can you build a good, real business in half a year? It takes a long time. And I am willing to be part of this journey.
One more word on Eric Brock: he is an active, involved CEO, and you can see it easily - his X platform and the interviews he gives show his strength as a CEO. And one thing that is a green flag for me: every time they reach a point where they do another positive thing, like a huge deal (you can see it in the release stating that over the past four weeks they closed $70 million in contracts) - he immediately writes "well done, but there is still work to do."
(Important to stress: this is my personal opinion only, I am in a position at the time of writing, I may be wrong, and there is risk.)
Success Points and Failure Points
Proven success points (from the filings):
- Execution above guidance - Q1 beat the company's own top end by 25%.
- Consistently improving gross margin: 35% -> 42% -> 49.2%.
- Real U.S. penetration: Mistral as a prime, a $982M IDIQ, World View in a Navy program, and now Sentinel with customers like the USAF, Army, Navy, USSOCOM, DARPA and NASA.
- A partnership ecosystem rare for a company this size: Palantir (SkyWeaver), Lockheed Martin (Sanctum), Heidelberg (ONBERG in Germany), Nammo.
- Financial firepower: $1.48 billion in cash, almost no debt - "ammunition" for acquisitions without near-term dependence on capital markets.
Failure points and risks (eyes open):
- Ongoing dilution - 464M -> 570M shares in four months, a $1.1 billion warrant liability, and an almost weekly stream of resale (424B7) filings at the SEC. Holders need to understand: there may be more.
- The DZYNE lock-up expires in early 2027 - more than half of the consideration shares (out of ~$676 million) are locked for only six months. A known, pre-scheduled supply point.
- Integration overload - 15+ acquisitions, and now a company the size of DZYNE too. In March I called this "management insanity risk," and it has only grown. Any integration failure will surface late.
- The losses are still here - negative adjusted EBITDA set to deepen in Q2 per the company itself. If the H2 ramp is delayed, the profitability point (Q1 2027-Q1 2028) recedes, and the market will not forgive.
- Dependence on backlog conversion - $457 million of backlog is a promise, not revenue. The conversion pace in the second half is the real test of the $525M guidance.
- A multiple that still requires faith - 8.7x sales is cheaper than 27x, but expensive versus traditional defense names; the valuation still prices in near-perfect execution.
The Bull Case vs. the Bear Case - July Edition
The bulls will say: this is the company that analysts (8 buy ratings, zero sell, an average target of ~$20; Northland at $18, Oppenheimer at $16) call "the only scaled independent autonomous-defense platform" - in a market undergoing a revolution of "affordable mass" and rearmament that we wrote about in our defense coverage. The growth is proven, the backlog provides certainty, the cash funds the expansion, the software (SkyWeaver) builds a moat, and the multiple was cut 70% while the business tripled. The dilution? The price of admission to the big leagues.
The bears will say: this is an acquisition machine funded by printing shares, whose net income is an accounting illusion, that has not yet proven a single quarter of positive company-level EBITDA, and whose entire guidance rests on an unprecedented integration of 15+ companies at once. The history of defense roll-ups is full of entities that collapsed under the weight. And the DZYNE lock-up expiring in early 2027 is a cloud everyone can see.
The debate in one line: the bulls see a platform that proved execution, tripled guidance and compressed in valuation - a rare combination. The bears see a dilution machine with deepening losses and an integration impossible to prove in advance. Both agree on one fact: the second half of 2026 will decide - the orders are already in backlog, now they must deliver.
The Sector Catalyst: What Anduril Could Do for the Whole Field
And one more angle, fresh from today. A report has emerged of Anduril in talks to raise at a $100 billion valuation - and it recalls exactly what SpaceX did for the space sector: when the private leader gets a premium valuation, the entire public cohort gets a new reference frame, and attention and capital begin to trickle down to the smaller players.
And here is the point for ONDS. While private valuations in the sector soar, many of the small public companies in the field have actually suffered sharp declines - not a few of them over 50% off their highs - amid dilutions, raises and volatility. Ondas among them, down ~56% from its June peak. In other words, a clear valuation gap has formed: the private leader trades at a dream premium, while the public cohort - which you can actually buy - trades at a deep discount. If and when the new reference frame trickles through, and a catalyst like Anduril can accelerate it, the companies hit hardest on valuation are the ones with the most room to close the gap. This is not a promise - it is a sector dynamic worth watching, and it adds another layer to our thesis.
Summary: What I'm Watching From Here
Four months after the research, my central thesis - "the autonomy factory" - not only survived the test, it accelerated beyond what I estimated. But this review also recalls the side that demands humility: the dilution I foresaw as a risk materialized fully, and it - not the business - is what dictated the share price. From here, four watch points:
- The Q2 report (expected mid-August): the backlog-conversion pace, the depth of the loss peak, and reaffirmation (or a raise) of the $525M guidance.
- The Cyberhawk close in Q3 and progress on integrating DZYNE into Sentinel.
- The capital structure: warrant exercises, the pace of resale filings, and the DZYNE lock-up expiry in early 2027.
- The flagship contracts: the Army IDIQ ($982M), the border programs, and any first major U.S. government contract for Sentinel.
This is a company in the middle of the biggest leap of its life - and those leaps, by nature, are not clean. Anyone who holds (and I hold, as disclosed) should know exactly what they hold: not a stock of a stable company, but a ticket to a managed bet on the decade of defense autonomy - with all the potential, and all the risks, that entails.
Sources: Ondas Inc.'s Q1 2026 report as filed with the SEC (the 8-K of 5/14/2026 and its exhibits) and the 10-Q; the official investor presentation of July 20, 2026 and the DZYNE acquisition deck of July 6 (fact sheet, 8-K); the company's official releases from ir.ondas.com (May-July 2026), including the order announcements of 5/29, 6/22, 7/20 and 7/22; market and analyst data from ongoing financial coverage. All figures are current as of the date of writing. The chart is shown in real time via TradingView.
Extended disclosure: the author holds shares of Ondas (ONDS) as of the date of publication, at an average entry price above the current market price, and may buy or sell shares at any time without notice. Nothing in this review constitutes investment advice, investment marketing, or a substitute for personalized advice that accounts for each individual's circumstances. The author is not a licensed investment advisor. The review relies on public sources believed to be reliable but does not purport to be complete, and includes estimates and forecasts that are "forward-looking information" whose realization is not certain. See the full disclaimer in the original Hebrew article.