Ambarella develops chips and software that enable cameras and machines to perceive and understand their surroundings, while processing data directly on the device. The company designs its own chips, outsourcing their actual manufacturing. Its technology—which integrates image processing, video compression, and artificial intelligence computation with low power consumption—is utilized in security cameras, vehicles, drones, and industrial equipment.
I evaluated the company based on its processing power per watt, image quality, software infrastructure, and the total system cost it offers to its customers. What particularly caught my interest in Ambarella is its design capability—the ability to adapt its technology across a diverse range of products—as well as its existing customer relationships.
My investment thesis is built on the expansion of artificial intelligence into the physical world, with Edge AI, the Internet of Things (IoT), and robotics at its core. I anticipate a shift toward an era where cameras, vehicles, factories, and robots process more data locally, gain a better perception of their environment, and perform an increasing number of tasks on-site. In this transformation, factors such as power consumption, heat management, latency, and connectivity requirements are critical; Ambarella’s products are designed specifically to address these needs.
The company’s primary revenue stream comes from the sale of processors. Once an Ambarella chip is integrated into a manufacturer’s design, it generates shipments and sales revenue as that product enters mass production. My growth outlook is predicated on the adoption of these chips in a larger number of customer products, the integration of higher-value AI chips, and the eventual translation of this volume into earnings per share.
A concrete basis for this today is the growing sales of the CV72 and CV75 series. For instance, i-PRO’s new cameras, which utilize the CV72, are capable of detecting specific scenarios—described through natural language—directly on the device itself. In the drone sector, the Antigravity A1 utilizes Ambarella’s CV5 chip. In robotics, there is Weilan’s CV72-based quadruped robot. These serve as examples of how the technology is being transformed into actual products. The company does not separately disclose the specific revenue generated from its robotics segment.
The CV7 and CV8 represent two key components of the new product cycle. The CV7 combines image processing with more powerful AI computing. For the 2-nanometer CV8—currently under development—management anticipates initial production revenue in fiscal year 2028 (the period spanning February 2027 to January 2028). The transition of these products to regular shipments is a cornerstone of my growth thesis.
The X7 opens up a distinct market segment for Ambarella; it can be integrated as an AI accelerator alongside Arm or x86 processors from other manufacturers. This allows the company to offer a product that brings its technology into systems that do not rely on Ambarella’s own main processors. The integration of AI into existing industrial systems and new device designs is the opportunity that interests me here. As of the September 2026 announcement, the X7 was in the sampling phase; the next steps toward commercial growth involve its integration into customer products and the commencement of mass shipments.
The agreement with Hanwha represents another tangible pillar of the company’s long-term opportunity. It encompasses supply and joint development across the fields of security, robotics, industrial automation, and life sciences. The projected potential revenue exceeds $800 million over a period spanning more than a decade. The realization of these sales depends on customer programs transitioning into production and firm orders; the breakdown between existing and new revenue streams has not been disclosed.
Furthermore, the company’s relationships with Macnica and Capgemini serve as channels that facilitate access to a broader customer base. In fragmented industrial and robotics markets, engineering, software, and installation support are required alongside the chips themselves. Management expects meaningful revenue from this new indirect sales channel within 2–3 years; in my thesis, this channel represents an area for longer-term growth.
Ambarella develops its own CVflow architecture and product platforms. For semi-custom projects like the CV8, customer requirements also help shape the design. In FY2026, R&D expenses reached $238.5 million—approximately 61% of revenue. The economic justification for a development infrastructure of this scale lies in new products reaching high sales volumes. For profitability to strengthen, both sales and gross profit must grow at a faster rate than development expenses.
I have evaluated the financials within this context. For the 12 months ending in July 2026, revenue stood at $417.8 million. After deducting investments in tangible/intangible assets and payments for licensed software from operating cash flow, approximately $2.3 million remains. During the same period, stock-based compensation expenses totaled $91.3 million; for shareholders, this manifests as dilution and an impact on value per share. The valuation basis for my position rests on the expectation of stronger growth, margin expansion, and cash generation in the coming years.
Customer concentration is a significant business risk. Insta360 is the largest disclosed end-customer, though its specific share of revenue is not publicly stated. The top 10 end-customers account for approximately 67% of revenue for the first six months. Meanwhile, the distribution channel—represented by "WT"—accounted for a 60.2% share in the second quarter. Reliance on Samsung for manufacturing, price competition, and the fact that customers are developing their own chips are also risks that I have taken into account.
The indicators that would validate the investment thesis are clear: the transition of new chips to mass production, an expanding customer base, recurring orders, and an increase in cash per share following capital investments. Conversely, factors such as persistent product delays, the loss of customers, or a failure to generate cash alongside growth would weaken the thesis. Ultimately, the key to the stock's performance will be whether the actual improvements meet the expectations already baked into the current price.
From a technical perspective, the stock has been trading within a long-standing horizontal range. I initiated a position when the share price was around $65, with plans to add to my holdings incrementally at the $56 and $44 levels. In the context of a "Darvas box" approach, a breakout above the upper bound would serve as confirmation for an entry; however, I took a position ahead of this breakout based on my assessment of the company’s fundamental potential. A sustained, high-volume breakout above the $97–$100 range would provide the technical justification for my expectation of a strong upward move. Currently, the price remains within this range.
I hold LEAPS (Long-Term Equity Anticipation Securities) contracts expiring in January 2029. For this position, the timing of when the market prices in the company’s growth is just as critical as the growth itself. The return on these contracts is shaped by the underlying share price, the option premium, the time remaining until expiration, and volatility. The relationship between the timeline for long-term product opportunities and the expiration date of the options is a defining characteristic of this position.
This is not investment advice.
