To The Moon

$Ambarella(AMBA)$  If you're looking for a company in the robotics sector that still offers reasonable valuation, hasn't been wildly over-hyped, and is perfectly positioned in the long-term robotics megatrend, then you shouldn't miss Ambarella.This is AI's next breakout area. It’s not a short-term concept play, but the very beginning of a multi-year secular industry cycle.The demand potential is enormous. Goldman Sachs estimates that by 2035, the addressable market for humanoid robots could reach $38 billion — more than 6x their previous forecast of $6 billion. Morgan Stanley is even more bullish, projecting up to 1 billion units by 2050.Even with conservative assumptions, the demand for components is already staggering. For every million robots, you need tens of actuators per unit, plus large numbers of sensors, controllers, precision reducers, and vision systems.NVIDIA has repeatedly emphasized Physical AI, calling it a $50 trillion opportunity. In the next 3–5 years, robots will be everywhere. Capital, industry, and tech giants are all pushing AI from the digital world into the physical one.In this environment, the real winners are often the "picks and shovels" companies deep in the supply chain.Many robots use dedicated, low-power vision processing chips for tasks like object detection, depth mapping, and scene understanding. The company I want to talk about today is one of the best in the world at providing specialized AI vision processing solutions for robotics.

Ambarella – The Robot’s Eyes and Local Brain

If robots are the body of AI, Ambarella (AMBA) wants to be the eyes and the local brain right next to them.Their low-power vision AI chips (notably the M8 series and newer generations) are widely used in cameras, drones, robots, automobiles, security systems, and industrial equipment. These chips process visual data on-device, run AI inference locally, and output actionable intelligence.Think of it this way: a security camera records video → Ambarella’s chip processes the images with AI → and immediately triggers alerts (e.g., someone entering a restricted area, a vehicle parking illegally, a worker in a danger zone, a drone identifying an object, or a robot detecting an obstacle).I’ve liked this company since around 2020, when it was still relatively under the radar. Back then, it was best known as GoPro’s main chip supplier. High-quality, low-power video processing for action cameras and security systems was its original foundation.That GoPro-era legacy — being one of the best video processing chip companies in the world — is now one of its strongest moats for AI vision.

Key Competitive Moats

1. Superior Image Signal Processing (ISP) + AI Integration

AI vision starts with seeing clearly. Ambarella has over a decade of deep expertise in low-light performance, HDR, wide-temperature operation, noise reduction, and image stabilization. They integrate top-tier ISP with their own AI inference engine (CV Flow architecture) on a single chip.

Competitors often need a separate image processor plus an AI chip. Ambarella’s integrated solution cuts cost and dramatically reduces latency.2. Excellent Price-Performance

Not every robot needs a $2,000 NVIDIA chip. Mass-market robots — lawn mowers, floor cleaners, industrial arms — need capable vision and local control chips priced between $15–$100. Ambarella dominates this huge middle market.

Its next-gen CV7 chip aims to integrate vision, AI decision-making, and even robotic arm control on a single SoC, priced roughly $10–$75.3. High Customer Stickiness & Software Ecosystem

In the B2B chip market, switching costs are extremely high. Ambarella’s COOPER development platform provides on-device robotics control software, compilers, and SDKs. Once a customer builds their system around Ambarella’s tools, migrating to Qualcomm, AMD, or others takes years and massive R&D investment.

They also signed a 10+ year, $800 million+ long-term agreement with South Korea’s Hanwha covering security, industrial, and robotics — a major validation with very sticky, long-cycle revenue.4. On-Device AI = Low Latency + Privacy

Processing vision locally eliminates cloud upload delays (critical for robot obstacle avoidance, drone flight, automotive safety) and addresses strict privacy regulations in Europe and the US, especially for security and in-cabin cameras.

Financial Snapshot

In fiscal 2026, 78% of revenue came from IoT (including security, industrial, and robotics) and 22% from automotive. Roughly 80% of total revenue is already from edge AI applications.

Revenue grew 37% for the year, with 46 million AI chips shipped. CV72 and CV75 chips are already in security and industrial robots; the CV7 targets full robotic control.Gross margin is a healthy ~60%, typical of a company selling differentiated, technology-rich vision AI chips rather than commoditized parts.

Current market cap is around $3.5 billion — still relatively small, which gives it significant upside leverage if robotics, edge AI, and automotive vision take off.

Recent Earnings & Stock Reaction

The stock sold off after the latest report not because results were bad, but because the market wanted more. Q2 guidance came in slightly below the highest expectations after the stock had already surged ~44% in the prior 30 days. Management’s fiscal 2027 outlook implies revenue growth of 10–15% (roughly $4.3–4.5 billion range), with gross margins staying at 59–62%. It’s not explosive growth yet, but it reflects an early-stage transition waiting for new products and applications to accelerate.

Three Major Growth Drivers

Robotics, Drones & Warehouse Automation

This is the clearest path to “Physical AI.” Ambarella is already seeing real deployments (e.g., N1-655 chips in large warehouse automation systems). Drones contributed their first full-quarter revenue recently. Success here would reposition the company from a “security camera chip maker” to a true Physical AI platform.

Automotive

Currently focused on safety and ADAS cameras, but expanding into higher-level perception with CV-AD series, radar, and acquisitions (Vizelab for software, Oculet for radar). They have ~$13 billion in won or bid automotive projects through 2032 — very long-cycle, sticky business.

Edge Infrastructure

Local AI inference units, edge gateways, and smart sensors. This could evolve Ambarella into a broader local AI infrastructure provider.

Investment Perspective

At around $78, the stock is in an interesting zone — pulled back from its 52-week high of ~$96.69 but still above major support.

$78–80: Observation / light accumulation zone.

$72–75: More attractive entry for longer-term believers.

Above $88–90: Bullish short-term signal; breakout above $96–100 would be very strong.

Bottom line: If you believe robotics is moving from hype to real volume production, don’t just chase the companies that tell the best stories. Look at the critical underlying components that every robot will need. Ambarella is one of those names worth following closely.