Nicholas Mugalli
Institutional TMT & Macro Research, AI/Semiconductors/Policy
One Stop Systems' 52% gross margin, against a 41% estimate, shows it sells mission-critical compute customers can't source elsewhere, not boxes. The quarter beat on revenue, EPS and margin and guided above consensus. That margin on ruggedized edge hardware reflects software and IP mix. Defense autonomy and battlefield edge AI inference run on multiyear, highly classified procurement contracts rather than cyclical tech markets, a supply chain separate from hyperscaler data centers. x.comÂ
One Stop Systems is a small, high-risk way to play rugged edge-AI computing for the military. Revenue grew 62% to $9.3M, bookings hit a record $15.1M for a 1.7x book-to-bill, and growth guidance was raised to 25-30%. A new DoD intelligence platform is expected to bring about $44M over four years, and a new Navy relationship began in August. With a $234M market cap and shares 55% below their high, it is the highest-risk name on the list. x.comÂ
One Stop Systems is a pure-play edge AI defense company, and edge AI in defense and heavy industry is still early. Standard servers fail under field conditions, altitude and vibration, yet fighter jets, armored vehicles, missile defense and field medical units need on-site inference; OSS builds rugged, liquid-cooled platforms around accelerators such as NVIDIA's. Divesting low-margin units like Bressner lifted Q1 gross margin to 51.6%, and a 1.8x book-to-bill supports predictable earnings. x.comÂ
One Stop Systems looks inexpensive relative to its growth, though it remains a risky play. FY26 growth guidance rose by 10-15% to about 40%, roughly four times the growth rate of rugged defense electronics peers such as Mercury Systems and Leonardo DRS. OSS trades at about 2.8x FY27 forecasts, versus Mercury's roughly 4.6x next-twelve-month sales. x.comÂ
One Stop Systems is a profitable AI compute small-cap whose Russell 2000 inclusion in 48 hours forces index buying that nobody is pricing in yet. It supplies rugged edge AI for drones, ships, autonomous vehicles and robotics, and holds an $8.4M defense contract with $44M of lifetime potential. x.comÂ
One Stop Systems is a compact way to gain upside from high-performance compute infrastructure. x.comÂ
Synaptics stands to benefit as AI moves out of data centers into physical devices, a read-through from Micron's results. Micron said L2+ vehicles carry 5x the memory and storage of average cars and humanoid robots carry 10x that of L2+ vehicles. Physical AI needs local compute, sensing, connectivity and low-power inference, which is Synaptics' lane. x.comÂ
AI on the edge is barely factored into Nvidia's and AMD's valuations, yet it is likely to become a multi-trillion-dollar market. As Larry Ellison explains, real-time models for autonomous driving, robotics and industrial automation need ultra-low-latency compute inside the device, using chips very different from data center GPUs. Investors see the data center buildout but are still missing edge AI. x.comÂ
Ambarella has become a pure-play edge AI chip company as AI inference migrates from the cloud to devices. Edge AI silicon is now 80% of the business, and revenue rose 37.2% to $390.7 million last fiscal year. Latency, bandwidth costs and privacy push inference onto drones, cars and robots, which its CV3-AD chips serve locally. A Hanwha agreement could bring up to $800 million, and non-GAAP EPS of $0.11 marks a profitability inflection, though GAAP results still show a loss. x.comÂ
SiTime is an under-the-radar AI infrastructure winner posting triple-digit growth while likely trading below 40x NTM EPS. Its silicon timing chips replace quartz across data centers, robots, drones, defense and robotaxis. Q3 guidance implies ~250% YoY revenue and ~315% EPS growth; the Renesas deal adds ~$85M, leaving ~30% organic QoQ growth. It sells more into each rack, expects 1.6T optical revenue to double in 2027 with higher share and ASPs, and customers order 12-18 months out. x.com x.com … Show more
