Circuit Mind announced $25 million in financing on October 28, 2026, alongside the launch of AVE, its Agentic Verification Engine. The funding package includes an $18.5 million Series A led by Norwest and a previously unannounced seed round, bringing the company’s total funding to $33.4 million, according to its announcement.
AVE extends Circuit Mind’s platform from generating circuit designs into reviewing them. The company says its AI agents check electronic schematics against component datasheets and an organization’s design rules, aiming to catch problems before engineers manufacture a prototype.
That puts Circuit Mind in a consequential corner of the AI industry: automating engineering work where an overlooked constraint can require another physical build. Its approach combines agents with structured component models and a physics-based decision engine. Those architectural details deserve attention separately from its advertised speed and cost savings, which measure different tasks and come with different levels of supporting evidence.
What the $25 Million Financing Includes
Norwest led the $18.5 million Series A, with participation from existing investors Third Prime, Differential and Blackhorn, plus new investor Lightscape VC. Third Prime led the previously unannounced seed round included in the $25 million announcement. Norwest general partner Dave Zilberman will join Circuit Mind’s board.
The distinction matters: $25 million is the combined financing being announced, not the size of the Series A alone.
Circuit Mind plans to use the money to expand its platform across the electronics design lifecycle, hire across engineering, sales and customer success, and establish a U.S. headquarters for the North American market. SiliconANGLE’s coverage identifies testing and firmware as areas of intended expansion.
Co-founder and CEO Tomide Adesanmi positioned the company alongside software-development assistants such as Cursor, Claude Code and Cognition, describing the goal as bringing hardware teams “AI workflows that professional engineers can rely on.”
The comparison explains the workflow ambition. The harder technical question is how an agent’s suggestions become defensible electrical decisions.
How AVE Uses Component Data and Engineering Rules
Circuit Mind describes three foundations for its platform: agentic AI, machine-readable models of electronic components, and a patented physics-based decision engine. Customer-specific engineering context sits alongside them, including approved manufacturers, preferred parts and design rules.
Norwest’s investment explanation provides more detail about the component models. They capture electrical behavior, connectivity rules, operating limits and other characteristics that determine whether a part is suitable for a circuit.
That is a more constrained task than asking a general-purpose chatbot to read a datasheet and suggest a design. Structured data gives the system explicit properties and limits to reason over. Company rules add requirements that may never appear in a manufacturer’s datasheet, such as preferred components or internal review standards.
The platform has two applications:
| Application | Main job | Described inputs and outputs |
|---|---|---|
| ACE: Assistant to Circuit Engineers | Generate and compare circuit designs | Takes a hardware block diagram and requirements, then produces schematics, a bill of materials and supporting analysis |
| AVE: Agentic Verification Engine | Review an electronic design | Checks the design against component datasheets and organizational rules to surface potential issues |
Circuit Mind says ACE can optimize design options for cost, size, power or supply-chain availability. AVE addresses a different task: checking whether a design respects the requirements of its components and the organization building it.
According to SiliconANGLE, AVE’s agents trace circuit connections and buses, check them against pin tables and operating voltages, and critique their findings before presenting them to an engineer.
That last step is a useful workflow detail. A checking system that produces many weak findings can create another review backlog. Having agents critique findings is intended to reduce that burden, although the process alone says nothing about how many errors survive or how many valid connections get flagged.
Where AVE Fits Alongside Existing ECAD Checks
Circuit Mind says its platform integrates with existing electronic computer-aided design, or ECAD, tools, libraries and product lifecycle management systems. Its website also describes support for customers’ symbol and footprint libraries, reusable design blocks, naming standards and review checklists.
The intended role is an additional review layer inside an established engineering workflow.
Conventional checks differ by tool and design stage. A schematic electrical-rule check is not the same thing as a board-layout check for physical spacing. It would be misleading to describe all existing design-rule checking as merely geometric.
AVE’s proposed distinction is its use of component-specific documentation and organizational context. The checks described in the available coverage include:
- Pin and connection compatibility: comparing circuit connections with component pin information.
- Operating-voltage compliance: looking for mismatches between connected circuitry and component limits.
- Company-specific requirements: applying internal engineering rules and review checklists.
For illustration, a connection may be present in a schematic while still being unsuitable for a particular component’s operating conditions. A review system needs the relevant component limits and design context to identify that problem; connectivity alone is insufficient.
My view is that this is AVE’s strongest practical pitch. Reviewing an existing schematic gives teams a bounded task to evaluate, without requiring them to adopt automated design generation at the same time. That follows from Circuit Mind’s stated support for bringing existing designs into its platform for verification, described on its website.
Integration still matters. A useful finding must be understandable in the context of the engineer’s actual design, libraries and internal standards. Those inputs are part of the verification task, not incidental setup.
What the Performance Evidence Actually Measures
Circuit Mind’s announcement includes substantial performance claims, but they should not be treated as interchangeable benchmarks.
The most concrete external evaluation described in the opened sources concerns design generation, not AVE’s verification accuracy. Norwest says Los Alamos National Laboratory completed roughly 90% of a production-representative PCB schematic design in just over four hours, compared with an estimated 60 to 80 hours of manual engineering work. SiliconANGLE reports the same comparison and says the lab cleared the report for public release.
Circuit Mind’s announcement summarizes the case as four hours versus two weeks. Norwest’s description adds two important qualifications: the result covered roughly 90% of the schematic, and the manual comparison was an estimate.
| Published result | Attribution and conditions | What readers can take from it |
|---|---|---|
| Roughly 90% of a representative schematic completed in just over four hours, versus an estimated 60–80 manual hours | Los Alamos evaluation, as described by Norwest and SiliconANGLE | Evidence of faster generation on one reported design task, not an AVE accuracy benchmark |
| Potential reduction of more than 90% in architecture-exploration effort |
Frequently Asked Questions
3 questions
1How much did Circuit Mind raise?
Circuit Mind announced $25 million across an $18.5 million Series A and a previously unannounced seed round. It reported total funding of $33.4 million.
