EdgeCortix has raised more than $110 million. The Tokyo-based fabless semiconductor company announced the second close of an oversubscribed Series B in November 2025, bringing its total disclosed funding to that mark [EdgeCortix, November 2025]. That capital, combined with strategic credit and government grants, is funding a specific wager: that the most demanding AI inference will happen far from the cloud, in satellites, drones, and defense systems where power and latency are non-negotiable constraints.
A Software-First Wedge into Hardware
The company's pitch is a full stack, sold with a software-first methodology. Instead of designing a chip and then building tools for it, EdgeCortix uses what it calls "hardware and software co-exploration" to develop its runtime-reconfigurable processors [PERPLEXITY SONAR PRO BRIEF]. The goal is efficiency at the edge, a term that here means environments with severe physical limits.
Its product suite reflects this integrated approach. The Dynamic Neural Accelerator (DNA) is the underlying processor IP. The SAKURA-II is a physical accelerator card built on that IP, designed to handle generative AI models like Llama and Stable Diffusion within an 8-watt power envelope [Hakuto Co., Ltd.]. The MERA compiler and software framework completes the package, aiming to let developers compile and deploy models without touching a command line [ZEROTH SOURCE].
Why Japan, and Why Now
The company's Tokyo headquarters and founder profile are not incidental. Founder and CEO Dr. Sakyasingha Dasgupta brought two decades of AI research experience from institutions like RIKEN and IBM to the venture when he established it in 2019 [Sota Tsutsuki, Retrieved 2026]. The location taps into Japan's deep industrial and manufacturing heritage, a natural customer base for edge AI in robotics and automation.
More importantly, it positions EdgeCortix at the center of a national strategic push. In February 2025, the company was awarded a ¥3 billion (approximately $21 million) project by Japan's New Energy and Industrial Technology Development Organization (NEDO) to develop a low-power AI inference and learning chiplet platform [PERPLEXITY SONAR PRO BRIEF, February 2025]. A second ¥3 billion NEDO award followed in May 2025 [Chiplet Marketplace, 20250528]. This government backing is both non-dilutive capital and a powerful signal of domestic priority.
Traction in the Most Stringent Verticals
EdgeCortix is not selling to startups. Its announced customers and partners read like a roster of industrial and defense giants, suggesting its wedge is finding purchase where requirements are extreme.
- Defense & Aerospace. The company became the first Japanese semiconductor firm awarded a contract with the U.S. Defense Innovation Unit (DIU) in May 2025, benchmarking its hardware for potential deployment across Department of Defense applications [EdgeCortix, May 2025]. It later demonstrated its platform in an exercise with the U.S. Air Force [EdgeCortix, June 2026].
- Industrial Heavyweights. A multi-year, multimillion-dollar agreement with Kawasaki Heavy Industries, announced in 2026, aims to advance next-generation AI-enabled aerial defense systems [News Hub, 2026/09].
- Space. The company has a collaboration with lunar exploration company ispace for AI-enabled missions and has tested its earlier SAKURA-I accelerator for radiation resiliency in space environments [EdgeCortix, January].
- Strategic Semiconductors. Renesas Electronics is listed as both a long-term customer and a strategic investor, collaborating on AI-inference solutions [PERPLEXITY SONAR PRO BRIEF].
This customer progress is mirrored by a diverse and deep capital table. The Series B round attracted a mix of financial and strategic players from across the globe.
| Metric | Value |
|---|---|
| NEDO Project Grant (Feb 2025) | 21 M USD |
| NEDO Project Grant (May 2025) | 21 M USD |
| Mizuho Credit Facility (Nov 2025) | 10 M USD |
| Disclosed Equity (Total to Nov 2025) | 58 M USD (estimated) |
The Competitive and Commercial Pressure
The field of edge AI inference is crowded with well-funded specialists. Competitors like Mythic, Syntiant, and Kneron have also raised significant capital to chase similar promises of low-power AI. EdgeCortix's answer to this is its full-stack, software-first approach and its early traction in defense and aerospace, sectors with long sales cycles but formidable barriers to entry once a supplier is qualified.
The larger commercial risk is scaling from design wins and development agreements to volume production. Fabless semiconductor companies face immense capital expenditure and execution risk in moving from prototypes to mass-market chips. The company's use of project funding for its next-generation SAKURA-X chiplet platform is a deliberate hedge against this, using non-dilutive capital for expensive R&D [Technologies.org, 2026].
The $10 million unsecured credit facility from Mizuho Bank, extended in parallel to the Series B, is another telling data point. Banks typically lend against assets; an unsecured facility suggests the lender sees a viable, creditworthy business trajectory beyond the venture capital gamble [The AI Journal, 2026].
The Next Twelve Months
The roadmap is clear. Proceeds from the Series B and the NEDO grants are earmarked to accelerate R&D, scale global sales, and ramp mass production for the SAKURA-II accelerator and the in-development SAKURA-X chiplet platform [Technologies.org, 2026]. The multi-year agreement with Kawasaki Heavy Industries, with an initial scope spanning 2026 to 2028, provides a visible revenue pipeline and a flagship industrial deployment.
The key metric to watch will be the transition from development agreements to recurring product revenue. The bet from investors like SBI Investment, TDK Ventures, and Jane Street Global Trading is that EdgeCortix's software-integrated hardware can own the AI compute layer in environments where a cloud connection is impossible, a millisecond matters, or a watt is too expensive. With over $110 million in equity and grants now deployed, the question is whether the edge of the network is big enough to build a semiconductor powerhouse.
Sources
- [EdgeCortix, November 2025] EdgeCortix Closes Oversubscribed Series B, Bringing Total Funding Over $110M | https://www.edgecortix.com/en/press-releases/edgecortix-closes-oversubscribed-series-b-bringing-total-funding-over-110m-amid-growing-edge-ai-demand
- [PERPLEXITY SONAR PRO BRIEF] Company and product overview
- [Hakuto Co., Ltd.] Product specification reference
- [ZEROTH SOURCE] MERA Studio product description
- [Sota Tsutsuki, Retrieved 2026] Founder background via LinkedIn
- [PERPLEXITY SONAR PRO BRIEF, February 2025] NEDO project funding announcement
- [Chiplet Marketplace, 20250528] Second NEDO award announcement
- [EdgeCortix, May 2025] U.S. Defense Innovation Unit Contract Award | https://www.edgecortix.com/en/press-releases/edgecortix-becomes-first-japanese-and-semiconductor-company-awarded-u.s.-defense-innovation-unit-contract
- [EdgeCortix, June 2026] U.S. Air Force Demonstration | https://www.edgecortix.com/en/press-releases/edgecortix-demonstrates-edge-ai-platform-readiness-with-us-air-force-and-receives-success-memorandum-from-the-defense-innovation-unit
- [News Hub, 2026/09] Kawasaki Heavy Industries Agreement | https://www.edgecortix.com/en/press-releases/edgecortix-signs-multi-year-multimillion-dollar-agreement-with-kawasaki-heavy-industries
- [EdgeCortix, January] Space suitability testing
- [Technologies.org, 2026] Use of proceeds and business commentary
- [The AI Journal, 2026] Commentary on Mizuho Bank credit facility