About Melten.ai
Melten turns customers’ product ideas into tested chip designs. Its AI software agents and chip engineers work together to design, test, and combine components. Specialist manufacturing partners produce the physical chips, an approach known as fabless manufacturing. Melten also plans a family of its own products based on technology developed through customer work. The agents share design knowledge across a project, while automated checks test work throughout development. The company starts with AI infrastructure, with later applications in robotics, industrial automation, and aerospace.
Melten is based in Campbell, California. It has completed full chip designs for data processing and AI computing, plus work on storage, connectivity, and secure processors. Chief executive officer Jonguk “Jeffrey” Song has more than 18 years in chip architecture and custom chip design. He held design leadership roles at Micron and SK hynix. Chief technology officer Yongjoon Lee has more than 16 years building production systems. At Zoox, he built validation systems for the company’s first public robotaxi launch.
Funding Rounds (SEC Form D)
| First sale | Filed with SEC | Amount sold | Investors | Avg. check | Terms | Filing |
|---|---|---|---|---|---|---|
| Oct 2, 2026 | Oct 19, 2026 | $2,285fully sold | 1 | $2,285 | Equity, Options, Securities on exerciseRule 506(b) | Form D ↗ |
| Apr 23, 2026 | May 4, 2026 | $2,199,999fully sold | 1 | $2.1M | EquityRule 506(c) | Form D ↗ |
| Total, two offerings | $2,202,284 | |||||
First sale is the date the first investor in the offering committed to invest. Companies must file Form D within 15 days of that date, so filing dates run later.
Why Investors Backed Melten.ai
Melten offers custom chip design work, starting with AI infrastructure. It targets both AI calculation and the movement of data inside servers. Its examples show design and verification before physical manufacturing.
The problem it targets
Melten argues that complex chip development takes too much time and money, keeping custom hardware out of reach for many applications. AI infrastructure needs processors that run models and separate chips that move data. Purpose-built AI processors are intended to address the cost and power demands of deployment.
What the product does
One complete design keeps data moving between a server’s network and storage. It combines networking, storage, and security so the server can devote more computing power to AI. It also includes acceleration and control. A second design runs AI model calculations. That chip combines computation, memory management, and security, with command control and supporting software. Other projects cover storage controllers, processor connections, and shared-memory links. The portfolio also includes secure processors.
Progress so far
On November 2, 2026, Melten’s agents completed a custom computer-interface controller in seven agent-days. A programmable test chip read encrypted storage at 28.4 gigabytes per second and booted Linux. In October, the system built a storage front end in seven days, reaching 13.9 gigabytes per second each way and 3.3 million small requests per second. Another project built a three-port switch in 17 days; two standard drives ran at 87% to 99% of rated speed. Separate work redesigned an AI processor in 52 hours and passed 11,226 exact output checks. An AI model accelerator moved from requirements to lab testing in 122 hours.
We wrote this summary from public statements by Melten.ai and its investors. It is not investment advice.
Leadership Team
Executive officers and directors listed on Melten.ai's Form D filed Oct 19, 2026.
Sources
- SECmelten.ai, Inc. filing history
- CompanyAbout Us — Melten.ai
- CompanyBlog — Melten.ai