Mythic, an AI chip startup that last November reportedly ran out of capital, rose from the ashes today with an unexpected injection of fresh funds.

Mythic this morning announced that it closed a $13 million financing round led by existing investors Atreides Management, DCVC and Lux Capital alongside latest investors Catapult Ventures and Hermann Hauser Investment. While the tranche is a fraction of the startup’s previous raise ($70 million), Mythic claims that the tranche will enable it to bring its “next-generation” product — an improved energy-efficient AI processor — to market.

“While economic conditions are difficult at once, this latest funding will help Mythic concentrate on its technology offering, go-to-market strategy and customer acquisition,” Dave Fick, Mythic’s newly appointed CEO, told TechCrunch in an email interview.

Co-founded by Fick and Mike Henry on the University of Michigan under the name Isocline, Mythic developed chip tech that stores analog values on flash transistors. While digital processors “pause” to swap data out and in of dedicated memory, Mythic’s hardware can perform calculations in parallel without stopping, resulting in performance and efficiency gains, particularly for AI applications — or so the corporate claims, at the very least.

Mythic initially worked on projects for the U.S. Air Force through its Small Business Innovation Research (SBIR) program, specializing in computer vision for high-altitude drones and GPS signal acquisition. Once the contracts wrapped up, the startup decided to pursue the enterprise capital route, raising over $170 million across multiple rounds from investors, including Hewlett Packard Enterprise and BlackRock.

With its first industrial chip, M1076, Mythic doubled down on computer vision use cases, constructing a system that may also help detect small objects from faraway distances in fewer than 33 milliseconds. It’s perhaps these capabilities that attracted Mythic’s largest customer up to now, Lockheed Martin, whose corporate arm, Lockheed Martin Ventures, became a significant investor within the startup.

So what went fallacious? Well, Mythic was competing in a really crowded field. Dozens of startups were — and are — developing chips to run AI efficiently at the sting, including SiMa.ai, Axelera, Flex Logix, NeuReality, EnCharge, Hailo and Kneron. But funding is drying up. According to The Register, global VC equity for semiconductor startups in 2022 declined 46% to $7.8 billion, reflecting increased scrutiny for the capital-intensive firms.

Mythic wasn’t the just one impacted by the bust. Last yr, AI chipmaker Graphcore, which reportedly had its valuation slashed by $1 billion after a cope with Microsoft fell through, said that it was planning job cuts attributable to the “extremely difficult” macroeconomic environment. Meanwhile, Habana Labs, the Intel-owned AI chip company, laid off an estimated 10% of its workforce.

To claw back to stability, Fick — who stepped up as Mythic’s CEO from CTO after Henry, who CEO, left to pursue other interests — prioritized efficiency. For instance, Mythic now leverages more development partners and off-the-shelf components than it did previously, cutting costs and (with luck) reducing the time to market.

Fick asserts that the restructuring has the additional benefit of creating Mythic’s R&D “more nimble” because it prepares to release its next-gen chip, the M2000.

“Given the brand new macroeconomic environment, we expect many startups will need to take an identical approach, especially for systems corporations which have to deliver many components beyond their core technology,” Fick added. “But a bonus of Mythic’s analog computing approach is that we will use more mature process nodes. These have more supply chain availability than bleeding-edge nodes and are also much more cost effective.”

Beyond cost cutting, Mythic revamped its go-to-market strategy, returning to its roots with a “renewed focus” on defense (and to a lesser extent public safety, industry and consumer verticals), in line with Fick. He wouldn’t elaborate. But presumably, meaning chasing after more customers with government contracts, like Lockheed.

“Cloud computing is usually not available on the battlefield, making a strict edge computing environment. Advanced computer vision is being applied in lots of places, including large and small drones, land-based autonomous systems, radar, augmented reality headsets and more,” Fick said. “Because the M2000 reduces the scale, weight, power, and price to deploy high-performance computer vision, these technologies may be applied in additional applications.”

Critics might deride Mythic’s pivot to defense as opportunistic, but there’s no denying that there’s a wellspring of capital there. As PitchBook noted in a recent report, after years of shunning investments in military and security-related tech startups, VCs have begun raising their profile within the sector because the U.S. attempts to achieve an upper hand against threats from adversaries, including China and Russia. In 2022, there was $7 billion invested in VC-backed U.S. aerospace and defense corporations through October 13, putting the sector on the right track to surpass 2021’s record deal value of $7.6 billion.

As Matt Ocko, managing partner at DCVC, put it in fairly apocalyptic language: “Mythic’s processor delivers data center GPU performance into the scale, power, and price of practical long-run-time edge systems that may defend America’s schools and public places from violence and terror, and our allied nations’ troops on the battlefield. Mythic’s technology will close the gap between what AI researchers publish and what systems makers actually deploy, and delivers revolutionary capabilities for homeland defense and national security.”

Parsing the hype, investors are smitten by Mythic’s latest defense-forward direction — at the very least for now. Time will tell whether it was the fitting one.