Valkryn raises $17 million for autonomous defense testing | Magica
Valkryn launches with $17 million for independent defense testing
Editorial Team
••📖6 min read
Valkryn disclosed a $15 million seed round and early military testing contracts. Its evaluator-focused Battle Lab faces existing commercial tools, public testing capabilities and funding constraints that software alone cannot resolve.
Valkryn disclosed $17 million in total funding, including a $15 million seed round led by Andreessen Horowitz.
The startup says its Battle Lab gives government evaluators comparable records of autonomous systems’ performance.
Existing field-testing tools and public-sector funding, data and staffing constraints narrow the opportunity.
Valkryn, a Washington, D.C.-based startup building infrastructure to evaluate military drones, robots and other autonomous systems, emerged from stealth on November 17 with $17 million in total funding. Its pitch is to serve the government organizations deciding what to field, drawing a commercial boundary between evaluating weapons and selling them.
The company’s launch announcement disclosed a $15 million seed round led by venture investor Andreessen Horowitz’s American Dynamism fund, with Cantos Ventures, Ambition Capital and Caffeinated Capital participating. An earlier $2 million pre-seed accounts for the balance, a launch report said. The $17 million is cumulative financing, rather than the size of the new round.
The timing matters because software updates can change a system’s behavior between physical test events. Valkryn wants to make those evaluations repeatable and their results easier to compare. But neither continuous testing nor independent evaluation starts with this company: commercial suppliers already offer field tools, and government organizations have their own modernization efforts. Its opportunity depends on integrating those functions for evaluators and demonstrating that the results improve decisions.
Valkryn’s company-reported total funding comprises a $15 million seed round and an earlier $2 million pre-seed. Source: Tectonic Defense.
Founders who built autonomy and its testing tools
Valkryn co-founder and CEO Hugh Zabriskie previously led development of autonomy testing tools at Applied Intuition, a supplier of autonomy software and infrastructure. Co-founder and chief technology officer Andrew Hoff previously led mission autonomy engineering at defense technology company Anduril, according to .
Microsoft is pitching Copilot Workspaces as a shared environment for people and AI agents. Team collaboration is already in early access, but personal Autopilot integration is still coming soon, with final agent pricing unannounced.
That experience connects the founders to both the systems being assessed and the tools used to assess them. Zabriskie says Valkryn’s distinction is its customer: the evaluator responsible for deciding how to field equipment, rather than the developer making it.
Founded in 2025, Valkryn says it secured multiple funded contracts with Pentagon testing organizations in less than 12 months. Zabriskie described the contracts in the product reporting as operational work, mostly through other transaction agreements, rather than Small Business Innovation Research awards. The launch materials disclose neither contract values nor pricing.
Valkryn also says it deployed with the U.S. military in the Middle East within three weeks of its first technology demonstration. The product reporting describes counter-drone evaluation in the field; the separate launch report describes work with a task force but says further details are scarce. Neither account supplies test results sufficient to measure savings or evaluation accuracy.
What the Battle Lab measures—and what already exists
Valkryn says its Battle Lab combines physical test environments, instrumentation and analysis software. The environments are designed to expose systems to conditions such as radio-frequency denial, dense counter-drone threats, adversarial systems and weather variations. Software interfaces and telemetry capture how systems and operators behave, producing a timeline that evaluators can inspect in a web application.
The company says its software creates auditable records and automatically scores results against government-defined criteria. It argues that consistent measurement allows comparisons across programs, sites and exercises. Those are claims about the platform’s capabilities, rather than independently published evidence that different tests produce comparable answers.
The analysis layer is intended to reveal patterns between software versions, tactics and mission outcomes. Zabriskie attributes gaps between claimed and actual performance to inadequate measurement and learning, rather than bad intent by suppliers or government. That is his explanation of the problem the company seeks to solve, not a finding about any particular weapons program.
Commercial alternatives cover substantial parts of the workflow. The product reporting identifies test-data company Nominal’s internal collection and analysis tools and Applied Intuition’s simulation capabilities.
Applied Intuition also operates beyond simulation. Founded in 2017, the company supplies tools, infrastructure, operating systems and autonomy across defense and civilian industries. Its March 31 launch of Applied Edge described a mobile platform for developing, testing and operating autonomous systems. The company said Edge packages power, rugged computing and communications with software for replaying logs, evaluating behavior and streaming live telemetry. It said the platform was available then, with data and workflows persisting between test events.
Persistent field-testing infrastructure was therefore already being offered months before Valkryn’s launch. Valkryn’s narrower case is evaluator-focused integration and its decision not to build or sell the systems under test. Its own announcement acknowledges federally funded research centers and university labs as established independent evaluators, while arguing that those institutions are not designed to scale commercial technology.
The Government Accountability Office, Congress’s audit and investigative arm, has documented barriers that support the need for better testing while limiting what a commercial platform can fix.
A December 2025 audit of testing modernization found Pentagon-wide policies did not fully reflect practices including early tester involvement, iterative testing, digital engineering and ongoing user feedback. The shortfalls were especially pronounced in acquisition pathways accounting for most weapons-acquisition spending; software and urgent-capability pathways already reflected some of those practices.
The audit also documented existing public capabilities. The Army Test and Evaluation Command was digitally modeling vehicle ranges and had conducted a test combining live and virtual assets across the country. The issue is therefore broader adoption and integration, rather than a complete absence of modern testing.
Money and access also constrain adoption. Air Force and Navy officials told GAO that testers depend on acquisition programs to fund tools, leaving some unable to invest in digital infrastructure. Air Force developmental test officials said that, in some instances, obtaining contractor data and models required an extra year of negotiations. These are funding and contracting obstacles that better analysis software alone cannot remove.
The policy changes themselves were contested. The Navy pointed to existing coordination requirements when rejecting GAO’s recommendation for end-user agreements; GAO said those requirements did not provide the same process for ongoing feedback. The audit examined policies and a nongeneralizable sample of eight weapons programs, not the performance of Valkryn or autonomous systems as a category.
A separate June 2026 audit found that a reorganization streamlined some processes but reduced capacity at the Pentagon’s Office of the Director, Operational Test and Evaluation, which oversees operational and live-fire testing. The office’s 2025 oversight list covered 173 weapons programs—a broader portfolio than autonomous systems. Staff reported excessive workloads and gaps in subject expertise; a support-contract work stoppage also disrupted access to testing data for about two months.
These findings separate demand for better tools from the government’s ability to buy, use and act on them.
What the next contracts need to demonstrate
Zabriskie said in the launch coverage that Valkryn primarily operates using existing government land and airspace, which he considers underused. Access to public facilities is thus part of the operating model, even with venture funding for the company’s own technology and team.
Valkryn plans to spend the funding on technology and expansion across Washington, New York and Los Angeles. The next evidence needed is whether evaluators can obtain repeatable comparisons across systems and software versions, how much each evaluation costs, and whether the records shorten decisions without weakening scrutiny. Early funded work provides a starting point; the disclosed financing and deployments do not yet answer those questions.
Google is rolling out file references and sharing for Gemini skills as adult personal accounts lose the ability to use Gems. Automatic conversion preserves supported material, but file limits and missing tools can interrupt existing workflows.
Muddy Waters disclosed a short position in Nvidia while questioning a possible connection involving senior adviser Ajay K. Puri. Its own account leaves the key identity question unanswered, separating its new claims from an existing server-diversion prosecution.
Nvidia reported $115.8 billion in quarterly revenue in its first earnings release since co-founder Jensen Huang resigned. Strong demand gives interim CEO Colette Kress a financial cushion, while a campaign for Huang’s return collides with unresolved investigations and the board’s reported opposition.
OpenAI launched GPT-6.2 Astra and Sol for professional work, cutting Astra’s standard input and output token prices by 20%. The models offer cheaper paths through selected tasks, but research evaluations, reporting failures and unfinished independent safeguard testing complicate the case for broader autonomy.
Microsoft’s first live speech-to-speech model offers OpenAI API compatibility and planned enterprise integrations. Its published evaluations show a low-cost option, with several rivals completing more agentic tasks.
Canberra-based Trellis Data is offering its 220-billion-parameter Agentaus 1 model to business and government. Its pitch goes beyond keeping data in Australia, but performance, deployment costs and a planned capital raising remain key tests.
Perplexity is adding direct collaboration between frontier models to Computer, letting a lead agent coordinate up to 16 supporting agents. Company-reported benchmarks suggest mixed teams can outperform repeated copies of one model, but leave customer costs and production reliability unresolved.
Marsh reported 105 client cyber incidents in the last two weeks of October, against 48 in the final two weeks of May. The broker links the increase partly to AI, while warning that insurance panel access may not guarantee an available response team.
Meta’s Muse personal AI agent has launched in Australia, with some local shopping and booking connections to follow. Its browser and payment tools broaden its reach, while the company’s technical disclosures qualify its privacy promises.
Microsoft and Salesforce are extending their sales software alliance with Copilotforce, combining CRM workflows with Microsoft 365 context under existing permissions. The Copilot plugin enters public preview, while Teams remains in pilot; earlier Microsoft tools and a Claude beta raise the bar for proving its value.
Microsoft expands MAI Image 2.7 from an editing preview into image generation and API access, with cheaper Flash pricing and a Canvas workspace. Benchmark claims depend on the workload, while the MAI Labs interface carries restrictions on commercial use.
OpenAI is widening applications for its legal AI configuration after improving research scores and expanding its search index. Access still requires approval, and benchmark results show substantial limits alongside the gains.
Microsoft has made Word, Excel and PowerPoint creation generally available inside Copilot, combining editable files with workplace permissions and collaboration. The broader shift toward autonomous work remains staggered across previews, while its claimed cost advantage rests on company-funded testing.
Knit Health’s NEA-led Series A funds deployment of Med-BEAM for care navigation and hospital operations. Its early company benchmark beats three Claude models at predicting clinical actions, while better patient outcomes remain unproven by that test.
Science Corporation’s Series D follows its European retinal-implant launch. Clinical results support improved central vision, but U.S. approval, eligible patients and commercial terms remain consequential questions for the startup, reportedly valued at $3.5 billion.
Google will let Gemini users save conversations while opting out of generative AI training on future chats. The global rollout also extends storage of history-off and temporary chats from 72 hours to around two months; human review and longer retention can still apply.
Google is removing the subscription requirement for Gemini’s task mode in six countries, while adding U.S. payment, calling and messaging features. The expansion broadens access to delegated work, but usage limits, permissions and website compatibility still govern what it can finish.
Eligible Pixel owners can join a Play Store beta to generate home-screen widgets from descriptions, but device, language and regional restrictions still apply.
Microsoft’s new in-house coding models compete with outside suppliers inside GitHub Copilot. Company tests show lower task costs at promotional prices, while its continuing financial ties to OpenAI complicate the rivalry.