Neura's $1.4B Round Has 4 Cities to Poach. None of Them Is San Francisco.
Neura Robotics closed a $1.4B Series C on June 10. The humanoid talent pool it now needs sits in four European cities, not on LinkedIn.
On June 10, 2026, Neura Robotics in Metzingen closed a Series C of up to $1.4 billion led by Tether, with Qualcomm, Amazon, NVIDIA, Bosch, Schaeffler, the European Investment Bank, and Lingotto Horizon on the cap table. Bloomberg pegged the valuation at roughly $7 billion. If you recruit for humanoid robotics in Europe, your week just got harder: the talent Neura needs to deploy that capital lives in four cities, knows each other, and is already employed.
This is the largest venture round in German history and the largest ever raised by a full-stack robotics company. It will not pull European engineers to the Bay Area. It will pull them across the Isar.
The pool is a triangle, not a continent
Headlines call this a "European" raise. That framing is wrong for sourcing. The operational center of humanoid gravity is a tight quadrilateral: Metzingen, Munich, Stuttgart, and Zurich, with Paris (Wandercraft, Enchanted Tools), Barcelona (PAL Robotics), and the UK (Humanoid, Engineered Arts) as peripheral nodes. Inside that quadrilateral, the same lab pedigrees keep appearing: DLR (German Aerospace Center), ETH Zurich (especially Marco Hutter's RSL), TUM, KIT, DFKI, and Fraunhofer IPA.
If you run a country-wide LinkedIn boolean for "humanoid" + "Germany," you will get hundreds of results and miss the people Neura actually wants. The signal is lab pedigree, not job title. Agile Robots SE was spun out of DLR in 2018 by Peter Meusel and Zhaopeng Chen. ANYbotics came out of ETH Zurich's RSL in 2009. Neura's own engineering bench is heavily DLR and Fraunhofer IPA flavored. Source by where they trained, not by where they live.
Why US-centric tooling fails here
Most sourcing stacks were built for the Bay Area and Seattle. They overweight LinkedIn activity, English-language signals, and US employer normalization. They underweight three things that matter in this triangle:
- Lab affiliations buried in PhD theses and IROS/ICRA author lists.
- GitHub commits to ROS 2 packages, MoveIt, Drake, Pinocchio, and proprietary middleware forks.
- German and Swiss labor-market norms (notice periods of three to six months, strong non-poach culture between Bosch and Schaeffler suppliers).
A recruiter who pastes a Bay Area humanoid req into LinkedIn Recruiter will get back a shortlist of generalist software engineers with "robotics" in their about-section. A recruiter who searches by lab and contribution graph will find the 600 people who actually matter.
Who Neura will poach, and who will poach Neura
The cap table is also the org chart of your competition. Bosch and Neura announced a humanoid software collaboration in January 2026. In April, Neura partnered with Dassault Systèmes on sim-to-real. Bosch Corporate Research and Schaeffler advanced engineering will now operate as talent corridors into Neura. That is good news if you are sourcing into Neura's adjacents and bad news if you are trying to hire away from them.
The real near-term poaching risk to Neura is not 1X in Palo Alto or Figure in Sunnyvale. It is Agile Robots, twenty minutes away in Munich, which launched its first bipedal humanoid Agile ONE in late 2025 and recruits from the same DLR + TUM pipeline. Two DLR-pedigreed humanoid programs are now fighting over the same alumni network.
The other names that should be on every humanoid recruiter's whiteboard:
- Agile Robots SE (Munich): direct competitor, same lab pedigree.
- ANYbotics (Zurich): quadruped, but the manipulation-ML and locomotion-controls people are exactly the profile Neura wants for 4NE1.
- Honda Research Institute Europe (Offenbach): a quiet, deep bench of manipulation researchers.
- PickNik Robotics (remote-friendly, MoveIt maintainers): the ROS 2 motion-planning experts.
- Wandercraft (Paris): bipedal walking expertise, freshly funded by Bpifrance and Renault.
- PAL Robotics (Barcelona): mature humanoid platform, long-tenured engineers ripe for relocation.
- Mercedes-Benz (Stuttgart): already piloting Apptronik Apollo, building internal humanoid integration knowledge.
The cap table is also the org chart of your competition.
The three tiers of "robotics engineer" Neura needs
The biggest mistake in humanoid robotics recruiting is treating "Robotics Software Engineer" as one role. Internal index data on senior robotics engineers with ROS skills across Germany and Switzerland shows the title is by far the most common label. The rare profiles inside it are not.
Tier 1: Manipulation-ML and whole-body controls
Whole-body model predictive control. Contact-rich manipulation. Vision-language-action model training on actual humanoid hardware. These people number in the dozens across the whole triangle, not the hundreds. They publish at CoRL, RSS, and Humanoids. Their GitHub shows Drake, Pinocchio, MuJoCo, and Isaac Lab commits, not just rospy tutorials. Most are still in PhD programs or postdocs at ETH, TUM, KIT, or DLR. Neura's $1.4B will be spent disproportionately to win this tier.
Tier 2: Real-time controls and perception
EtherCAT, real-time Linux (PREEMPT_RT or Xenomai), sensor fusion across IMUs, force-torque sensors, and depth cameras. These engineers are at Bosch CR, Schaeffler, Festo, KUKA, and the Fraunhofer institutes. They are the bridge between the research tier and the integration tier. Headcount in the low hundreds across the triangle.
Tier 3: ROS 2 integration and platform
Behavior trees, MoveIt, Nav2, simulation pipelines, CI for embedded targets. This is the tier most "ROS engineer sourcing" searches actually return. Healthy supply, especially out of KIT, TU Darmstadt, and the Fraunhofer IPA orbit in Stuttgart. Hireable in weeks if you know where to look.
If your req says "ROS engineer" and means tier 1, you will spend six months interviewing the wrong people. If it says "ROS engineer" and means tier 3, you should already have ten in your pipeline.
This is the friction we built Refolk to remove. You describe the engineer in plain English, including the pedigree and the rare skill, and get a ranked shortlist across GitHub, LinkedIn, paper authorship, and the open web. For humanoid sourcing in particular, the cross-source join (GitHub commits to a specific motion-planning library + thesis advisor at a specific lab + current employer) is what separates a usable shortlist from noise.
Stop searching for "ROS." Start searching for stacks.
Top European humanoid teams have largely moved to ROS 2 plus custom middleware, or to proprietary stacks entirely. Neura's Neuraverse is one example. Agile Robots runs a heavily customized stack. ANYbotics has its own on-robot software. A recruiter who searches "ROS" alone will surface academic CVs and miss the people who actually ship.
Better signals:
- GitHub: commits to
ros2,moveit2,nav2,drake,pinocchio,crocoddyl,mujoco,isaac_ros_*,cartographer. Bonus if the commits are in C++ and touch real-time code paths. - Paper authorship: Humanoids, ICRA, IROS, CoRL, RSS in the last three years, with at least one paper that includes hardware results, not just simulation.
- Lab affiliation in the email or thesis metadata:
@dlr.de,@ethz.ch,@tum.de,@kit.edu,@ipa.fraunhofer.de,@dfki.de. - Conference attendance: International Humanoid Forum in Biel, RoboCup German Open, Munich Robotics Network meetups.
Refolk lets you stack those signals in a single plain-English query, which matters because the engineer you actually want is rarely findable from any one source alone. The same person might be invisible on LinkedIn, prolific on GitHub under a handle, and the third author on a 2024 Humanoids paper out of TUM. You need all three views to confirm it is one person.
What this means for your Q3 plan
If you are hiring against Neura, three things change this quarter.
1. Comp benchmarks will move, but not as fast as you think
Germany's average labor cost was €43.40 per hour in 2024, the seventh highest in the EU. That ceiling does not move because Tether wrote a check. Neura's $1.4B will buy hiring velocity and equity upside, not radically higher base salaries. The competitive lever for adjacents is signing bonuses, relocation packages from Zurich or Paris to Munich, and accelerated equity vesting. Lead with those, not base.
2. Notice periods just became your enemy
Three to six month German notice periods mean a candidate Neura signs in June starts in October or later. If you want to outpace Neura on a specific target, the move is to identify them now, build the relationship across two or three conversations, and have the offer ready before Neura's recruiter calls. This is exactly the scenario where a fast sourcing loop matters. We built Refolk so that the time from "I need three people like this" to "here are 40 ranked candidates with context" is minutes, not a week.
3. The corporate strategics are now your channel
Bosch, Schaeffler, NVIDIA, Amazon, and Qualcomm all signed onto this round. Their internal humanoid and robotics teams now have a roadmap reason to talk to Neura, which means their engineers will too. That is a two-way door. If you are not Neura, the engineers in Bosch CR's robotics groups, NVIDIA's Isaac team in Munich, and Amazon's robotics group in Berlin are now actively comparable to Neura on mission. Source there.
The 600-person reality
Across Metzingen, Munich, Stuttgart, and Zurich, the realistic count of engineers who can ship on a humanoid program at Neura's tier is in the low hundreds. Add Paris, Barcelona, Offenbach, and a handful of UK profiles and you are still well under a thousand. Neura's $1.4B is not chasing a labor market. It is chasing a roster.
Treat it that way. Name the labs. Name the advisors. Name the GitHub handles. The companies that win the next eighteen months of humanoid robotics recruiting will be the ones whose sourcing looks more like scouting a football academy than running a job board.
FAQ
Where should I focus sourcing first if I only have a week?
Start with DLR alumni in Munich and ETH RSL alumni in Zurich. These two pipelines feed Agile Robots, Neura, ANYbotics, and Honda Research Institute Europe directly, and they are the highest-leverage profiles for tier 1 and tier 2 roles. Pull recent Humanoids and ICRA author lists, cross-reference to current employer, and prioritize anyone who has shipped on hardware in the last 24 months.
Is it worth trying to pull US-based humanoid engineers to Europe?
Rarely. The relocation friction is real and the comp gap with 1X, Figure, Apptronik, and Tesla Optimus is wide. The better play is the reverse: targeting Europeans who moved to the Bay Area in 2023 or 2024 and may be open to coming home to Munich or Zurich, particularly those with family ties. They are a small group, but they convert at a much higher rate than cold US targets.
How do I tell a tier-1 manipulation-ML engineer from a tier-3 ROS integrator on paper?
Look at three things: published papers with hardware results in the last three years, GitHub commits to Drake, Pinocchio, Crocoddyl, or Isaac Lab (not just rospy or roslaunch), and the specific lab they trained in. A tier-1 candidate will have all three. A tier-3 candidate will have a clean LinkedIn, a long list of integration projects, and limited public research output. Both are valuable. They are not interchangeable.
What is the single biggest mistake recruiters make on humanoid roles?
Searching by job title instead of by pedigree and stack. "Robotics Software Engineer" is the most common title in the pool and the least informative. Search by lab, by library, and by paper venue. Then verify with a 20-minute conversation about a specific technical problem, not a behavioral screen. The pool is small enough that one mis-targeted week of outreach burns goodwill you cannot get back.