Deep Signal: Skild AI Series B Funding Round

Skild AI’s $4.5B Series B: Foundation Model Valuations Meet Verification Gaps

What Happened

Skild AI closed a Series B funding round in summer 2025 at a reported valuation of approximately $4.5 billion. The round size itself is contested: Crunchbase reports $135 million raised, while TechCrunch cited figures closer to $500 million. That $365 million discrepancy is not a rounding error — it represents a fundamental transparency problem that runs through the company’s entire financial narrative.

The Pittsburgh-based AI software company, founded in 2023 by Carnegie Mellon researchers Deepak Pathak and Abhinav Gupta, is building what it calls an “omni-bodied” foundation model — a single AI system capable of controlling quadrupeds, humanoids, tabletop arms, and mobile manipulators without robot-specific training. The Series B followed a Series A and preceded a reported $1.4 billion Series C in January 2026 led by SoftBank, pushing total disclosed capital to over $1.83 billion and valuation above $14 billion within roughly 18 months of founding.

Why It Matters

The Skild Series B is a useful marker for where foundation model valuations stood in mid-2025: high on narrative, thin on verification. The company reported approximately $30 million in unaudited 2025 revenue, described as accumulated within “a few months.” At the $4.5 billion Series B valuation, that implies a revenue multiple of roughly 150x. At the subsequent $14 billion Series C valuation, the multiple expands to approximately 467x.

Metric	Value
Series B valuation	~$4.5B
Series C valuation (Jan 2026)	>$14B
Reported 2025 revenue	~$30M (unaudited)
Revenue multiple at Series C	~467x
Total capital raised	>$1.83B (Crunchbase) / >$2B (CEO)
Series B size (Crunchbase)	$135M
Series B size (TechCrunch)	~$500M
Skild Brain deployment status	LIMITED
Named enterprise customers	0 publicly disclosed
Safety certifications disclosed	0

The valuation trajectory — tripling in roughly seven months between Series B and Series C — reflects investor conviction in the generalist robotics control thesis rather than demonstrated commercial scale. That thesis is technically coherent: per-robot, per-task training pipelines are expensive and brittle, and a universal control layer would compress deployment costs significantly. The problem is that no independently verifiable production deployments, customer references, or standardized benchmarks have been made public.

Who Is Affected

Physical AI platform competitors face the most direct pressure. Field AI, which focuses on autonomous navigation in unstructured environments, and 1X Technologies, developing humanoid platforms with their own control software, are both pursuing adjacent generalist control approaches. Skild’s capital advantage — over $1.8 billion versus significantly smaller raises at both companies — creates a meaningful R&D and data acquisition gap if Skild converts its runway into labeled robot interaction data at scale. Flexion, pursuing similar foundation model positioning, faces comparable dynamics.

Robot hardware OEMs occupy an ambiguous position. Skild Brain is hardware-agnostic by design, which means Boston Dynamics, Unitree, Agility Robotics, and others could theoretically become distribution channels rather than competitors. However, each of those companies has incentives to develop proprietary control software to avoid commoditization of their hardware margins. The strategic investor roster — Nvidia NVentures, Samsung, LG — suggests Skild is actively courting hardware ecosystem alignment, but no co-development agreements have been publicly disclosed.

Zebra Technologies is directly implicated by an unverified acquisition report suggesting Skild acquired Zebra’s robotics automation business, including the Symmetry Fulfillment orchestration platform. If confirmed, this would give Skild an installed base in warehouse automation and existing customer relationships — materially changing the revenue verification picture. The acquisition remains unconfirmed from accessible sources. HIGH CONFIDENCE that confirmation or denial will surface within 12 months given the capital scale involved.

Enterprise end-users in warehousing, manufacturing, healthcare, and data centers are the ultimate validation layer. Strategic investors Schneider Electric (manufacturing/energy infrastructure), CommonSpirit Health (healthcare), and Salesforce Ventures (enterprise software) each represent potential go-to-market channels. MODERATE CONFIDENCE these relationships convert to pilot deployments within 18 months; LOW CONFIDENCE they convert to disclosed production contracts at scale within that window.

What to Watch

Q3 2025 – Q1 2026: Confirmation or denial of the Zebra Technologies robotics automation acquisition. This is the single highest-impact near-term disclosure.

Q2 2026: Whether any of the Series C strategic investors (Schneider Electric, CommonSpirit, Samsung, LG) announce named commercial deployments of Skild Brain. Absence of announcements by mid-2026 would be a meaningful negative signal given the 18-month window since the Series B.

Full-year 2025 revenue verification: Crossing $100 million ARR with audited figures would begin to justify the valuation structure. Remaining at sub-$50 million unaudited revenue through 2026 would intensify down-round risk.

Competitive benchmark publication: If Field AI, 1X, or any incumbent publishes standardized cross-morphology control benchmarks before Skild does, it shifts the narrative framing on who owns the generalist control category definition.

The broader industry pattern is clear: 2025’s $13.8 billion in robotics startup funding, up 77% from $7.8 billion in 2024, is producing valuations that require production-grade proof points to sustain. Skild’s next 12 months are a test of whether foundation model capital efficiency translates to enterprise deployment velocity — or whether the verification gap widens.