Sakana AI vs Skild AI
Side-by-side comparison of AI visibility scores, market position, and capabilities
Tokyo AI lab co-founded by Llion Jones (Transformer paper co-author); Japan's most valuable AI startup at $2.65B; nature-inspired model merging and evolutionary AI; backed by Khosla and NEA.
About
Sakana AI is a Tokyo-based AI research laboratory co-founded by David Ha and Llion Jones, the latter a co-author of the original Transformer paper that underpins modern large language models. Established to explore nature-inspired approaches to artificial intelligence, Sakana takes its name from the Japanese word for fish — evoking swarm intelligence and emergent collective behavior as a design philosophy for AI systems rather than scaling a single monolithic model.\n\nThe lab's research focuses on evolutionary and compositional AI architectures: building capable AI systems by combining and evolving smaller specialized models rather than training ever-larger ones. This approach has produced novel techniques in model merging, neural architecture search, and AI-generated AI research. Sakana's work targets both academic contribution and practical deployment, with research that attracts attention from leading institutions globally.\n\nSakana AI has become Japan's most valuable AI startup, reaching a $2.65B valuation backed by top-tier investors including Khosla Ventures, NEA, and In-Q-Tel. Its prominence reflects Japan's strategic push to develop sovereign AI capabilities and the global research community's interest in alternative scaling paradigms. As foundation model costs climb, Sakana's nature-inspired compositional approach offers a potentially more efficient path to capable AI — making it one of the most intellectually distinctive labs in the 2025–2026 AI landscape.
Universal robot brain startup raised $1.4B Series C at $14B valuation in Jan 2026 led by SoftBank with Nvidia and Bezos; $30M 2025 revenue; deployed at Foxconn
About
Skild AI is building a universal robot brain — a foundation model for physical intelligence that can power a broad range of robot types without requiring task-specific training for each deployment. Founded to solve the fragmentation problem in robotics AI, where every robot type and task requires separate model development, Skild's approach trains a single generalist model on diverse robotic data and fine-tunes it rapidly for specific deployments. The company was founded by robotics AI researchers who identified the model reuse gap as the primary barrier to scalable robot deployment.\n\nSkild's generalist robot model has been deployed across more than 30 distinct robot types — spanning manipulation arms, mobile platforms, and humanoid form factors — demonstrating the cross-hardware generalization that most robot AI systems lack. The platform targets robotics manufacturers, logistics operators, and industrial automation companies that need AI-capable robots but lack the internal ML infrastructure to develop foundation models themselves. By offering a model-as-a-service layer, Skild enables robot OEMs and systems integrators to add AI capabilities without building the underlying research infrastructure.\n\nSkild AI raised a $1.4 billion Series C in January 2026 at a $14 billion valuation, led by SoftBank with co-investment from NVIDIA and Jeff Bezos. The round was one of the largest in robotics AI history and reflects institutional conviction in the physical AI market's scale. With $30 million in 2025 revenue and accelerating enterprise deployments, Skild is building the financial foundation to match its valuation. The SoftBank-NVIDIA investor combination positions Skild at the center of the global robotics deployment wave.
AI Visibility Head-to-Head
Key Details
Capabilities & Ecosystem
Capabilities
Also Compare
Track AI Visibility in Real Time
Monitor how your brand performs across ChatGPT, Gemini, Perplexity, Claude, and Grok daily.