Conviction capital for frontier AI

Investing where AGI research becomes durable advantage.

Investment firm focused on companies turning advanced research into infrastructure, products and durable returns. We combine capital discipline, technical diligence and a long-duration view of innovation cycles in artificial general intelligence.

Capital deployed, performance and leverage.

14 Portfolio startups

AGI companies funded since 2021 across the globe.

3.1× Average ROI

Weighted return on investment across the entire portfolio.

$87M Follow-on raised

Total capital raised by our startups in subsequent rounds (Series A+).

An investment thesis
grounded in research.

Angel investor with a core conviction: artificial general intelligence emerges at the intersection of deep reinforcement learning, neuro-symbolic reasoning and internal world models — and the startups mastering this convergence will define the next computational paradigm.

The focus is on teams that move beyond simple LLM scaling to explore the composability of latent representations, causal attention mechanisms, hierarchical sub-goal generation and continual learning without catastrophic forgetting. AGI will not be a single model, but an ecosystem of specialized modules orchestrated by a recursive meta-learning loop.

Beyond funding, technical support is provided to founders: multi-agent system architecture, evaluation strategies on ARC-AGI-type reasoning benchmarks, intrinsic reward design for autonomous exploration, and navigating regulatory challenges related to general-capability systems.

Paradigmatic convergence

AGI emerges from the convergence of compute scaling, persistent episodic memory architectures and counterfactual causal reasoning systems.

Technical support

Direct involvement in architectural decisions: computational graph topologies, model distillation strategies and multi-task training curricula design.

Empirical rigor

Startup evaluation based on reproducible metrics: out-of-distribution generalization capability, adversarial robustness and performance on zero-shot compositional reasoning tasks.

Scalable alignment

Priority to teams integrating iterative RLHF, formal verification of safety properties, mechanistic interpretability and corrigibility protocols from the earliest development phases.

Where we believe
durable value is created.

Pre-seed & Seed

Investing at the earliest stage, where fundamental innovations in knowledge representation and world model learning are still at proof-of-concept stage.

  • Pre-product funding for fundamental AGI research
  • Supporting teams building abstract reasoning benchmarks
  • Investment horizon compatible with long research cycles
  • Co-investing with funds specialized in deep tech and frontier AI

Beneficial superintelligence

The trajectory toward ASI (Artificial Superintelligence) demands investments now in computational governance frameworks, safe shutdown protocols and value alignment for recursively self-improving optimization systems.

  • Research in corrigibility and interruptibility of autonomous agents
  • Formal verification of safety invariants for adaptive systems
  • Containment protocols for AI with superhuman-level capabilities
  • Decentralized governance mechanisms for distributed AGI systems

The technology layers
where we take conviction positions.

Neuro-symbolic reasoning

Kahneman's System 2 × deep networks
$$K(x) = \min_p \{\,|p| : U(p) = x\,\}$$

Hybrid architectures combining Bayesian probabilistic inference with differentiable symbolic reasoning modules. Focus on program synthesis, gradient-based constraint solving and hierarchical concept abstraction via primitive library learning (DreamCoder, neural λ-calculus).

Recursive OODA agents

Self-improving Observe-Orient-Decide-Act loops
$$V^*(s) = \max_a \left[R(s,a) + \gamma \sum_{s'} P(s'|s,a)\,V^*(s')\right]$$

Multi-agent systems with hierarchical planning through chain-of-thought task decomposition, dynamic tool use and sparse-attention working memory. Exploring cognitive scaffolding architectures where agents build and revise their own decision heuristics through meta-learning.

Mechanistic interpretability

Reverse-engineering internal computational circuits
$$\text{Attn}(Q,K,V) = \text{softmax}\!\left(\frac{QK^T}{\sqrt{d_k}}\right)V$$

Research in feature decomposition via sparse autoencoders, induction circuit mapping in transformers, mesa-optimizer detection through internal gradient analysis and automated red-teaming protocols for systems with unanticipated emergent capabilities.

World Models & Simulation

Generative models of causal dynamics
$$F = D_{\text{KL}}\!\left[q(\theta|x) \,\|\, p(\theta)\right] - \mathbb{E}_q[\log p(x|\theta)]$$

Architectures learning compressed representations of the physical world through self-supervised prediction. Video prediction models, intuitive physics models, mental simulation via latent diffusion and imagination planning in structured latent spaces (Yann LeCun's JEPA, Genie).

Optimal Compute & Scaling

Chinchilla laws, MoE and structured sparsity
$$C_{\text{opt}}(N) \approx 6ND \;\;\text{FLOPs}$$

Infrastructure for frontier model training: tensor/pipeline/expert parallelism, mixed-precision quantization (FP8/INT4), mixture-of-experts architecture with conditional routing, ML compilers (XLA, Triton) and low-latency communication fabrics for multi-datacenter clusters.

Multimodal grounding

Sensorimotor anchoring of representations
$$\mathcal{L}_{\text{CLIP}} = -\log \frac{e^{\text{sim}(z_i, z_j)/\tau}}{\sum_k e^{\text{sim}(z_i, z_k)/\tau}}$$

Unified cross-modal encoding systems via contrastive learning (CLIP, ImageBind) extended to proprioception, haptics and vestibular flow. The symbol grounding problem hypothesis resolved through embodied AI and active interaction with the physical environment via sim-to-real transfer.

What drives an
investment decision.

The Team
01

The most important criterion: exceptional founders.

  • Tier-1 conference publications (NeurIPS, ICML, ICLR)
  • Experience training models with >10B parameters
  • Mastery of frontier MLOps stack (FSDP, DeepSpeed, Megatron)
  • Understanding of information theory and mathematical foundations
  • Track record in reasoning, alignment or world models research
The Vision
03

A credible roadmap toward systems with general cognitive capability.

  • Articulated technical thesis on the critical path toward AGI
  • Milestones defined in terms of measurable emergent capabilities
  • Compute and data strategy for the next 24 months
  • Alignment and safety plan integrated from the design phase
  • Clear positioning in the frontier AI ecosystem

Building a company that genuinely expands the frontier of capability?

We invest early in teams that combine technical edge, commercial ambition and execution discipline. Infrastructure, models, agents, safety and production tooling: if the technology creates durable advantage, it is relevant to us.