Superlearner by Ineffable Intelligence: status and timeline

Milestones

Next · First model results

Company foundedNov 2025Complete.
Record $1.1B seed round closes27 Apr 2026Complete.
First model resultsTarget 2026Current milestone.

Most important updates

Ineffable Intelligence raises record $1.1B seed round at $5.1B valuation with Sequoia and Lightspeed27 Apr 2026

Upcoming

2026First model results (next)

Current obstacles

Open-ended discoverySelf-play works in games; extending to science requires new exploration and evaluation methods.
Validating novel knowledgeDetermining if AI discoveries are correct and novel (not hallucinated) is critical.

Physics limits

Semantic groundingSelf-play can discover logical patterns, but translating them to real-world problems requires the system to understand physical laws, which self-play alone may not provide.
Evaluation in open domainsGames have clear win/loss conditions; scientific discovery lacks obvious reward signals, making it hard to train an RL system that discovers genuinely useful knowledge.

How it works

3 parts

Self-generated problems

Create its own learning challenges

Generates problems and puzzles that test understanding, rather than receiving curated training data.

Reward signals

Learn from success and failure

Adjusts approach based on solution success; builds understanding of underlying principles.

Pattern discovery

Emergent knowledge

Discovers general principles and scientific frameworks that emerge from autonomous exploration.

Spec sheet

Spec	Superlearner
Seed funding	$1.1 billion^R (reported)
Post-money valuation (April 2026)	$5.1 billion^R (reported)

^R reported by the company

Update log

1 update

Mon 27 Apr

Major: Ineffable Intelligence raises record $1.1B seed round at $5.1B valuation with Sequoia and LightspeedPressFunding

About Ineffable Intelligence

The team behind Superlearner

Ineffable Intelligence

Superlearner, reinforcement learning, self-play AI

Ineffable Intelligence builds a 'superlearner' AI that discovers knowledge through self-play and reinforcement learning without human-labeled data, founded by David Silver (AlphaGo, AlphaZero).

Record $1.1B seed round in April 2026 at $5.1B valuation, Europe's largest ever.
Co-led by Sequoia and Lightspeed; Nvidia, Google, UK Sovereign AI Fund participated.
David Silver (AlphaGo lead) + 4 DeepMind alumni. First models expected late 2026.

ineffable.ai (opens ineffable.ai)Company profile →

Founded: 20251 yrs
Headquarters: United Kingdom
Status: Private
Valuation: $5.1BprivateApr 2026
Raised: $1.1B1 round
Last round: Seed · $1.1B27 Apr 2026
Works in: AIFrontier models
Coverage: 1 program · 1 updatelatest 27 Apr 2026checked 11 Oct
Lead investors: Sequoia CapitalLightspeed Venture Partners
Partners: Google DeepMind (opens deepmind.com)
Investors: Nvidia (opens nvidia.com)
People: David Silver (opens en.wikipedia.org)Founder & CEOWojciech CzarneckiCo-founderLasse EspeholtCo-founder

Spec

Superlearner

Seed funding

$1.1 billion^R (reported)

Post-money valuation (April 2026)

$5.1 billion^R (reported)