Comet

Opik becomes the de facto open-source LLM evaluation standard (think Grafana for AI observability), driving enterprise cloud subscriptions to $20-30M ARR by 2028 and making Comet an attractive acqui-hire target for a cloud provider or observability platform (e.g., Datadog, New Relic, Elastic) at 15-20x revenue, implying a $400-500M exit — roughly 75% above the $284M valuation, though the $88M preference stack absorbs ~18% of any exit proceeds before common stockholders see a dollar.

Comet holds a small but loyal mid-market user base while Weights & Biases dominates enterprise and MLflow/SageMaker absorbs price-sensitive users; Opik gains modest traction in LLM eval but fails to break through crowded competition from Langfuse, Arize, and Braintrust; revenue reaches $8-12M ARR by 2027 but at a compressed multiple (8-12x) yields a $80-150M exit or a down-round recap — at or below the $284M current valuation, and after $88M in liquidation preferences, common shareholders recover little.

AWS deepens SageMaker experiment tracking (already free for Prime users), Google Vertex AI bundles LLM evaluation natively, and Weights & Biases — with 10x the funding and brand recognition — absorbs the remaining mid-market; Comet's revenue stagnates below $5M, runway runs dry with no new funding since March 2022, and the company is sold for parts or shuttered; after $88M in liquidation preferences, a $50-100M distressed exit returns nothing to common stockholders, representing a near-total loss on employee equity.