Scaled Inference raises $13.6M at a valuation of about $60M from Khosla Ventures and others, to offer machine learning software as a cloud service
Scaled Inference Raises $13.6 Million to Build Out Machine Learning — Artificial intelligence and machine learning are back in vogue. Tweets: @felicisventures Tweets: Felicis Ventures / @felicisventures : A new $13.6M round for our machine learning co @ScaledInference led by @khoslaventures via @WSJ - http://blogs.wsj.com/... cc @svangel @DCVC
Context & Ripple Effects
Scaled Inference's $13.6M round at a ~$60M valuation is one of the first venture bets on selling machine learning purely as a cloud service rather than shipping it as in-house software, with GraphLab's $18.5M raise days later confirming investors were funding several ML-as-a-play at once.
The bet looks small against what the category became: the same thesis — rent out the hard part of AI instead of building it yourself — later carried Scale AI from its $155M data-labeling round in 2020 to a $1B Series F at $13.8B in 2024, with a tender offer reportedly targeting up to $25B.
First-order effects
- Scaled Inference gets the capital to stand up machine learning as a hosted service, while lead investor Khosla Ventures and Felicis Ventures take positions at a ~$60M entry price on the inference-operator thesis.
- Companies wanting ML capability gain a new procurement path: buying it over the cloud rather than hiring researchers and building pipelines internally.
Second-order effects
- Competing ML startups like GraphLab/Dato are forced into the same hosted-service framing within weeks of this round, turning 'ML as cloud service' into a contested pricing and packaging battleground rather than a differentiator.
- Cloud providers become the default distribution and billing layer for these services, capturing margin on every ML workload sold through their platforms.
Third-order effects
- If the pattern holds across later waves, durable value in applied AI accrues not to model authors but to the operating layer that serves models and manages data at scale — the position Scale AI's multi-billion-dollar trajectory now occupies.
- Venture firms like Khosla Ventures that seeded this early round establish a repeatable playbook: fund the infrastructure-of-intelligence layer early, then hold exposure as the category re-rates orders of magnitude higher.
The trend: Machine learning is migrating from an in-house research discipline to rented cloud infrastructure, with each venture wave — model serving, then data operations — repricing the operator layer upward.