Scale AI, which offers data-labeling providers to corporations that need to prepare machine studying fashions, has raised a $1 billion Collection F spherical from a slew of big-name institutional and company traders that embody Amazon and Meta.
The fundraise is a mixture of major and secondary funding, and is the most recent in a slew of huge enterprise capital investments in AI. Amazon lately closed a $4 billion funding in OpenAI rival Anthropic, and the likes of Mistral AI and Perplexity are additionally within the course of of elevating extra billion-dollar rounds at lofty valuations.
Earlier than this spherical, Scale AI had raised round $600 million in its eight-year historical past, together with a $325 million Collection E in 2021 that valued it at round $7 billion — double the valuation of its Collection D in 2020. Three years on, and regardless of headwinds that prompted it to put off 20% of its employees final yr, Scale AI is now valued at $13.8 billion — an indication of the occasions, as traders scramble to get forward within the AI gold rush.
The Collection F was led by Accel, which additionally led the corporate’s Collection A and took part in subsequent enterprise rounds.
Other than Amazon and Meta, Scale AI has attracted an assortment of recent traders: the enterprise arms of Cisco, Intel, AMD and ServiceNow participated, as did DFJ Development, WCM and investor Elad Gil. Lots of its current traders additionally returned: Nvidia, Coatue, Y Combinator (YC), Index Ventures, Founders Fund, Tiger World Administration, Thrive Capital, Spark Capital, Greenoaks, Wellington Administration, and former GitHub CEO, Nat Friedman.
Banking on the rising significance of information
Knowledge is the lifeblood of synthetic intelligence, which is why corporations specializing in knowledge administration and processing are faring nicely proper now. Simply final week, Weka stated it raised $140 million at a $1.6 billion post-money valuation to assist corporations construct knowledge pipelines for his or her AI purposes.
Based in 2016, Scale AI meshes machine studying with ‘human-in-the-loop’ oversight to handle and annotate giant volumes of information, which is significant for coaching AI methods throughout industries corresponding to autonomous autos.
However most knowledge is unstructured, and AI methods discover it troublesome to make use of such knowledge off the bat. It must be labeled, which is a resource-intensive endeavor, particularly with giant knowledge units. Scale AI offers corporations with knowledge that has been accurately annotated and primed for coaching fashions. It additionally specializes for various industries with totally different wants — a self-driving automobile firm will seemingly want labeled knowledge from cameras and Lidar, whereas pure language processing (NLP) use-cases will want annotated textual content.
The corporate’s prospects embody Microsoft, Toyota, GM, Meta, the U.S. Division of Protection and, as of final August, ChatGPT-maker OpenAI, which is tapping Scale AI to let corporations fine-tune its GPT-3.5 text-generating fashions.
Scale AI says it would use the brand new money to assist speed up the “abundance of frontier knowledge that can pave our street to synthetic normal intelligence.”
“Knowledge abundance just isn’t the default — it’s a alternative,” Scale AI’s CEO and founder, Alexandr Wang, stated in a press launch. “It requires bringing collectively the very best minds in engineering, operations, and AI. Our imaginative and prescient is one in all knowledge abundance, the place we’ve got the technique of manufacturing to proceed scaling frontier LLMs many extra orders of magnitude. We shouldn’t be data-constrained in attending to GPT-10.”