MLOps startup Vessl AI eyes 50% sales from US in 2025

Vessl AI, a South Korea-native MLOps platform-operating startup, is envisioning a takeoff in the US next year after building a concrete client base at home in the four years since its inception.

“We have set a goal to raise 50% of our revenue from the US market next year,” An Jaeman, the chief executive officer of Vessl AI, said in an interview with The Korea Economic Daily last week.

The Korean MLOps platform engineering startup, founded in 2020, currently operates mostly in the local market but is aiming higher -- targeting about $8 million in annual recurring revenue in 2025.

MLOps is short for machine learning operations and is designed to offer end-to-end services for machine learning model development, deployment, maintenance and management covering the lifespan of artificial intelligence functions.

Its demand is expected to surge as companies across the board are eager to roll out their own unique machine learning models in the wake of the sensational debut of OpenAI’s ChatGPT and amid accelerating digitalization in corporations worldwide.

Global industry research company Allied Market Research forecasts the global MLOps market will grow to $37.4 billion in 2032 from 2023 at a compound annual growth rate of 39.3%.

“As long as the AI market expands and the need for AI and computing increases, the market for our business (in computing infrastructure) is also expected to grow rapidly,” said An.

“This is a hot market but not many competitors are armed with highly advanced (MLOps platform) technologies like us.”

The Vessl AI CEO is confident the Korean startup will repeat its home-ground success in the US.

“I am 100% sure about our business in the US, where demand (for MLOps) is vast. There are not many platforms that can support all the demands (in the US) partly because our competitors in the country have been acquired by bigger ones,” said An.

“Our platform is highly competitive even on the global stage and many clients in the US, including startups, are highly interested in our product.”

'HOT MARKET' STILL IN INFANCY

An founded Vessl AI in 2020 after encountering multiple challenges in implementing machine learning models in a real business environment as an engineering vice president at a Korean healthcare AI startup.

He joined hands with former DevOps and machine learning engineers of Sendbird and Google to start his own company and develop MLOps in anticipation of the high demand for the service amid the rapid arrival of the AI era.

Vessl AI offers an end-to-end MLOps platform that provides a unified ecosystem across the entire AI life cycle to increase productivity, speed and reliability while reducing risk.

It enables users to train, develop and deploy tens of generative AI and large language models (LLMs) with machine learning in any cloud environment.

The software-as-a-service (SaaS) platform is a rare MLOps system supported by multiple cloud and hybrid cloud services that can operate in any environment including on-premises, according to the Korean startup. 

“Because we use multiple cloud services, we can offer the most optimized graphics process unit (GPU) infrastructure to save our clients up to 70% in computing costs,” said An.

MULTIPLE CLOUD AND HYBRID CLOUD SERVICES

Vessl AI supports cloud services of not only multinational players such as Amazon Web Services (AWS) and Google Cloud Platform but also Korean firms such as Naver Cloud, NHN Cloud and KT Cloud.

Thanks to its versatility, it has secured major conglomerates including Hyundai Motor Co., Samsung and Cognex Corp., as well as leading academic organizations such as Korea Advanced Institute of Science & Technology (KAIST) AI and Seoul National University as its clients since the release of the beta service in March 2021.  

Hyundai Motor’s autonomous driving technology department and Korean startups like ScatterLab, Wrtn and Yanolja also use Vessl AI’s MLOps platform to train various AI models.

Global big tech companies Oracle Corp. and Google Cloud have also partnered with Vessl AI on various projects to develop more powerful AI platforms and for joint marketing, An said.

About 6,000 engineers across the world use Vessl AI’s MLOps platform, the company said.

The Korean MLOps developer set up its US entity, which now serves as its headquarters, in 2021 to better serve its international clients and partners.

It currently operates MLOps Now SF to expand its network in the US.

An will move to the US in September to lead the US operations with three other Vessl AI employees.

SERIES A FUNDING ROUND TO RAISE UP TO $15 MILLION

For its global push, Vessl AI has been in talks with several investors, both local and foreign, to raise up to 20 billion won ($15 million) in a Series A funding round.

“We are expected to complete the Series A round in August,” said An, adding that the company will use the proceeds for R&D and sales and marketing in the US.

Of Vessl AI’s 35 staff, nearly 70% are software developers and engineers.

Since its inception, Vessl AI has raised a total of 6 billion won in seed and pre-A rounds from investors such as KB Investment Co., Mirae Asset Venture Investment Co. and SpringCamp.

Earlier this year, Vessl AI was also selected among 20 Korean startups for support from Asan Voyager, a program designed to assist Korean startups entering the US by providing acceleration, coaching, community learning and long-term residency support.

ARTIFICIAL GENERAL INTELLIGENCE

Vessl AI’s ultimate goal is to build an artificial general intelligence (AGI) system, An said.

Current AI technologies function within a set of pre-determined parameters and need to be trained but AGI possesses autonomous self-control, a reasonable degree of self-understanding and self-teaching abilities.

AGI with human abilities remains a theoretical concept and a research goal.

But the AGI era could come earlier than expected, An said.

“The next two to three years are the most critical period,” said the Vessl AI CEO.

"The future many are anxious about, where a lot of jobs are replaced and decisions are made by AI, could become a reality sooner than expected," says An.

"Vessl AI will take prompt action to ride the AI wave over the next three to five years.”

Write to Sookyung Seo at skseo@hankyung.com

Jennifer Nicholson-Breen edited this article.