FriendliAI appoints Brian Yoo, former Moloco COO, as Chief Business Officer to drive Next phase of hypergrowth. AI industry veteran brings proven track record of scaling revenue 500x, securing $180M+ in capital, and leading global operations at the ~$4B machine learning company to accelerate adoption of The Frontier AI Inference Cloud. Apr 13, 2026 Prev Next 1 of 42,839 FriendliAI, The Frontier AI Inference Cloud, announced the appointment of Brian Yoo as Chief Business Officer. Joining from applied AI leader Moloco, Yoo will lead global commercial operations, go-to-market strategies, and partnerships as FriendliAI rapidly scales its high-performance, cost-efficient AI inference infrastructure worldwide. Yoo brings exceptional operational and strategic scaling expertise to FriendliAI. He most recently served as Chief Operating Officer at Moloco, where he built its global operations entirely from the ground up. Yoo created, managed, and scaled the Finance, Marketing, HR, BizOps, Legal, IT, and Workplace Operations functions. His leadership was the engine behind the company's massive expansion: he successfully grew revenue more than 500x to over $250 million, scaled headcount from a 10-person startup to a global workforce of over 600 employees, and led pivotal fundraising efforts that secured more than $180 million in capital, helping drive Moloco to a ~$4 billion valuation. "Brian's track record of building the operational engine behind an AI-driven startup and scaling it into a multi-billion dollar global powerhouse is simply remarkable," said Byung-Gon Chun, CEO of FriendliAI. "As we experience surging demand for our frontier inference infrastructure, Brian's unmatched expertise in hypergrowth, global commercial operations, and capital strategy is exactly what we need to accelerate our expansion and deliver unparalleled value to our customers." "As AI moves into production, performance at the inference layer directly determines how many tokens you can generate - and ultimately the margins you can capture," said Brian Yoo, Chief Business Officer at FriendliAI. "FriendliAI is positioned to maximize both, delivering industry-leading throughput and efficiency so our customers get the most out of every GPU. I am thrilled to join Byung-Gon and the incredible team here, especially now, as businesses that raced to build with large language models are beginning to heavily scrutinize their inference costs." Accelerating Market Leadership in AI Inference Yoo joins FriendliAI during a period of explosive momentum and enterprise adoption. FriendliAI's core technology, built by the researchers who pioneered the now-industry-standard "continuous batching" optimization technique, powers massive production workloads for rapidly growing AI companies. Under Yoo's commercial leadership, the company will focus on rapidly expanding its market share among AI-native startups, SaaS companies, and enterprises looking to escape inefficient open-source setups or expensive closed model APIs. FriendliAI is already delivering transformative results for its partners: Twelve Labs, a leader in video understanding, partnered with FriendliAI for production inference at scale, while NextDay AI reported processing trillions of tokens monthly while cutting GPU requirements by 50% after migrating to the platform.

VAST Data and TwelveLabs have partnered to bring advanced video intelligence capabilities to large-scale, secure video archives beyond public cloud deployments. The collaboration introduces TwelveLabs' first customer-managed deployment on the VAST AI Operating System, enabling video search, analytics and reasoning workflows where governance and data sovereignty are critical. TwelveLabs' foundation models, including Marengo for multimodal search and Pegasus for video understanding, will run on VAST's infrastructure designed to manage unstructured data at exabyte scale. The system features unified global namespace, trillion-vector scale storage and real-time data orchestration. Target industries include media and entertainment for content archive management, financial services for surveillance and compliance, and public sector agencies requiring on-premises video intelligence. The partnership addresses demand for deploying AI closer to where video data is created whilst maintaining regulatory compliance and cost efficiency.

TwelveLabs unveils Marengo 3.0: next-gen video understanding model now on Amazon Bedrock. In a groundbreaking move for small business owners managing video content, TwelveLabs has unveiled Marengo 3.0 during the AWS re:Invent conference in Las Vegas. Marketed as the world's most advanced video understanding model, Marengo 3.0 promises to revolutionize how businesses leverage video data, helping them utilize this previously unwieldy asset more effectively. TwelveLabs, a leader in video search and understanding, has designed Marengo 3.0 to provide deep insights into video content by "reading" it in a way that mimics human understanding. The platform processes audio, visual elements, text, and movement as interconnected components, making it easier to analyze and search for specific moments or themes within extensive video libraries. The advantages for small business owners are compelling. Marengo 3.0 offers a 50% reduction in video storage costs and doubles the speed at which videos can be indexed. This means businesses can manage larger volumes of video efficiently while lowering operational costs. As Jae Lee, CEO and co-founder of TwelveLabs, highlights, "Video represents 90% of digitized data, but that data has been largely unusable... Now, Marengo 3.0 shatters the limits of what is possible." The promise of immediate ROI is particularly appealing for small businesses looking to maximize every dollar spent. Among the standout features of Marengo 3.0 is its "native video understanding," meaning it was built specifically for video, as opposed to merely adapting image processing models. This model enables businesses to navigate complex scenes, track events over time, and connect dialogue with visual cues, making it an invaluable tool for sectors like media, entertainment, advertising, and public security. Small business owners can harness this technology to enhance various operational needs. For instance, in sports, the model allows teams to track player actions and jersey numbers, streamlining highlight identification for promotional content. Similarly, companies in advertising can gain insights into audience engagement by analyzing viewers' emotional reactions, thus improving content targeting and ROI. While these features excite many small business leaders, there are practical considerations and potential challenges. The deployment of Marengo 3.0 requires a robust technological infrastructure, particularly for those lacking IT expertise. Fortunately, its availability via Amazon Bedrock simplifies integration into existing AWS environments, a significant bonus for businesses already familiar with Amazon's cloud service. For non-technical teams, a learning curve exists in understanding how to fully utilize the model's capabilities. While TwelveLabs provides documentation and support, owners must invest time and resources into training their teams. Small businesses often operate under tight budgets, so weighing these costs against the anticipated benefits will be crucial. With the ability to conduct multimodal queries, users can combine text with images in their searches, enabling a more thorough exploration of video content. This functionality can unlock new insights across industries, establishing a more informed approach to content strategy. Moreover, Marengo 3.0 supports multiple language inputs, broadening its appeal and utility for diverse businesses operating in a multilingual landscape. Nishant Mehta, VP of AI Infrastructure at AWS, noted, "TwelveLabs' work in video understanding is transforming how entire industries manage their video capabilities." This statement underscores the importance of adopting advanced technologies to remain competitive. For small business owners, collaborating with innovators like TwelveLabs may open new revenue streams, as well as improve operational efficiencies. In a world increasingly dominated by video, leveraging tools like Marengo 3.0 is no longer just an option but a necessity. Small businesses can enhance their competitive advantage by adopting this powerful model, optimizing video content from various angles, and ultimately transforming the way they engage with their customers. TwelveLabs' latest offering is available through their platform and via Amazon Bedrock. With its API-first design and improved video support, the model stands ready to support a multitude of business needs while driving significant cost savings and efficiency gains. For further information on Marengo 3.0's capabilities and to explore how it can specifically benefit your business, visit the original announcement. Sarah Lewis is a small business news journalist and writer dedicated to keeping entrepreneurs informed on the latest industry trends, policy changes, and economic developments. With over a decade of experience in business reporting, Sarah has covered breaking news, market insights, and success stories that impact small business owners. Her work has been featured in prominent business publications, delivering timely and actionable information to help entrepreneurs stay ahead. When she's not covering small business news, Sarah enjoys exploring new coffee shops and perfecting her homemade pasta recipes.