Over 80% of enterprises are turning to open-source AI fashions. Nevertheless, the problem isn’t accessing highly effective fashions, fairly it’s deploying them effectively and securely. In the present day, firms should sew collectively greater than 10 totally different elements simply to start deployment, with every optimization requiring 1000’s of engineering hours. In the present day, AI infrastructure startup Pipeshift introduced their $2.5M seed spherical to resolve this problem, launching a new-age Platform-as-a-Service (PaaS) that allows engineering groups to orchestrate AI workloads throughout any infrastructure – cloud or on-premises – with unprecedented velocity and management.
This spherical was led by Y Combinator and SenseAI Ventures, with further participation from Arka Enterprise Labs, Good Information Ventures, Nivesha Ventures, Astir VC, GradCapital, and MyAsiaVC. Seasoned Silicon Valley angels like Kulveer Taggar (CEO of Zuess), Umur Cubukcu (CEO of Ubicloud and former Head of PostgreSQL at Azure), and Krishna Mehra (former Head of Engineering at Meta and co-founder of Capillary Applied sciences) additionally joined the spherical.
Not like current gamers who’re GPU-brokers providing one-size-fits-all options, Pipeshift understands the enterprise want for management and suppleness of infrastructure and affords an end-to-end MLOps stack for enterprises to coach, deploy, and scale open-source GenAI fashions — LLMs, imaginative and prescient fashions, audio fashions, and picture fashions — throughout any cloud or on-prem GPUs. Consequently, enterprises can deploy their AI workloads in manufacturing sooner and extra reliably. Moreover, as we see extra mannequin and {hardware} architectures coming into the market, Pipeshift future-proofs infrastructure investments by providing flexibility by means of their modular MLOps stack that permits enterprises to carry down their GPU infrastructure prices with none further engineering effort.
“2025 marks the 12 months when GenAI transitions into manufacturing and engineering groups are witnessing the advantages of utilizing open-source fashions in-house. This affords excessive ranges of privateness and management alongside enhanced efficiency and decrease prices. Nevertheless, it is a advanced and costly course of involving a number of elements being stitched collectively.” stated Arko Chattopadhyay, Co-Founder and CEO of Pipeshift. He added “Pipeshift’s enterprise-grade orchestration platform eradicates the necessity for such in depth engineering investments by not solely simplifying deployment but additionally maximizing the manufacturing throughput.
Pipeshift’s founding staff has been engaged on this downside for over a 12 months. Co-founders, Arko Chattopadhyay, Enrique Ferrao and Pranav Reddy met throughout their undergraduate research at Manipal Institute of Expertise, the place they have been main a defence robotics non-profit supported by NVIDIA, Dassault Techniques, and SICK Sensor Intelligence. The staff was targeted on the deployment of machine studying fashions on the cloud and edge for processing real-time sensor information and task-specific imaginative and prescient fashions. In 2023, they scaled a Llama2-powered enterprise search app inside a company of over 1000 workers, utterly on-prem. Whereas constructing this, they noticed the challenges of working and scaling personal AI workloads in manufacturing, pushing them to start out optimizing the identical at Pipeshift.
The timing for Pipeshift is important. As AI reshapes markets and redefines competitors, firms know the rewards for seizing the second are immense. Nevertheless, safety and information privateness dangers loom massive, demanding safety for proprietary IP. These challenges compound in a quickly evolving know-how panorama the place missteps result in costly delays and misplaced alternatives. Including to this complexity is the distinctiveness of each enterprise downside. No two AI methods are the identical, and each deployment should align with the distinct wants of the group. Pipeshift solves this by bringing within the flexibility and precision of open-source AI fashions and the scalability of their enterprise MLOps platform. Companies overcome these challenges whereas managing useful resource calls for and guaranteeing compliance — all with out shedding sight of their broader targets.
“Enterprises choose open-source GenAI for the advantages of privateness, mannequin possession, and decrease prices. Nevertheless, transitioning GenAI to manufacturing stays a fancy and costly course of requiring a number of elements to be stitched.” stated Rahul Agarwalla, Managing Companion of SenseAI Ventures. He added, “Pipeshift’s enterprise-grade orchestration platform eliminates the necessity for such in depth engineering investments by not solely simplifying deployment but additionally maximizing the manufacturing throughput.”
Anu Mangaly, Director of Software program Engineering at NetApp stated, “Pipeshift’s potential to orchestrate current GPUs to ship >500 tokens/second for fashions like Llama 3.1 8B with none compression or quantization of the LLM is extraordinarily spectacular, permitting companies to cut back their compute footprint and prices in manufacturing, whereas delivering enhanced consumer experiences which can be additionally personal and safe.” She additionally shared, “At NetApp, we understood the enterprise want for a single information cloth throughout cloud, on-prem, and hybrid arrange. Pipeshift’s orchestration permits for enterprises to unlock the identical potential from the brand new technology of AI fashions all inside their infrastructure.”
Yash Hemaraj, Founding Companion at Arka Enterprise Labs and Basic Companion at BGV added: “We invested in Pipeshift as a result of their progressive platform addresses a important want in enterprise AI adoption, enabling seamless deployment of open-source language fashions. The founding staff’s deep technical experience and observe file in scaling AI options impressed us immensely. Pipeshift’s imaginative and prescient aligns completely with our give attention to transformative Enterprise AI firms, notably these bridging the US-India tech hall, making them a super match for our portfolio.”
Having already labored with over 30 firms together with NetApp, Pipeshift goals to grow to be the trusted companion for organizations seeking to unlock AI’s potential whereas sustaining management of their infrastructure and information.
