SK Telecom has entered into a strategic partnership agreement with GPUaaS provider Lambda, their expansion into the AI infrastructure industry in full swing. SKT and Lambda’s AI data center will open in Seoul this December, and Lambda’s Korea region will be the first of its kind launched in APAC. Furthermore, SKT will be launching ‘GPUaaS(GPU-as-a-Service)’, an AI cloud subscription service using Lambda’s GPU resources.
Founded by AI engineers in 2012, Lambda is a GPUaas company that provides AI cloud services to many clients including global big tech companies. SK Telecom Newsroom interviewed Stephen Balaban, CEO of Lambda, on the strategy that made them the emerging AI company it is today, his view on the AI infrastructure market, and the significance of their partnership with SKT. Balaban will also be visiting Korea to attend the “SK AI Summit 2024” held on November 4th-5th at COEX, and will be giving his keynote presentation on the first day of the event.
AI Infrastructure Market Innovator ‘Lambda’ CEO says – To Win in Competitive Market, Simply Build Best Product
Stephen Balaban, CEO and co-founder of Lambda
Q. Lambda is successfully expanding its business in the fiercely competitive AI infrastructure market. What do you think is the most important factor?
At the end of the day, the best way to win in a competitive market is to build the best product and let customers vote with their feet. The number one important factor in winning is how well we understand our customer and how much effort we put into building a product for them.
We’ve strived to be the first in the industry to allow AI/ML teams to seamlessly and affordably build, test and deploy AI products. We were the first to offer self-serve multi-node GPU clusters for weeks at the time, instead of years. We have also consistently been one of the fastest companies to deploy NVIDIA’s latest chips to our cloud.
Q. Lambda started as a deep learning company and has evolved into an AI infrastructure provider. What is the background of this change, and what industry trends have influenced it?
Before Lambda evolved into an AI infrastructure provider, we worked on several AI projects, including image and face recognition, wearable electronics and even image generation. While working on these, we continuously ran into the problem of not having access to a large quantity of affordable GPUs to power our research. We realized this challenge wasn’t unique to Lambda and that was the birth of that pivot.
Q. The AI ecosystem is rapidly forming. How do you expect industries to change due to AI in the future? What role do you see AI DC or AI infrastructure playing in this ecosystem?
AI is going to fundamentally restructure every aspect of our economy. I’m a software engineer so I’m acutely aware of the productivity boost that is offered by the current state-of-the-art LLMs. It’s clear that this trend of scaling is going to continue, and that our entire field is about to get turned on its head.
We’re also seeing the gap between proprietary and open source models narrow. This lets companies build powerful AI applications while maintaining control of their data. Of course, Lambda benefits from the fact that open models generate even more demand for our compute infrastructure.
Q. How do you think the development of AI technology and industry will impact the daily lives of general consumers? What do you think is the most important factor in this process?
We’ve already seen ChatGPT become the fastest growing application in the history of capitalism. This is because it improves the lives of those who use it. As things like ChatGPT, Apple Intelligence, and other AI based systems become more and more mainstream, we’ll see a shift in how people work, entertain themselves, do science, and conduct commerce. The decision making in each of those domains is getting augmented with AI, and even the core logic of software is getting gutted and replaced with AI.
Q. How does Lambda view the AI infrastructure and GPU market in Korea and Northeast Asia? What are your expectations?
There’s a rapid pace of AI innovation happening in Korea. Every day, I see a cool new research paper or robotics demo from somebody at KAIST or SNU. Those research advances are powered by GPU infrastructure.
On the next layer up, forward-thinking leaders like James at SKT are leading the way in terms of integrating AI into their enterprise and showing other enterprises how to integrate AI into their business.
We expect to see growing GPU demand in both Korea and the broader Asia-Pacific markets.
Lambda Prepares for New Economy with SKT – CEO to Give Keynote at ‘SK AI Summit’
Ryu Young-sang, CEO of SK Telecom (pictured left) and Stephen Balaban, CEO and co-founder of Lambda, pose for a commemorative photo after discussing specific cooperation plans at Lambda’s San Jose headquarters in January 2024.
Q. It seems that you have partnerships with various companies, including SK Telecom, in the AI business. What is the significance of such partnerships in the AI era?
SKT and their leadership team have been astute investors in the AI space, with investments in Perplexity, Anthropic, and Lambda. SKT has deep and long-lasting relationships with every major company in Korea and with other telecommunication operators across the world.
SKT is well-positioned with data center infrastructure that is ready to host GPUs to be made available to both local customers in Korea as well as the rest of the world.
Q. What is the significance and expected effect of this collaboration with SKT?
As we continue working toward a world where access to compute is as easy and ubiquitous as electricity, there’s no better partner than SKT – a telecommunications company that connects people around the world, and truly understands how AI works.
Beyond that, this partnership with SKT gives us the opportunity to expand cloud services in Korea, a growing hub for tech and AI innovation. We’re looking forward to joining forces with SKT to accelerate AI deployment and innovation across the region.
Q. How do you evaluate SKT’s approach and actions in the AI industry?
SKT recognizes that the entire world is going to be transformed in the coming decade – and they’re making major investments to make sure we’re positioning both Lambda and SKT as core infrastructure and service providers in the new economy being built by AI.
SKT’s CEO James Ryu is a visionary who truly understands AI and its massive opportunity, and has made really smart investments in companies like Lambda, Perplexity, and Anthropic. He has the unique ability to tell where the puck is going and is skating in that direction – paired with SKT’s understanding of infrastructure and decades of expertise, SKT is well-positioned to become a global AI powerhouse.
Q. I understand that you will be giving a keynote presentation at ‘SK AI Summit’. What kind of presentation will you cover?
In my keynote presentation, “Take Control of AI with Open Models and GPUs As-A-Service,” I’ll be exploring two key shifts in the AI landscape over the last few months:
1. How the landscape has evolved from a single proprietary AI, to many proprietary, to shorter and shorter cycles between quality Open Source models, along with what that means for enterprises.
2. How reasoning models are driving a shift in how compute usage improves the quality of model output. We’re shifting from improvements being driven by compute used at training time to compute used at both training and inference time.
I’ll also be participating in a panel discussion with SKT CEO James Ryu and Penguin Solutions CEO Mark Adams, where we will be diving into the present and future of AI infrastructure.
Q. The ‘SK AI Summit’ is a large-scale international AI event held in Korea. What are you expecting from this event?
Just look at the lineup of speakers – it’s a who’s who of researchers, engineers, and business leaders. I think that SK continues to position itself well to keep winning in the AI industry. It’s an awesome opportunity to meet other people across the world who are all just as passionate about AI.
