A New Era of Context Memory with Val Bercovici from WEKA

**Val Bercovici** (0:00)
100,000 tokens, one megabyte of data, translates to 50 gigabytes of KVCache working memory. And that's because you're vectorizing, you're adding 10 to 20,000 dimensions of intelligence to this 100,000 tokens, this one megabyte of data. So if one megabyte equals 50 gigabytes, and that's just at the beginning of one session for one user.

**Vikram Sekar** (0:28)
Welcome to the Semi Doped Podcast. I'm Vikram from VIX Newsletter on Substack, and with me is a special guest today, Val Bercovici, Chief AI Officer at Weka. Weka is a California based technology company that provides an AI native data platform designed to solve the massive data bottlenecks found in modern high-performance computing. Founded in 2013, the company is currently valued over $1.6 billion by focusing on the infrastructure needs of the AI era. In today's episode, we'll go over Weka's technologies, especially in the context era of AI, and what it means for the future of influence at scale. Thanks, Val, for being on the podcast with us. How are you doing?

**Val Bercovici** (1:11)
Really good. Excited to be here. Very enthusiastic new fan of the newsletter.

**Vikram Sekar** (1:17)
Great. Thank you so much for giving feedback on my article on context memory storage. It is a big area, and I learned a lot from all the articles on the Weka website, and so it was fantastic.
The whole thing I learned about context storage is really fascinating, and I think we'll get into some of that today with our listeners too and break down a lot of what's going on. How does that sound?

**Val Bercovici** (1:44)
Absolutely. My favorite topics. We should go. Let's go.

**Vikram Sekar** (1:47)
Awesome. Let's start with a brief background. You were a former CTO at NetApp. My Internet travels, I should be recall the Clouds are for leading the Cloud Storage SaaS Strategy. And now you're a Chief AI Officer at Weka. So what's your transition here? How did you get to working on storage solutions? Or is that something you've always been doing?

**Val Bercovici** (2:11)
Yes, and yes, and yes.
It feels like deja vu because I remember I left NetApp actually, joined a small Cloud native storage startup called Solid Fireout of Boulder, Colorado, and then got reacquired back basically into NetApp, surprisingly enough. And by then we were doing Cloud storage and that was after VMware, a golden era of enterprise storage and these NAS and SAN acronyms of the past. And I figured Kubernetes was really cool. I actually got involved with the Borg team at Kubernetes, Craig McLuckie and Sarah Novotny, Joe Beda and others, and we created the Cloud Native Compute Foundation under the Linux Foundation together. I got on the board, the first board, the governing board of the CNCF. So I thought around 2017, there was nothing left to do in storage and I left it. But it was a very disruptive time back then, 2012 to 2015 It's hard to remember that now, but Cloud was hyper-controversial, kind of as controversial as AI-eating software today. Back then, it was Cloud software eating the world. So it's deja vu in that it's just big transitions all over again. And at Weka, I joined again on the promise of actually not joining another storage company. And what I mean by that is that we were using the term data platform last year. And fundamentally, that confuses some people because it's an overload of term. There's so many things up and down the stack, which we'll get into. They can be called platforms. But fundamentally, what Weka provides is high-speed storage and memory for AI infrastructure, which ultimately is the key bottleneck we'll be diving into right now.

**Vikram Sekar** (3:47)
That's awesome. It's amazing that you're saying that in 2017, nobody really thought about storage solutions too much. It was considered boring and probably commodity. But fast forward, not even a decade. Things have turned around entirely. We have entire generative AI and people running agents all over the place. And we want to store infinite context. And the need for storage has completely changed within a short span of a decade. Who would have thought, right?

**Val Bercovici** (4:16)
Exactly, exactly. And yeah, it's, you know, it changed even three weeks ago, which we'll get into, you know, with Jensen's latest announcements at the CES keynote.

**Vikram Sekar** (4:26)
Yeah, let's get into that straight away then, because for those who missed that announcement, it's at CES 26, Jensen announced that they are introducing an inference context memory storage platform as a first-class citizen into the whole Rubin platform. And this is something that's going to be here for the future and is here to stay because we need a lot of context going forward, especially in the agentic AI era. But maybe we should just start with what context memory is and why we need it so badly today.