Feross on TBPN: Socket's Series C and the State of Software Supply Chain Security
Feross Aboukhadijeh joins TBPN to discuss Socket's $60M Series C, 500%+ ARR growth, AI's impact on open source, and the rise in supply chain attacks.
- Sarah Gooding

Socket CEO Feross Aboukhadijeh joined John Coogan and Jordi Hays on TBPN to discuss Socket's $60 million Series C led by Thrive Capital, the company's 500%+ ARR growth over the past 12 months, and why software supply chain security has moved to the top of the priority list at nearly every company.
The 10-minute conversation covers three forces converging right now: AI generating more third-party code than ever before, frontier models surfacing massive volumes of vulnerabilities across operating systems and open source libraries, and attackers increasingly targeting open source components as a way into thousands of organizations at once.
Feross also walks through what's coming for open source maintainers as AI-driven vulnerability discovery enables a future with far more disclosures than the ecosystem has ever had to absorb. When maintainers can't keep pace with patch review, vulnerable code sits exposed with no upgrade path for downstream users. Socket Certified Patches close that gap so teams don't have to wait on upstream fixes. He also covers the recent cluster of supply chain attacks that hit in a single day.
Watch the full conversation below.
TBPN Hosts: Up next, we have Feross from Socket back on the show to discuss a Series C by none other than Thrive Capital. We'll bring in Feross from Socket. Get that gong ready. Welcome to the show. How you doing? Feross: Good to see you guys. Thanks for having me back on. TBPN Hosts: I think last time you were on, we said we'd see you soon, and here you are. Give us the news first. What happened? Feross: Socket announced its $60 million Series C at a billion dollar valuation, led by Thrive Capital. It's a big day for us. TBPN Hosts: Absolutely massive. Is what you do important right now? Feross: You couldn't have really asked for a better perfect storm. You have the confluence of AI, cybersecurity, and all the different attacks we've been seeing, and we kind of built the perfect product for this moment. It's been something we've been building towards for a while. It seems like, okay, we just have the right thing right now, but the reality is we've been building towards this for four years plus. It's an overnight success that took four years. TBPN Hosts: Overnight success. I love it. What was the key unlock for the new round? Growth, needed the capital for specific expansion, just the broad tailwinds of the industry? What's changed in the business? Feross: The business is just inflecting. Every year I feel like, okay, this is great, certainly we're gonna get a regression to the mean or something — but it just keeps accelerating. Over the last 12-month period, we've had 500%+ ARR growth, and it keeps going. I think we're seeing three forces converging all at once. First, AI is generating more code than ever before. You've got developers, and increasingly non-developers too, pulling in open source dependencies and third-party code at unprecedented velocity. More code is being written than ever before, both by humans and AI agents. The maybe counterintuitive thing is it actually brings in a lot of third-party code, and that code is vetted less than it's ever been before. The second thing is the frontier AI models like Mythos are finding thousands of high-severity vulnerabilities across every major operating system and open source library. The total volume of vulnerabilities in code is exploding right now, and it's gonna keep going as we throw these models against this backlog of code that's been out there for a long time. The third piece is that attackers have started realizing they can exploit the software supply chain to get into companies. They're not coming in through finding vulnerabilities — they're realizing that if they go into an open source component, that's an easy way to get into an organization, and not just one, but usually thousands of organizations that all use that same component. It's a perfect storm of all these factors right now. TBPN Hosts: Walk me through how your business model interacts with the open source community. When we see things like Mythos finding bugs in open source repos, we hope everyone will submit bug reports early for the good of the world. But I'm interested in the economics of patching open source — if there's a really important package out there that maybe doesn't have a Linux Foundation behind it, at the very least there's inference costs. Who's paying to patch all of the open source software that so many different companies rely on? If you go to a big company, they might patch their dependencies, but how does this all flow together into an actually safer internet? Feross: I've been in open source for over 15 years now, and I maintain a bunch of open source packages and have a real soft spot for the open source community. Let me tell you, they were suffering under an enormous burden even before the current AI trends have made it even worse. They don't get much support, whether financial or just in terms of the number of people working on this really critical infrastructure — it's super low for how critical it is. We started seeing a lot of AI tools being used to create these slop PRs, these slop issues, and that was already causing a burden. And now you have the various frontier labs whose models are finding a lot of vulnerabilities that were always there, but the community didn't know about. I will say they're doing a good job of providing a patch along with the bug they've identified, so they're doing a lot of the work for the maintainer. But it still takes effort to review the PR. The thing people don't realize is when you accept a pull request as a maintainer, it's not this one-time cost. You're actually accepting a burden indefinitely into the future to maintain that code, make sure it's secure, and keep it up to date with all the other changes. So I think what we're gonna see, and we're already starting to see, is that despite the best efforts of the frontier labs — and they are making a good effort — maintainers are going to fail to actually accept these patches. We're gonna see a lot of libraries that have a vulnerability with code sitting on GitHub that anyone can look at and use to generate an exploit against that vulnerability. There's not gonna be an easy path for companies and developer teams to actually patch their libraries because there's no version they can upgrade to. The maintainer's just sort of sitting there on that patch and hasn't accepted it. It's creating this real risk where there's more vulns than ever before, and there's not really easy paths. This is actually something we try to solve. We built a solution to this called Certified Patches. It's basically a deterministic way to, in one click, make your vulnerabilities in your open source code go away. We use a whole bunch of AI to produce these patches that make the vuln go away without any work on the part of the developer. There's no burden of upgrading packages — in a click, the vuln is gone from your dependencies, and you can keep the rest of your code and your application the same. It's a quick fix to the problem that we hope is gonna be part of the solution. And by the way, we're giving away all the critical patches to the community for free, so we can disseminate this widely to as many people as possible. TBPN Hosts: Are you finding any companies that are completely asleep at the wheel? Or is everyone giving this the necessary attention? I imagine part of why the business is growing so quickly is this is a hair-on-fire problem every single day. You're seeing vulnerabilities or issues pop up. I'd have to imagine every software company online where the technical leadership uses the internet is leaning into this problem. But from an outbound sales and marketing standpoint, are there people who are like, "Ah, this just isn't that important for us right now"? Or would a no be, "We have a solution, it's with someone else"? Feross: Okay, so we've met one company that has literally no software. They produce toilet paper. And they didn't have any — no smartphone, no business software. I think it's become a universal problem. When we started, it was a more niche problem where people said, "I've got a lot of other problems that are more important than this." We initially had a really good uptake of the product in the cryptocurrency community because they have a lot of software supply chain risk, and when these hacks happen, it's an irreversible thing for them — they lose all their funds. Then I'd say the AI labs were the next to pick it up, and SF tech companies — the people who are ahead of the curve. Now it's become a top one or top two concern at nearly every company we talk to. It's a board-level concern a lot of the time now, and CISOs and heads of engineering are being asked to figure out, what is our solution to make sure the next time one of these things happens, we're protected? And the next time is probably gonna be tomorrow. It's literally — yeah, every day. Literally, our end of quarter was the end of last month. We were already super busy, and then there were three major software supply chain attacks that happened that same day. It's wild. It's more than we've ever seen, and it's totally unprecedented. TBPN Hosts: Was the GitHub issue that I saw a supply chain attack, or was that something else? I saw a couple posts but didn't get to dig into it. Feross: I don't think they've released that information yet. There's definitely been some speculation. The timing right now — every time a company gets popped or their source code leaks, it's the first question people think of now. The group I believe claimed responsibility is this Team PCP group, which has been responsible for a lot of the attacks. I wouldn't be surprised if that's what we learn, but I don't think that's something we know right now, and I wouldn't want to speculate. TBPN Hosts: I wonder if there will be movements to find the compute infrastructure for Team PCP at some point, or the hackers themselves, because at a certain point you'd imagine they're — Feross: A lot of this is still social engineering, right? So they don't necessarily need a country of geniuses' data center. They can just — TBPN Hosts: A couple dudes with phones. Who knows? Well, congratulations on the progress. Incredible progress. Thank you so much for coming on the show. Thanks for keeping us all safe. We'll talk to you soon. Feross: Thanks, guys. Have a good one. Appreciate it.




