Anti Metal raises $20M Series A to automate cloud infrastructure as AI makes deployment the new bottleneck

that different what you would find in a large campus because it all has to flow together. The packets must flow end to end. Wait, what's the difference between nonGPU data centers and GPU data centers in this context? Yeah.

So, so yeah, in GPU stuff, there's actually a lot that happens between two GPUs and between clustered GPUs. Got it. That's a different type of networking. There's a lot of work that's happening.

Obviously, Nvidia with their Melanox four years ago and what's happening now with ultra Ethernet, but just everything else is actually relatively same. What was the what was the key word you just said? Ultra. Ultra Ethernet. Ultra Ethernet is a new thing that's happening.

Is this I remember Cat 5, Cat 6, I've heard about Cat 7. Is this like cat 8? What are we talking here? This is like a whole different ballgame. So uh if you if you look at you know uh the joke in the networking world is uh you know uh Ethernet is dead, long live Ethernet.

Uh you know every sort of 15 20 years uh Ethernet is predicted that it will be the end of it. uh but you know somebody comes up with innovations on how it can scale up on speed on throughput on bandwidth on security.

So ultra ethernet is this new version that's coming out for you know next phase of all the data center buildout that's happening because you know roughly the US has more data centers than the rest of the world combined. Wow roughly let's go here for America that's a white pill right there.

Everyone says we're falling behind but not today. But the rest of the world is trying to catch up to get to that and the US is trying to 3x where we are. So there's a lot we're undefeated. We support meters international strategy. So it's kind of nuanced. We're pro but also anti as long as we're benefiting.

Anyway, yeah, there's a lot that goes into it. Yeah. Yeah. Yeah, that makes sense. I mean, we we've experienced it here.

there's uh like new generations of just streaming cameras that do power and video over Ethernet and and you and you know we're thinking about kind of the V2 of the studio and there's a ton of insane things that you can do do over Ethernet. It's no longer just LAN cable and so must be fun to be at the heart of that.

Uh Jordy, any more questions or should we let them get back to building? Congratulations on the the funding milestone. Congratulations and uh yeah, check in next time. We'll we'll check back anytime. Thank you both. See you soon. Bye. Cheers. Uh our next guest will be joining in just a few minutes.

Uh in the meantime, it is an absolute banger day. Lots of fun. I know. But I was just going to say it's actually insane. We've been able to keep the show up with every the entire It is crazy. We Yeah, we got lucky today. Uh did did have we ever has the stream ever actually gone fully down?

We had Zoom go down one day and we were kind of kind of kind of wrecked by that. Um well uh let's bring in our next guest. Ian in the chat says uh Mix Panel, Shopify, and GCP are still down for him. All the services that he's in. Honestly, if you're worried about the show going down, do not worry.

Just send us your mailing address. We will print the show onto vinyl and we'll send you a record of the show so you can listen to it. Same day shipping. Exactly. So that you can get the show almost live, you know, a couple hours. Couple hours later. Yeah.

But um yeah, we'll also do a lot of innovation on the vinyl pressing because I think that takes a while too. Anyway, our next guest is here. Welcome to the stream. How you doing, Trish? What's going on? Good to see you. How are you doing? Good to see you guys. Thanks for having me on. It's great to have you.

It's great to have you. Uh big day. You guys announced a series A, new website, and you took down the whole internet. You took down the whole internet. I can't confirm or deny it was part of the launch strategy, but good timing was. Uh, give us the stats. How much you raised? Yeah, raised $20 million. There we go.

Congratulations. Who did you raise it from? Yeah, sound led the round participation. Hit it again for sound. Hit it again. Congratulations. Um, amazing. Sorry. Continue. We got a little excited. That was a good gong. Yeah.

participation from Buckley, Nabal, Arj Vowsi, founder of Dropbox, Arvin from Perplexity, bunch of other awesome angels. Yeah. Very cool. Very cool. Uh, break down the pitch. Uh, like like how are you positioning the company right now because I know you've been doing a lot of different stuff.

They pivoted from pizza to something. I know them as the pizza delivery company. Uh, there there was that Zoom company that was doing pizza on demand. They that didn't go so well. So, obviously a huge white space. Well, they just did pizza for a day. they sold out and then they moved on to the next thing.

So give us an update on on the platform. I know how how are you positioning it? For sure.

So I think for us always the mission of anti-metal has been how do we simplify infrastructure but current moment what's really different is you know software engineering the bottleneck used to always be programming like if you could build a software you could ship it you could distribute to everyone suddenly your product is live now you have all this AI assisted stuff like really awesome products like you know cursor lovable vote v0ero and increasingly like the latter half of that equation which is deploying that code maintaining it making sure it scales like going from one user to 1 million is increasingly the bottleneck on like how companies, you know, are actually going to grow and sustain.

Um, like you know, the short of it is like writing code is no longer the hard part, right? It's all this operations and and maintenance. And so what we're really indexing on is building this product that helps you better manage, you know, and automate infrastructure.

So, a lot of that is, you know, the nuance of infrastructure versus code is a lot of it is knowledge that is baked into, you know, people's heads. Like you have that one senior engineer who just knows how everything works. that's your go-to guy when something breaks or you need to spin up something.

Uh, but that's not super sustainable, right? And this like notion that comprehension can just scale linearly as you hire more hire more people isn't super tenable anymore. We're like emitting a bunch of data. All that data is super fragmented. Everyone's super specialized.

And so for us what we're trying to do is you know start going into companies plugging into their entire infra surface area whether it's like source code their cloud their ticketing system whatever it might be really understand why things are failing um and basically pull in data and give them these golden pathways as to what's going on why is it happening what prescriptive action should I take that is going to help me fix this and along the way sort of learning what are the ingrained practices your organization uh that part is super important in the sense that we don't want to build like an opinionated platform from says like this is good or this is bad, right?

That's pretty subjective to each company. So, it's a little more about learning, okay, given this tenant, what's going to be good for them? How do they prefer to operate? You know, what's going to make their systems perform the best? Makes a ton of sense.

How much how much are you spending time thinking about from a product standpoint? trying to solve things that are sort of human nature to some degree versus like it feels like that the nature of software development is changing so quickly.

But if you can kind of focus in on on what will always be true and and you know many people would say that's human nature, you can create something like durable and valuable out of that. Yeah.

I think a big thing for us is, you know, some companies try to position this like, hey, we're a full-time employee and we're going to like automate this functional way for you.

I think for us it's a little more of how do we take all this knowledge that exists in a bunch of people's heads, institutionalize it into a system, encode it into a system such that it's reusable.

Like the dream for us as companies go from asking this question of like yo, can we keep this running to what can we build next and really taking away this time for maintenance uh that a lot of people are spending their time on.

So you know I think human nature like inevitably all info problems are kind of human problems like you know someone messed up a config or someone didn't really understand what's going on.

Uh, and in that sense, you know, working with our users to understand like what's their preferred behavior, why do they want their stack set up this way, why have they chosen the tools they did, um, and not trying to like force an opinion down their throat. That makes sense.

Uh c can you talk to us about um some of the newest problems that companies are facing with LLM inference costs just faults that come from this new paradigm of running agents and running tons of inference and and just taking advantage of all the tools.

Uh we saw Sacha Nadella at Microsoft Build talk about model routing being something that he was really pushing at Azure. GCP has a similar tool. Um, are you hearing from customers?

Like right now we're in this era where uh every new model release and price drop basically goes viral on X and so unless you're under a rock like it's pretty easy to move over to the latest and greatest thing, not leave a bunch of money on the table, but are there new problems that you foresee as things become a little more commoditized or a little bit more calcified that companies aren't using best practices or they're running into new problems from the kind of AI age generally?

Yeah, I mean I think the first one is like what you struck on the head which is all of this AI stuff is super nent. That means there's not a ton of like super robust tooling out there how to monitor evaluate these things. You know people are building their own frameworks on the fly.

I think like most companies that are really succeeding are not using something off the shelf. They're building it in house right so there's a lot of nuance that goes into these things. The other thing is you know models are another potentially single point of failure.

Like if cloud goes down like what happens to your product? Are you ready to do model routing or are you going to do something else?

uh but I think like you know putting them into like a broader system perspective they're not like new problems in the sense of these are things systems engineers or you know data architects have had to think about for a long time but obviously more and more companies are doing these things and they're really complex systems I I think the second part of that is actually just like you know we don't index on this super highly but something we're seeing is more companies are just pushing more software like highly verticalized niche to their internals like you know why go put something on a retool dashboard when I can like build it and bolt and probably get a more nuance tailored experience So I think that coupling of like hey we're working on a pretty nent space that there isn't a ton of structure around and we're trying to figure out at the same time combined with we just have more infra and more services running period is creating this like really chaotic environment where people are you know stuck in the trenches of like debugging things figuring out how do we scale them uh paying down tech dev sort of thing.

Is there anyone on the hyperscaler side that you think is like really underrated right now? Yeah. Um, honestly, OCI, I think, uh, Oracle Cloud, they're never part of the conversation. I think all these other clouds have really great offerings, really great services.

I think the thing about Oracle is, you know, they don't sell a bunch of services on top. They have a really good foundation. Like, these guys have been in the computing business for quite a while.

I don't think it's the sexiest option out there, but like from a price perspective, security perspective, I think they're pushing a lot internally on what that's going to look like over the next couple of years. Um, but yeah, I think underrated. That's amazing. the uh uh stock hit the all-time high today.

Let's hear it out for alltime highs and generational run. Fantastic. What's your guys I'm just curious since you guys are so deep in the weeds on this. Uh what's your what's your personal uh AI, you know, codegen stack right now? What are what are you and the team loving? Uh right now we have cursor for the entire team.

Uh some people preferred claude code depends you know different strokes for different folks. Um, but yeah, I mean like people are using chat GBT, claude, cursor. Cursor is probably by and far the biggest one.

We're finding a lot of success with some of the agentic stuff like cloud code, but I think the median for us is people are loving cursor. That's probably been the best coding experience. Very cool. mostly tennis.

Uh what about just go to market um top of funnel like you're it's an interesting super interesting company because you've been able to go like massively viral with what is an enterprise like software company like it would be different if you were selling like an energy drink.

I would kind of expect like a stunt and oh yeah everyone's talking about it but uh you've been able to break through in a really interesting way in in in the ex community in the venture capital community. Um, what are you seeing that's working on the top of funnel side? What viral loops are you taking advantage of?

Flywheels, just break down the marketing side. Yeah, I mean like you know growth is a big part of our DNA. I think largely from you know my co-founder Mattly I like a beast at growth and marketing.

Um, you know for us uh a lot of it at least in the early stage right now is you want to grab people's attention obviously attention economy. I think the the really nice thing for us is this is a problem like every engineer we talk to has faced or seen to some degree.

Uh even people on the operational side when they're like running a business, they're like, "Holy crap, like we couldn't push this feature out because we were stuck like fixing the service or paying down tech debt instead of launching the new thing. " Um but yeah, I mean we're doing all sorts of things.

I think candidly you're going to see a bunch of it over the coming months. Um we're going to have a you know bigger splash later this year, but yeah, more to come. Can't wait. Cool. Amazing. Well, you'll have to hop back on when that bigger splash happens. 100%. Uh, great talking to you.

We'll uh congratulate Great to Great to get the full update. I think we're both on the on the cap. Yeah, we'll talk soon. Love it. Awesome, dude. Thanks, guys. Talk soon. Have a good one. Uh, let's close out by telling you about Figma. Figma. com. Think bigger, build faster.

Figma helps design and development teams build great products together. You can get started for free. I have a challenge for the audience. your uh your personal website is probably bad. Go on Figma, make a new one, and just ship a website. Do it in in like five minutes. Fantastic.

Uh speaking of design, Signal's been on an absolute tear breaking down all the WWDC updates. Um uh he says, "Two quick observation observations after using iOS 26 and iPad OS 26 for a few days. iOS 26 feels great. The system is fluid. Buttons are larger. Everything responds with intention.

It's super enjoyable and fresh. iPad OS 26 is a breakthrough. Productivity finally feels native. File handling, window management, multitasking are intuitive and fast. The most fluid and powerful computing experience on the planet for most tasks. He says he does everything except code on an iPad.

And so I've I've every time there's a new update, I'm always like, "Oh, maybe I should pick up an iPad. Maybe I should. " I'm tempted. This might be the one that does it for me. It's a bad point in the release cycle to buy an iPad Pro because it's over 400 days since they launched the last one.

So, they'll probably refresh it at the next iPhone event. But, uh I'm I'm I'm definitely thinking about getting it uh on here. I think it could work really well for running the show. Got a couple iPads around the house that just kind of collect dust.

Collect a lot of dust, but maybe I'll upgrade and maybe you can get back in. I think you got to commit to that being How is uh your your uh iOS 26 experience so far? Have you run into any more problems? Uh, it's pretty good so far.

I mean, it's still, you know, my phone is like 6 years old, so it's a little bit slow, but overall it's been totally fine. I like it a lot. Yeah. Getting a six sixy old iPhone. I was asking him I was I was asking him like, "Show me it.

" And he starts swiping through his apps and one and on one screen he just has like 25 copies of What app was that? Uh, Chick-fil-A. Chick-fil-A. He He somehow got the Chick-fil-A. the one app. You're only supposed to be able to install the app once. Somehow he got 26 copies of it installed. You got to have 26.

If you're serious about If you're serious about Chick-fil-A, you should have at least 26. 26 copies just in case 24 of them break, I guess. Uh give us the update on the art history project. How's it going? Are you an expert yet?

Uh, by the way, my wife texted me and said that uh it's it's extremely embarrassing that we didn't know that Greek statues were painted because apparently that's common knowledge, but also she has a she's an art history. Okay. The art history major is common knowledge if you're if you're if you're a renowned expert.

Yeah. Yeah. Anyway, g give us the latest art history fact. How are you doing? Yeah. So, so I I took another quiz, another like assessment. Um, this time I got 50%. Oh, very good. So, that's 10% in only I think like an hour and a half. compound that you'll be at 50% more than 25% improvement.

Honestly, give yourself some credit. Yeah. 25% improvement. Yeah. Yeah. But, you know, if I get 10% every hour and a half, that's, you know, math checks out. Extrapolate a bit. I think it's like 7 and 1/2 hours. I should be at 100%. Fantastic. So, I should have do it again. You could be at 200%. 200%, 300%.

You could be the world expert. Uh, did you learn anything else interesting about art history? Do you have any more facts to share with us or are you tapped out? I've been reading a lot about this guy, um, Giorgio Vasari. Okay. So he's kind of he's 16th century.

He's kind of a polymath um author, historian, painter, sculptor. He's kind of like the uh the Plutarch of of art, right? So Plutarch writes uh noble lives of uh or parallel lives of noble Greeks and Romans, right? Um and then Visari kind of writes the same thing for uh kind of Renaissance art.

Um but kind of this controversial guy. He he writes about, you know, he he looks at art from the perspective of the artist.

um which is you know controversial in the sense that like later on this guy um Winkkelman I think okay he he criticizes Visari he says no you should you know separate the art from the artist you should look at art kind of in and of itself by itself um but yeah a lot that's where it came from I've heard the term separate the art Chris from Pace would just say absolutely not we're not going to separate it no we're going to look at it the artist is if we were to issue you a termination letter today would you feel more confident now that you've at least studied art.

Well, I I actually your job prospects. I mean, yeah. So, so so I I also looked at those other stats about just general employment and you know, so I'm actually a physics major. I'm not CS. Sure. Um and physics is actually the second most unemployed. So So they're at uh physics majors are at I think 7. 8%. CS was at 6.

1. So um I think I think I probably would go into art history. I mean I I don't got much else. So amazing. Underrated. underrated. I think we just discovered some massive alpha. It wouldn't really take that much to double the number of art history majors. You know, there's only 2,000 of them, right?

Or something like that. Yeah. Well, 2001 now. 2001. 2001. Um, amazing. Let's rip through some timeline. Great work, Tyler. Yep. We have some we have some news. Uh, congratulations to Crusoe. Uh, Chase Lock Miller been on the show.

Uh they are partnering with Brookfield on a $750 million debt facility to build the intelligent infrastructure of the future. Do we have to ring the gong? I think hit it, John. Hit it. That's a lot of That's a lot of pesos. That's a lot of pennies. Uh congratulations to the whole team over there. Fantastic.

We love big data centers. We love uh scaling laws, putting big training runs on the data center. Fantastic. We got another post from Z. What are some good baby names for someone just getting into nominative determinism? I got to say it came to me immediately.

Oh, I guess it's here as a comment, too, but I was going to say if it's a boy, Chad. Uh, it's just so obvious. Um, Chad, I think, is going to come back in a big way. I think it it sort of waned in popularity when it was more of a slur, but now now it's just massive alpha there.

It can kind of be an inside joke for you and the other dads. 20 years from now, your son, you know, little baby Chad turns into a man. Oh. Oh, he's a Chad. Huge surprise. You know, it was, you know, uh Johnny IV was in the news because he went to OpenAI.

Uh but he's technically Sir Johnny IV because he's been kned by by the uh by the the British royalty, right? And I was thinking we need an American version of sir of knighting.

And I think what we should do is we should take all the Chads and the ones who are really really impactful, the ones who have driven, you know, massive capital investment like Chad Buyers, they should be knited in an American way and they should earn the title Giga.

And so and so for his contributions, the president should take a sword and tap him on both shoulders and knight him. Giga Chad Buyers. Giga Chad Buers. I think I I I think that would really really it would be the American version of the nighting. Yeah. You know, and it's based purely on how much capital formation 100%.

There's no other criteria. That's that's the way it needs to go. I mean, so step one, name your son, Chad. Step two, get him into capital allocation on the private side ideally, but public markets will work too. You're still eligible. There's some more good comments in here. Grit Capital says King or Don. Donald.

Yeah, Don. Don. Donald. I never never thought about the dawn. Yeah, that's where dawn comes from. Yeah, I think so. Uh, in other news, Anan Swal says, "My first and only angel investment went to zero today. " Well, founder of stake. Yeah, should have made 50. I'm sure one of them would have worked out.

Made it all back in one trade. fascinating because he's the founder of CB Insights, which you would think he would be addicted to angel investing with all that data and stuff, but he knew his life's work was to build that company and that's what he focused on and he only invested in one company. Fascinating.

Very, very dealership guy responds, "Congrats. " Very funny. Uh anyway. Yeah. Yeah. I mean, he said he learned, "I don't really enjoy. " And if you don't enjoy it, you're not going to be good at it, so stay out of the game. Uh Paula says, "Diet Cokes or fridge cigarettes? " So true. Silly. So true.

Just drink them all day long. All right, this is a good place to end. I like how Zuck has been back against the wall like 37 times since 2007 and every times resoundingly wins, but each times he he has a crisis. People are like, "Oh man, Zuck is cooked. " Yep. So true. Yeah.

Do not count the AI talent crisis is is merely a bump in the road for for old Zuck. He's he's cooking. We've all We've all been there. We've all, you know, we've all tried. Yeah. We've all been running a trillion dollar company, a personal behemoth, you know. Yes. Yes.

We all have something you need to wrangle and get out. Gotta power. Whether it's a 4 hour stream or or a or a two trillion parameter model, you know, you got to get it out. Well, this was a great show. I got to give it up to the to the incredible production team while everything goes down internet.

We learned everything there is to know about art. We're experts, too. He's an expert. Yeah. Um, now we got to worry about poaching because I imagine a lot of the galleries are going to be watching this and they're going to be, "Oh, new art history just dropped. We got a picture. " Understands physics. Yep.

Live streaming and computer science. Knows everything there is to know about. He knows everything. The history of art. Anyway, thank you for watching. Leave us five stars on Apple Podcast and Spotify.