Latest / Oxide and Friends / The Oxide Supply Chain

Transcript

The Oxide Supply Chain · Oxide and Friends

• 0:00
  Speaker 1: Kate, it is great to have you here. Really, really excited about this. I thought it was exciting. I don't know if you saw Seth's tweet, but, our Seth, the partner from bank from from Eclipse Ventures that that is on our board. Man, when he saw that you were here, he was very excited about it.
• 0:17
  So, obviously, he's, appreciated your wisdom over the years as we all have. I thought you might start by talking about how we first met. Because, Adam, are you in the office when Kate first came by? And this would have been like, Kate, was this like January or February of 2020?
• 0:35
  Speaker 2: I think it was January 2020 before all the crazy lockdowns. And I remember meeting you, Steve, Jess, and Robert Nastocki specifically. But I don't remember meeting Adam.
• 0:48
  Speaker 3: I'm sorry.
• 0:49
  Speaker 4: I think this might have been I know this might have been, like, moments before I joined because Yeah. I joined sort of towards the
• 0:54
  Speaker 1: end of the day. Right. Okay. So this is and and it was kinda and Kate was obviously, like, super impressed. Kate, your background is in industrial engineering and, had spent time at a defense contractor and then spent a ton of time at Apple, but it was not meant to be because you were moving back home to Arizona.
• 1:12
  Right? You were it was like Yeah.
• 1:15
  Speaker 2: Well, first off, thanks for having me today. I'm really excited to be here. But, yeah, I'd love to tell the story of how I kind of ended up at Oxide. And I think it all started with that trip in January out to the Bay Area. I was actually on a business trip at the time and, stopped by for dinner with Jess and got to meet you guys in person.
• 1:33
  And at the time, we were really looking to move back to Arizona, which we ended up doing, in April of 2020. And so I thought, oh, well, oxide is really Bay Area only, and it's just not gonna work out from a timing perspective, in terms of geographies. But then, as we all know, remote work became a real Remote work. And it really opened a lot of doors.
• 1:54
  Speaker 1: It did open a lot of doors. And so we would and we knew that we were gonna be hybrid, local, remote. But and I would just I just feel so embarrassed about some of the, like, some of the ridiculous notions we had. I had this notion that, like, oh, operations needs to be based in the Bay Area. Like, I'm not sure, like and I think I I I would say that was, like, jointly shared.
• 2:16
  It wasn't just me with that idea, but I don't know. It felt like something like, oh, you've gotta be in the office for this. I don't know why.
• 2:22
  Speaker 4: No. That that made a ton of sense. It it really felt like the right thing even though it's turned out not. But I just wanna pause for a second to just note that Kate is the first guest to ever say she's been happy to meet her. I just want that to not escape our attention.
• 2:34
  Speaker 2: The record.
• 2:36
  Speaker 1: Right. As opposed to no. We so you're saying that it would really be more natural for folks if we started off with a complaint about Twitter spaces. We haven't complained about Twitter spaces at all.
• 2:43
  Speaker 4: That'd be more familiar.
• 2:44
  Speaker 1: Am I in the right space? We're talking about stolen cars, not how Twitter spaces is a disappointment. I so but, Kate, you're moving back to Arizona, and I remember thinking, like, oh, Kate's great, but it's just not to be. It's the it's it's not gonna be. And then, yeah, as you say, the pandemic hit.
• 3:00
  As it turns out, like, the gig in Arizona, maybe not the best.
• 3:05
  Speaker 2: Yeah. That didn't work out, but the gig that brought me back to Arizona turned out to be a business unit that was being sold off just a couple months later. And so I moved back to Arizona and started talking to Oxide about opportunities and turns out a rec went up that I happened to see for a technical program manager, and it seemed to fit a lot of my background and skill set. And I got really excited about that, especially the opportunity to kind of wear multiple hats, which has turned out to be really true with Oxeye.
• 3:33
  Speaker 1: Well, I remember that. Right? I remember you saying that, like, one of the things that frustrated you previously in your career is just getting siloed and and and seeing, like, the right answer but not being able to affect the right answer because you're in the, like, nope. That's the responsibility of a different organization, which is always a super frustrating feeling. And, I mean so, you know, Adam, obviously, you were on board, but Kate came on board because that was a couple months later.
• 3:58
  And Kate, I I remember, like, just early after you arrived, just in terms of I I mean, this I such a I was so ignorant of all the supply chain issues. I remember you you the the one of the first things you said is, like, we need to figure out what we're doing for long lead time parts. And I'm like, Kate, we don't know what we're building. We've got
• 4:20
  Speaker 2: Yeah. I remember distinctly coming in and having 2 big questions for the engineering team that was already there, which was a small, scrappy team at the time and it was, what is our roadmap? Like, what's the project schedule? Do we have a Gantt chart?
• 4:33
  Speaker 3: Do we
• 4:33
  Speaker 2: have dependencies? And where are long lead parts and where are we out in the build materials to start procuring? Because when I walked in, it was we're gonna ship 18 months later and that was the reality. I was like, if we're gonna do this, let's figure these two things out.
• 4:47
  Speaker 1: Yes. And and I would say 3 years later, we've done one of those two things. We've we have dealt with unequivocally. No. We don't we're shipping.
• 4:55
  But the, so it you were asking us about long lead time parts. I remember thinking like, that is that's we that's nuts. We don't know what we're shipping, but you were as we started digging in and you were starting to dig in on, like, alright. Look. Like, we don't know everything shipping, but can we make we can make some decisions.
• 5:12
  And Yep. I remember and this is when did the supply chain crunch really hit? Because this is definitely before the supply chain crunch. This is
• 5:20
  Speaker 2: Yeah. So I joined in June of 2020, and I think we were starting to feel some early inklings, you know, like toilet paper shortages and stuff like that, but it hadn't reached the tech sphere completely yet. We weren't in this crazy escalation path of on allocation for almost every part with 52 plus week lead times. We weren't quite there where the chips, shortage was top of the national media headlines. I think that really started to hit harder in q4, which actually is when we started procuring.
• 5:51
  So I joined in June. We kind of started getting to determination on some of the key components, not an entire bill of materials, but the really strategic parts that we couldn't live without. And, we started procuring in November of 2020 on those strategic parts like the Intel, AMD, Chelsea, Murata, like some of those key parts.
• 6:12
  Speaker 1: And I feel like the fans were early in there. Right? I think that we the San Yudanki fan Yes. I feel like we're San Yudanki. And we this is one of those where I it's like a total wake up call again where, you know, you're coming in about long lead time parts.
• 6:23
  I'm like,
• 6:24
  Speaker 3: really?
• 6:24
  Speaker 1: It's, like, really an issue. And then you we get to some of these lead times, and, again, before supply chain crimes, before this is top of mind for us visivally, and getting, like, 52 week lead times on fans. You're like, wait. What? And I I think I was like most people in that, like, I just didn't really think about the complexity of getting things from a manufacturer.
• 6:49
  The the complexity of manufacturing something and in pulling together all of its constituent parts. And ultimately, it's like endemic complexities of this and why something that feels simple, like a San Diego fan, is actually not simple at all. Like, why are these e times so long even in good times?
• 7:14
  Speaker 2: Yeah. So a lot of these, key suppliers, especially the ones that are building high quality parts, are working all the way up their supply chain to the lowest node. And so they're doing quality audits of their subcomponent suppliers, and they're really working on forecasting and building material pretty far out, which leads to some of these long lead times. Not typically 52 weeks. That would be in this constrained environment, but it's not unusual in this industry to see, you know, 6 plus month lead time on small parts because of that complexity of where their sub components come from and how many BOM items in the bill of materials.
• 7:52
  BOM is bill of materials. I guess maybe not everyone
• 7:56
  Speaker 3: in our audience knows that.
• 7:58
  Speaker 2: Where all their bomb items come from.
• 8:00
  Speaker 1: And they're not always transparent about where they come from. Right? I mean, it's sometimes it can be, like, you're getting quoted some outrageous lead time, and you really gotta take it apart to figure out what is actually causing this lead time.
• 8:09
  Speaker 2: Yeah. So San Diego Danke fans is actually a really good example of that. So they had, a pretty complex bill of materials, and we knew this was a strategic part. And we had already started designing that as one of our key components. We've done thermal simulations, with our mechanical engineering team that we're using at Benchmark Electronics, and we had started realizing that, wow, the system's pretty hot.
• 8:34
  We we need to take cooling very seriously, and the Senor Denki fans, fit what we need in terms of capability and the cost and the quality, most importantly, quality and capability. And we had started working with the Sanyard Denki engineering team because we were going to make kind of a semi custom, take a COTS fan and make some modifications specific to the oxide design. And so I
• 8:57
  Speaker 1: do talk about that modification because that modification is very near and dear to my heart.
• 9:02
  Speaker 2: Yeah. We, basically changed the PWM of the fan so that it doesn't sound like a jet engine taking off. So we reduced the PWM so that it, is a little bit slower and that allows us to meet our meet our thermal model, but also not have crazy amount of decibels of audio.
• 9:20
  Speaker 1: Yeah. And so just to to to kind of define some of these terms here. So PWM is pulse width mod modulation, and the this is kind of like your frequency of high versus low signals that indicates, like, how fast it's spin a motor, you know, the the the more on versus off that you have, the faster it will go. And 0% PWM is, like, all off. How fast does it go at all off?
• 9:44
  And the the the fans, San Diego fans, as quoted, have are 5,000 RPM fans when they are so, like, you can't go slower than 5,000 RPM. And, Adam, have you have you been around the 5,000 RPM fan?
• 10:01
  Speaker 4: No. Not not in not in the flesh. I've only heard the Oh, really?
• 10:04
  Speaker 2: It turns
• 10:04
  Speaker 1: Oh, yeah.
• 10:05
  Speaker 2: Mhmm. It turns out if you stand next to it with your Apple Watch on, it'll tell you that you're, like, gonna hurt your eardrums from
• 10:11
  Speaker 1: the front ear. And Yeah. But, Adam, you recall when I was wounded by one of the fans. Do you do you remember this?
• 10:20
  Speaker 4: No. Oh oh, it like took off part of your finger.
• 10:22
  Speaker 1: It did take off part of my finger. This is the way I I you know, maybe I I I know you should be sticking to software, but the, I was taking one of the fan. So we are using 80 millimeter fans, which is important because these are bigger fans that can actually are much more efficient. So the ability to go slower than 5,000 RPMs actually would be a feature. And the actually, to go ahead of it, the the feature that we got, the the the modification that we have, the Sandy, Becky fan, is the ability to go at 2 k RPM versus 5 k RPM at 0% PWM, which makes a huge difference.
• 10:57
  But I was testing this fan or the the a different fan but very similar at 5,000 RPM and then taking it all the way up to 12,000 RPM. And when I took that to 12,000 RPM, I hadn't really secured the fan very well, and it took off. Like, the fan was, like, was going across the bench. And I went and I'm like, this is one of these things where, like, left arm and right arm had 2 different thoughts simultaneously. Left arm's thought, which is really the much better thought, is cut the power to this thing.
• 11:26
  Right arm's thought was, prevent the fan from moving across the bench. And the the, the problem is that the right arm beat the left arm, and I caught the fan, the blades went right into my fingertip. And not only did they slice off a bit of my fingertip, but then there was this, circle of blood on the wall in my, like, home lab.
• 11:56
  Speaker 2: It it's It's just outside artwork.
• 11:58
  Speaker 1: It's outside artwork. I definitely have, like, left it there. I'm like, no. This is like I have I have bled for this for for the for for the stand modification. So, we definitely want and so it they are loud.
• 12:10
  And we have 6 of them. Right? So at 5,000 RPM, they are super loud, and I have always found it to be just grating when you plug in a a a Dell server, Supermicro server, and the first thing it does is go to full fan blast. I just feel it's like, oh, come on. We gotta do better
• 12:31
  Speaker 3: than that. So that's why we wanted to get to 2 k r. Yeah. Well That's
• 12:32
  Speaker 5: Ryan, recall that we we also figured out the air volume that would be moved when you turn up all 6 fans at 200%. And we actually decided try this out in the office by putting one of the servers on the ground on top of ping pong balls and then turning all the fans on. And in fact, it will propel itself.
• 12:53
  Speaker 1: Propel itself. You just need a bus bar long enough. You can do or you need power, and you can actually move. So yeah. Now these fans could definitely move a lot of air.
• 13:03
  This be it
• 13:04
  Speaker 2: Yeah. So back to the Yeah.
• 13:06
  Speaker 1: The first
• 13:06
  Speaker 3: thing I
• 13:06
  Speaker 2: was saying, you know, Dinky, where we're headed was, we had to dig down into their bill of materials and figure out exactly which component that made up the fans they were short because they were telling us they weren't gonna be able to deliver in time for our, EBT engineering build, which was pretty critical path on our schedule, to get through the engineering build cycles. And so we figured out what specific component in their BOM was short. We went around the supply chain to see if we could find inventory of that through any distribution channels or any strategic partners. And it turns out we found a bunch of these on the broker market, and we've seen that a lot over the past year and a half as supply chain tightens. Like, brokers are swooping up any inventory that comes into stock at the distys.
• 13:50
  Speaker 4: Okay. Could
• 13:50
  Speaker 1: you describe the broker a little? Because this is basically, like, at first, I'd tell, like, the supply chain mafia. It's the way it is. I mean, what is the broker market?
• 13:59
  Speaker 2: Yeah. Play at your own risk. So it's basically parties, either entire companies, which are more legit, that's actually the broker market that we like, that we went through, and I can elaborate on that, that are buying parts through distribution channels that are hot commodities that are on allocation, very constrained, and then they're marking them up and reselling them to people that need them. And they're holding some inventory there, And then there's the sketchy brokers that is just someone in a loading dock where they took a box off a truck or something like that. Or
• 14:44
  Speaker 4: So it varies from, like, the the folks who are trying to, like, force down arbitrage, like, jack up prices, and then the folks sketchier than that. Right. Like that like, that's the range.
• 14:56
  Speaker 2: Yes. The ones where, like, where this truck go, where this box go off the back of this truck, or why does this part number look slightly different?
• 15:03
  Speaker 1: Oh, because
• 15:03
  Speaker 2: right now it's
• 15:05
  Speaker 1: customers. Right.
• 15:05
  Speaker 4: Yeah. And and
• 15:06
  Speaker 1: what and I was, like, the the the issue counterfeit parts is a real issue. I think, Rick, you definitely opened my eyes this early. It's unbelievable all the stuff that's counterfeit that's out there. And, Kate, I mean, we must be be quite very carefully to obviously, we wanna avoid counterfeit parts. But counterfeit parts are a real problem.
• 15:23
  Is that right?
• 15:23
  Speaker 2: Oh, it's a real problem. Yeah. And I think that's why we've been very, very careful about using any broker supply. There's only 2 examples I can think of where we actually went forward with Broker Supply, and that was because it was with a trusted broker partner. Actually, they're called Fusion.
• 15:39
  And, they were on a short list of good brokers. 1 of our team members, Kirsten Neera, who's a supply chain badass and has a long history at Lenovo, had actually worked with them in the past. And so she kind of had already vetted them from that relationship. But we also went in and did some additional vetting, and they do, they're a long standing company with a good history in the broker market. They do third party testing.
• 16:04
  They honor the manufacturer warranty, 3 years in this case. They do authenticity and functionality checks before they send them to you. And so it's not just basically a pass through where they mark up the price and then send you some counter fit parts.
• 16:18
  Speaker 1: And and then we actually worked with San Yodenke to get them to honor their warranty with the brokered part, which, I mean
• 16:26
  Speaker 2: Right. Yeah. We went back to Sanjay Danki, said we found you the supply. If we buy it from this broker and dropship it to you, so that's what we ended up doing, buy the part from Fusion, and we shipped it to San Yodenke. We worked with their engineering team to qualify that to be a qualified part, going through their normal build process processes, their normal quality checks, and still honoring their warranty.
• 16:48
  Like, this is legit. If the fan fails 2 years from now, it's still within warranty.
• 16:54
  Speaker 6: And Kate, can you give a little bit of context on why that is not a common occurrence? Why, like, a a vendor that you're procuring a part from may not want you digging around in their bill of materials and then trying to second source their subcomponents?
• 17:09
  Speaker 2: Yeah. I mean, for 1, they don't want competition against, the limited inventory that's out there. But there's typically most manufacturers have an ABL, an approved vendor list, and that's companies that have ISO 9,000 certifications or various quality checks depending on what their criteria are to be an approved vendor and they only will buy parts from those lists within their procurement org. So that's kind of where us being smaller, scrappy startup and being able to leverage this broker market, but still do the necessary quality checks, not cutting any corners, we were able to get this part added to their approved vendor list and then input into their build process.
• 17:48
  Speaker 1: Yeah. When I also have to say, like, give I mean, first of all, hats off to you and the team and here and Harrison for actually being able to pull this off. This is really, amazing to pull this off. But I I think this is also a testament to us going deep on these partnerships and really and just it was great to have, you know, when Spenio Dekky was out visiting us and seeing the rack. I mean, for them, they expressed, like, this is unusual.
• 18:12
  This is not something we've done really for anyone else. But it was a relationship with Oxide that really I just the reason we were willing to do this is because of how diligently Kate, you, and your team were were working with them on it. So there's a there's a lot of value in building these relationships in the supply chain.
• 18:32
  Speaker 2: Yeah. We found making the getting the inside contact directly with the engineering team, working with the procurement team closely and being honest and transparent about our demand signals, not arbitrarily inflating them just to try to swoop up more allocation has really gone a long way. So just telling them like, this is truly what we need. This is our main quantity to be successful for upcoming build. What can you do?
• 18:55
  How can we get creative about getting that inventory to us on our schedule?
• 18:59
  Speaker 1: Yeah. Yeah.
• 19:00
  Speaker 6: Well, and and also before before just gonna say, like, my experience at a much, much, much larger company, before I mean, you know, Samsung had all the buying power in the world, and it was eye opening to see, Kate, you and Kirsten and team operate because my experience has been you go and you're told a lead time and you pound the table and you threaten to use another vendor, you know, which was the Samsung model. And you see if you can't chip away that lead time. And to see kind of this collaborative approach and then be able to start working with a vendor to find the subsystems that are the actual issue, that it's not the fan, it's a TI part or something else, and then kinda collaboratively work to help them go source that, source it securely, and get that integrated in was, that was definitely a first for me, and it was it was impressive to see.
• 19:51
  Speaker 2: Yeah. I think we've been really lucky in that we've had a team that's willing to build those relationships. That's been really a key part of who we've hired in the operations at Orchid OXIDE and how we've continued to select partners of who we want to include in the design. The strategic partnership and partnership with shared values really goes a long way.
• 20:10
  Speaker 3: Kate, have
• 20:10
  Speaker 4: you ever seen something like that happen before? Where, like, you dumpster dived a bomb and injected something into their own supply chain?
• 20:17
  Speaker 2: Honestly, no. I've never seen that. Yeah. That's awesome.
• 20:22
  Speaker 4: That's really cool.
• 20:23
  Speaker 1: Yeah. And it and I again, you know, it's it's funny because when you talk to these folks and you explain you know, we we get to explain that, you know, Oxide's approach in terms of deep partnership. They it's like, oh my god. What a breath of fresh air. Like, thank you so much.
• 20:35
  I mean, it's like we find so much, the the I mean, they're so positive about that. And especially when you use that as a lens for the partnership and you select partners that and we and we have had partners where we have, like, well, we actually, you know, where I this relationship is not where we want it to be, and someone else comes along with that better relationship. It's like, no. Sorry. We're gonna go with this better relationship because the relationship is really, really, really important to us.
• 20:59
  And it's for the ability to do things like this. That I've been really
• 21:03
  Speaker 2: We've actually been able to succeed with this approach, not just with San Diego Denki, but with, Murata in our power supply. We similarly asked them to walk through their bill of materials and figure out what was constrained and why we couldn't get supply. And there was 30 parts that were constrained out of their bill of materials. We went and searched the market and narrowed that down to 11 parts where they could focus on those 11 and eventually got it down to 1 part, which is a TI buck converter. I think Steve mentioned earlier, which we ended up being able to source that TI part, get it in this in this case, we didn't have to, like, get it through broker supply.
• 21:39
  We got it through, Disney supply. And so we were able to just keep sending them that part directly, and it actually unlocked their build and not just for Oxide. So their other customers that had backlog also got their supply much earlier.
• 21:54
  Speaker 1: Oh, wow. I actually I did not realize that. Wow. That is, that is amazing. Again, what a test to that.
• 22:00
  But, obviously, like, talk about strengthening a relationship.
• 22:02
  Speaker 4: Yeah. Let's talk about partnership. Absolutely.
• 22:04
  Speaker 1: Yeah. It's like, wow. Thank you, customer. Boy, we're really in this together. It's a the Murano one, I think, is also interesting, Kate, because when we were hitting and I think, you know, with each of these, when we've hit them, it's been this, like, all hands on deck kind of like because these are existential.
• 22:19
  Like, if we if we don't have a PowerShell, we don't have a product. If we don't have fans, we don't have a product. And we so you we are obviously folks and I know as the operations team is focusing on solving this the right way, the engineering team is fantasizing about solving it the wrong way. Namely, we were like, how hard is it to build a PowerShell? And I don't I don't know if you were in any of these conversations.
• 22:38
  No. Oh, no. Where where oh, no. Eric is like, you idiots, let me explain to you how complicated the PowerShell is. And let's just say that, like, we learned quickly that we will not be building a PowerShell.
• 22:50
  Speaker 2: I mean, we do we do have a lot of startups with Inoxides, but that is not one we wanna take on.
• 22:55
  Speaker 1: And that that is not one we wanna take on, and Eric was walking us through that. Just the complicated I know, Rick, you were also often trying to get on educating people about how complicated these things are. It's like, no. No. You actually, like, you no.
• 23:08
  Really. You wanna get the you like, let's help resolve their supply issue, please. Let us not we're not gonna do this, like, ever. So, that was but that and it so, Kate, that that kinda highlights another point that I think is, you know, that I'm so impressed with about your team. But it just seems to be true about supply chain in general.
• 23:27
  It's it is detail oriented. Like, the details really, really matter, and you've got to wade into the details. It feels like an obvious thing to say, but I feel like some of the it doesn't feel like everyone is is as detail oriented as as this team has been.
• 23:41
  Speaker 2: Yeah. So, actually, that's another unique thing about operations and supply chain at Oxide is I have never in my past experience in my career been engaged so early and so deeply with the engineering team. So operations often is this silo, like you mentioned at the top of the call, but where they get engaged in the final stages of the engineering build and then they're, okay, go work the commercial side, negotiate prices, figure out how to get us capacity and supply, but their design is locked. At Oxide, we really took an integrated approach from the beginning where I was in all the hardware huddles with the team as they were making design determinations. We were checking what the supply chain situation were was for those parts as we were going.
• 24:25
  And we were able to start engaging early in three way calls with our vendors. So the engineers, the FAE support teams, the operations team was in those conversations forecasting early and often to suppliers. And I think that's really helped us succeed as we've gone through EBT and through the engineering build cycles.
• 24:43
  Speaker 1: Yeah. I mean, there were and there were a couple of parts that I mean, we, I can think of it a bunch of times where we actually changed it up where it's like, hey. This part I mean, I think I think still the world record longest lead time we heard, on a Vitesse Pemiconductor part was a 93 week lead time, if I recall correctly. And it's like, at that point, you're like, is this thing, like, no longer for sale? Like, what is the 90% complete?
• 25:05
  But that was one of those where, like, okay. Obviously, we have to find an alternative. And, you know, again, a bunch of hard work. And as you say, collaborative work between engineers and folks on your team in the supply chain at the vendors kind of trying to triangulate and find the right part. And it's been hard on on a bunch of time, but it it feels like we've been able to endure and it's part of the reason, like, at this point, I'm gonna knock on wood, but we're in I mean, we've been in surprisingly good shape from a supply chain perspective.
• 25:34
  Speaker 2: Yeah. We've, gotten ahead a lot of issues. We did some early bulk buys to cover our demand for the next year and a half, 2 years of our strategic parts that are locked in the design. And so that has really helped us. And then we've continued to iterate on the design with that collaboration engineering and operations where we check inventory as we're putting something into the design.
• 25:56
  And sometimes we've had to select, I can think of some specific parts where we had to select basically a higher grade part in the same product family from a supplier. So this is especially electrical components. We had to select 1 with like a, a tighter temperature margin or something like that that might cost 50¢ more per part, but had a bunch of inventory on hand. And so we had to make those trade off decisions, like, do we want to build in the next 2 years or do we want to spend 50¢ less per part? And so a lot of times we've had to make those decisions and choose what goes in the bomb that way.
• 26:30
  Speaker 6: And, Kate, I think one one thing that, again, on the other side of it seems obvious, but, until you and your team started following this approach or indoctrinating this approach, it was, not quite as self evident, but it's the process of kind of getting the bill of materials and listing that all out. And then in the column next to that, listing out the current estimated lead times for all those parts and then having the pricing for all those parts and basically saying, we're gonna go buy further out inventory of the low cost parts that are longest lead time, which, again, it it it does seem straightforward. But I think the methodology by which you and your team kinda walk through that has put us in this great position where it is, like, we're gonna go buy 2 or 3 years of supply of the less expensive parts that have long lead times. And, you know, then those that are the most expensive part of the bomb, things like MBME or other, we're not gonna buy buy, you know, 2 years in advance, but we know the lead times are short enough to where we're gonna be able to meet the demand forecast.
• 27:33
  And that sort of sophistication and thinking about inventory management has been a big part of why we're in the position we're in.
• 27:41
  Speaker 2: Yeah. And even, being able to select different package types. So tape and reel versus tube versus trays. We've been working with our manufacturing partner to figure out what their ideal is, which is pretty much always a full tape and reel. But we've been able to sometimes get supply by buying in tubes and then repackaging into tape and reel, which is not always ideal if you might have some fallout, but it allows us to get supply that's otherwise unobtainable.
• 28:06
  Speaker 4: So can can you tell me what any of that means?
• 28:08
  Speaker 1: Yeah. Exactly. I I was for for those who are not regular listeners of Takeaways podcast I podcast Sorry.
• 28:14
  Speaker 2: That is a
• 28:14
  Speaker 1: good one. That I haven't read. Yeah. That's a good one. But to those who are not regular listeners of Takeaways podcast, what, what are those things?
• 28:20
  Speaker 2: Yeah. So a lot of our electrical components that go into the SMT line for the PCBA. So once all the components get added onto the PCB, they come in different packages. So a tray being literally as you would think, like a flat tray where the pieces are set in different little cubbies on there and the equipment can pick up the pieces out of the tray and the automation automated equipment can put it down on the PCB. Tape and reel is a roll basically where the small components are fed into a machine via the roll and then to a different package style.
• 28:56
  So it's basically just how the SMT machine gets the package and automates putting it on the PCBA.
• 29:03
  Speaker 4: So when you're talking about, like, transferring, say, a tape to, to a tray, is that a shorthand for, like, some dude sitting there and, like, picking it up?
• 29:12
  Speaker 2: There's 3rd party. Companies we shipped them out to. Actually, we got the inventory benchmark, shipped it out to a 3rd party repackaging company that is literally, yeah, some dude taking them off out of a tray and putting them on a tape in real.
• 29:23
  Speaker 4: I mean, not to be too reductive, but yeah.
• 29:25
  Speaker 1: I did not you know, actually, it's funny because Adam was asking that. I'm like, no. No. No. We didn't have the repack.
• 29:30
  Oh my god. We did have
• 29:31
  Speaker 2: the package. I
• 29:32
  Speaker 1: I did not know that. I The and so SMP is service not technology. These are these parts that and if you if you haven't, if folks have not, you should find YouTube videos of these machines cranking. I mean, they are amazing, to watch these things in in action. I mean, the sophistication required to actually get these things onto the the the the PCB is just amazing, and it's really, really neat to watch.
• 29:56
  So, yeah, it is you say so the these are, we had to get a bunch of those. So what I think I know I think in terms of repackaging it, are we sending in multiple trays to get onto onto tape and reel. Right? I mean, because tape and reel holds a lot more parts. Right?
• 30:11
  Speaker 2: Is that Yeah. Correct.
• 30:12
  Speaker 1: Right. Okay.
• 30:13
  Speaker 2: Yeah. Multiple trays or most commonly it was tubes. Benchmark can't accommodate tube type at all. And so those are the ones we had to repackage, but we were able to get inventory easier.
• 30:23
  Speaker 1: Got it. And then on all those things that, you know, that Steve is talking about where we hit those long lead time parts early, I think a lot of those long lead time parts became even longer lead time. And we look at it in hindsight, like, oh my god. Thank god.
• 30:36
  Speaker 3: We did that.
• 30:37
  Speaker 2: I remember complaining being like, this is 26 week lead time. This is crazy. And then we're going back through and walking the bomb now weekly, and we're seeing some of those lead times at, like, 60, 70 weeks. And so
• 30:49
  Speaker 3: Kate, I'd love to hear more
• 30:50
  Speaker 4: about this because, like, obviously, everyone hears about the supply chain crunch, and we experience it through, like, you know, weird runs on toilet paper or whatever. But, like, you've been doing this for a long time. So can you describe, like, are what what things have been shocking to you, or or what have you seen that's been particularly different or alarming?
• 31:13
  Speaker 2: So I think it seems no surprise here, but everyone's talking about it. It seems worse than ever now, like, the worst I've seen in my career, in terms of lead times and parts being on allocation. When I say that, I mean, you can't just go place an order. On allocation means the manufacturer who makes a part is allocating specifically each part that comes off the line to a specific customer, which is uniquely challenging for a startup because that means we're buying with the big Googles and Microsofts and all the big manufacturers of the world that probably have very large buys and contracts in place that get them the 1st allocation of those parts coming off the line. And so I think there's a lot of things contributing contributing pandemic, war situations.
• 32:00
  And I don't wanna go into, like, all the different things that the media is covering of what's causing it and more so what the impacts are and, how we're working around
• 32:11
  Speaker 4: it. Yeah. No. Totally. I I guess I just mean, were there any things that you've seen where you're like, really?
• 32:15
  Like, 52 weeks for this? Normally, it's it's 52 minutes or whatever.
• 32:19
  Speaker 1: Oh, god. All over the place.
• 32:21
  Speaker 2: All up and down the air bomb. So electrical components like capacitors and resistors, which we kind of held towards the end. Like, I remember as we were preparing for EBT, we were buying it. Like Brian mentioned, we bought the strategic big components and we're like, okay, caps and resistors are short lead time. We can wait until, I don't know, 6 to 10 weeks before the build and we can order those.
• 32:44
  That's never been a problem before. And then when we started getting close, Kirsten and I started looking at each revision of the bomb, and we started pushing the electrical engineers to be like, nope. You need to make a determination sooner. These lead times are now 20 plus weeks. And so Which is
• 32:59
  Speaker 1: I mean, that is for a cap and over that is that is off the hook. I mean, that is crazy. I mean and you think that, like, Kate, the late like, the so the STM 328753, Adam, that is the service processor. There are people who are like, how did you even get any of those to build out of? Because those are and, Katie, is there a point we had some packages there as well.
• 33:21
  Yeah. Right? Go start, Rick.
• 33:23
  Speaker 5: Well, I was just gonna say, like, I I have a friend who had a part that kept getting pushed further and further out for lead time to the point where the update emails from the distributor actually wrapped around and sent him a date in the past.
• 33:38
  Speaker 2: That's a by the end. I feel like we get so many, auto updates now on parts we have on order that I probably get 25 to 30 emails a day now of, like, update to your date changing of your delivery date of your order. It's just crazy how often they're changing.
• 33:56
  Speaker 1: And it is and even if the the folks that are big buyers don't necessarily have I mean, as you say, like, they they get big allocations, but they also have got harder problems to solve because they need to buy bigger numbers. And the I think another advantage we had, Kate, correct me if I'm wrong, but the the fact that we were on newer designs for and, you know, we were not we're using designs that are are the the manufacturers newest design or newer design, something they're featuring. The if folks that are on these old designs are and buying at high volumes, which which is to say, like, the automotive industry buying MCUs that, you know, for that that maybe haven't that are a trailing edge design, they get are are just getting annihilated. It's it's kinda my read. Is that a is that a fair
• 34:40
  Speaker 2: Yeah. That's totally fair. So we're typically on the latest gen on most of our parts because we're typically rolling from, like, a roadmap conversation with a strategic supplier to waiting buying one of their newest releases even in, like, the quarter is released. And so we're often through the engineering build cycle at low volume been able to get away with like here we'll send you a 100 piece sample of this new part that's hitting the market. And then we use that for our engineering build and immediately place a PO for our production demand.
• 35:09
  And so we're able to get ahead of kind of some of those large volume players moving to that new technology.
• 35:17
  Speaker 1: So, Kate, another thing I wanna ask you about was panic buying because I know we have engaged in panic buying. We've definitely become part of that problem where, you know, we're like, we don't know if we need this part or not, but, okay, we just found 700 of them available. Like, let's buy it. And then we on some of those things, we've, like, we sold them later. I know early in the supply and crate supply chain crunch, we thought lots of people were engaged in this.
• 35:41
  And so there would be this big release of parts and things would kind of alleviate, but we haven't really seen that happen. Is that my imagination, or has that not yet happened?
• 35:52
  Speaker 2: A little bit of both. So I definitely think there was panic buying, and I think other industries and other companies swooped up entire thousands and thousands of parts over what they're actually building Kirsten and I would be monitoring the disti sites. Like, I remember some Saturday mornings, like, 7 AM being, like, oh my gosh. 40,000 of this part just hit the site. And typically that's like someone had placed an order, via the distribution channel and then when their order became available to ship had decided that they didn't need it.
• 36:31
  And then they'll put that entire bulk that had been forecasted and allocated to them up available for grabs. And so we saw that happen a couple times, and that's when we did our panic buying to be like, okay, we haven't seen this available in 2 months. We need to buy it now. But it hasn't happened as much as I had hoped. I think people are still holding on to their supply.
• 36:52
  Speaker 1: Yeah. And I think in particular, we've had this idea through the entire supply chain crunch. Crunch. I mean, I remember hearing, like, alright. Look.
• 36:58
  It's bad now, but by mid 2021, it's gonna clear up. By late 2021, it's gonna clear up. By early 2022, it's gonna clear up. And now people have kinda stopped saying anything. Now it's just kinda like, it's bad.
• 37:09
  Speaker 2: Yeah. Just don't It's quite bad. With a date. We haven't seen memory prices coming down a little bit this quarter, which is a nice relief. And I'm glad, like Steve mentioned, we we bought kind of as we needed it in that space.
• 37:22
  And so we did not buy a 2 year buy ahead on memory because that is one of the highest cost parts of our bill of materials. So I'm hoping for that to continue to lower and correct and so kind of by quarter by quarter.
• 37:34
  Speaker 1: Yeah. And, of course, all that stuff is, you know, we're obviously nervous because, you know, figuring out, like, you know, what you stockpile and how much. And I remember one of those decisions that was, you know, the a big question around Intel, Tafino and how much we we buy ahead on that. Turns out we made the right decision on that one. Kate, I don't know if you've heard the kind of the latest that we've been hearing and talking to folks.
• 37:54
  We didn't know if other people are seeing this out there, but witches are unobtainable right now. And people are getting quoted, you know, 200 plus day lead times from Cisco, which is not something that you would expect normally.
• 38:10
  Speaker 2: Yeah. So I think buying ahead on Tufino too was really smart and strategic for us, and we have enough to get us through our 1st year of production. And I think that was the right choice because we're already placing future demand and we've been working closely. That's one of those strategic partners. We've been meeting closely with their team, providing forecast and getting ahead of that.
• 38:29
  And so I don't expect us to have a gap there on inventory of those.
• 38:35
  Speaker 1: And the the and are there some other big wins like that? Because I feel like we've had a couple of these that just feel like, wow. Thank god we did that. What a big win. And we, you know, we got lucky a little bit, but we were also we did our homework, and we made them the right decision.
• 38:49
  Are there any others that come to mind like that?
• 38:51
  Speaker 2: I think one of the ones that comes to mind that has had its peaks and valleys, I'd call it both a war story and a win is, with an XP on the LPC 50 fives. And I I hear you laugh because it has been kind of a roller coaster, but,
• 39:08
  Speaker 1: The LPC55, you may know from our pre previous 2 spaces on its vulnerabilities. So we have a complicated relationship with LPC 55 with the the vulnerabilities that Laura has found.
• 39:17
  Speaker 2: Yeah. So it's been as we've been chasing getting the security vulnerabilities fixed, we've also been chasing getting the right package on the supply chain side. And so the LPC 55 comes in various different package types and we wanted to get the specific one for our design and then we wanted to also get, the latest revision that had the security vulnerability patches in place And so in order to do that, what we were told by NXP is we just need to get the parts that are after a certain date code. So after a certain date code, all the parts I'd be shipping out would have the fix in place. And so you're as you would imagine, us being 20,000 vulnerability, it's important to us to have that in place ahead of our production builds.
• 39:58
  We knew that our engineering builds weren't gonna get it because of the schedule. But there was no clean way to do that on a procurement side. And we were directed back to the Disney channel, basically, just keep ordering parts until you get the ones with the right day code and RMA the ones that are bad, and that's what we had to do. Wow. Which is kind of like a That
• 40:19
  Speaker 4: sounds so funny.
• 40:20
  Speaker 2: Better way.
• 40:21
  Speaker 3: But
• 40:21
  Speaker 1: Oh, god. I know. And they're like, RMA the ones that are it's like, okay. Like, but you're okay. You're sending those to a customer?
• 40:27
  I mean, like, hey. You know, it's been, yeah, it's been a it's been a challenge, I I
• 40:32
  Speaker 2: would say. I think the DSCs caught in the middle weren't happy with it either because they basically that was the workaround. They just had to, we had to keep placing orders until we consumed all of their on hand inventory and started getting new, but new build parts. And so we did that and now we have the good parts. So that's a good, that's a win there.
• 40:51
  Speaker 5: I I was reassured by NXP that we were the only people who cared about those vulnerabilities. I don't know that they actually told anyone else.
• 41:01
  Speaker 1: Exactly.
• 41:02
  Speaker 7: I I was actually working for, one of my clients at the time time was using the NXP LPC 55 s, 28 and 69 parts. NXP did not reach out to the client at all. They did not reach out to us, the people who are doing the engineering. Not a word about it.
• 41:18
  Speaker 3: Wow. Charming.
• 41:20
  Speaker 1: Yeah. Come on, NXP. And this is I mean, this and actually, it was funny. Laura Abbott, you've been on our space a couple of times. Great blog entries.
• 41:32
  And Laura, I think she was on was, was it the the amp hour, Rick? I'm sure that she she went on a couple of podcasts. But when there was a great interview with Laura. They're just like, like, why are you still in this relationship? Like, why are you still with NXP?
• 41:46
  I'm like, well, you know, if you unfortunately, there are a bunch this this is a secure silicon is actually a really hard problem as it turns out. And, hard in it it it's, we'd welcome alternatives, but this is where we are. The l and the l p c 55, it should be said, it is used our root of trust, I think, for those of you catching folks up on that. So it's very this is a very important part for us. And it's important point of principle that this not have that's had the fix for the vulnerability that that Aura found.
• 42:16
  So, and I'm really sorry to hear that NXP is not, like, first came to that for the customers. It's really, aggravating. So, Kate, are we taking and have we been able to then RMA the ones back that we effectively,
• 42:40
  Speaker 2: was opened so we could check the date code and they won't take those back.
• 42:44
  Speaker 1: Oh my god.
• 42:44
  Speaker 2: The other box, we knew then because it came in the same shipment. It was part 2 of that shipment. We did not open the box, and, yes, we are.
• 42:52
  Speaker 4: Oh, that's so nuts. They're like, well, what's the date say? It's like, well, we checked and it says this. Like, well, we can't take that now. It's an open box.
• 42:59
  Speaker 2: So it turns out I think, Nathaniel got a huge box of LPC 55 shipped to his house.
• 43:06
  Speaker 1: I mean, we get we'll use the parts. Yeah. And, hey, if anyone wants to buy some LPC 5 parts, like, price is reasonable and we are not a shady broker. Like, these are
• 43:14
  Speaker 2: not captive parts. XP and me. So, yes, the packaging.
• 43:19
  Speaker 4: Well, we know how vulnerable they are.
• 43:22
  Speaker 1: Right. Well, these are right. Exactly. I guess these 2 it is true. Okay.
• 43:25
  Fine. Yes. These parts do have certain vulnerabilities, but we know what that vulnerability is. We have a work around for it.
• 43:30
  Speaker 7: I That client I mentioned, doesn't actually isn't actually affected by that issue. So if I hear anything, maybe I'll send them your way.
• 43:37
  Speaker 1: Yeah. Yeah. There you go. Yeah. That'd be great.
• 43:39
  With the world, we'll definitely and then Kate, how often does it because the other thing I have found is it feels like, and correct me if I'm wrong, this is not a kind of a world that people talk about very much. I mean, certainly, the big buyers are there are out there are not forthcoming about their supply chain, it seems.
• 43:59
  Speaker 2: Yeah. There's a lot of, secrecy around it and sometimes for good reason around product security or just the surprise and delight of a new product coming to market. But we found that kind of the transparency and the partnership doesn't really work as well if you're trying to hide some of those details. We still obviously get NDAs where appropriate and we're smart about that. But, like, if we were treating it that way, we we And one of the big things that has allowed us to succeed is kind of selling our partners and our vendors on what Oxide's mission is, what this industry is we're gonna go disrupt, kind of what our long term strategic play is, the scale of the business we're going after and how, while we might be tiny now, we will be mighty later.
• 44:57
  And if you get in early, you'll be in our design. And there's a great win and a unique use case about how we integrate your product into our product.
• 45:08
  Speaker 1: Yeah. And I think that also I I found that one of the values that that that our partners enjoy is that we because we are taking some of their newer designs, and we're using them the way that they kind of intended them to be used. And it's you know, you look at, like, Reda Sauce, which were, got the the Redesauce, the voltage regulators we're using there, or with Samtech and the the are used to the flyover cables. But they're like and actually, I didn't quite realize this until Samtech was looking at our design in our office where they're like, thank God. Like, someone is doing it this way.
• 45:41
  It's like, yes. But you you make a very good taste that everybody should be doing it this way. They're yes. Of course, we do. But, like, you're the only ones actually doing it.
• 45:49
  And I think that that also helps. Because what I feel like we're proving out their thesis as well, I think.
• 45:55
  Speaker 2: Yeah. I think there's a lot of really cool stories of us being kind of a use case or a story that they can tell as part of their marketing as well, seeing it succeed in our product. And Symtech is definitely one of those.
• 46:08
  Speaker 1: It's okay. Can can you speak a little bit to how you built the team too? Because you also have I mean, it's just so funny to kinda think about this crazy notion that we had pre COVID that this is gonna be a team in the Bay Area. Because it turns out your team is not in the Bay Area at all, and they are
• 46:30
  Speaker 2: us. And I, the strength of what we've been able to accomplish really is our team. We're a small, scrappy operations team. We're only 5 people, but the team we have assembled is really what has allowed us to succeed. This is definitely not all on me.
• 46:44
  And so we have, Eric Anderson who came to us from a small company, manufacturing company, and he's used to wearing many hats, and he is working as our PM, and he's local in Minnesota where we're manufacturing. And so he's working really closely with Benchmark, our contract manufacturer, and he's just been crushing it. And so that's a huge, huge help.
• 47:06
  Speaker 1: Okay. So that's a good one to take apart a little bit because ERIC is like boots on ground in Minnesota. And we are by and large, we are on shore. Most of the the the the assembly is certainly happening on shore. We got and, boy, we had a recent issue where we were dealing with someone offshore and And I can't I don't know if you I definitely have the flashback of, like, oh my god.
• 47:26
  This would have been
• 47:27
  Speaker 3: Oh.
• 47:27
  Speaker 1: This would have been impossible to do.
• 47:29
  Speaker 2: We've basically been living living in Asia the past 12 months had we not chose to keep it US based, through those engineering belts. I think this was definitely the right call, especially for the size team we have. I mean, the ops team that supports an engineering build cycle on a product at Apple is probably 300 plus people strong and we're a team of 5. And so it's very different.
• 47:53
  Speaker 1: And we've been able to basically keep it onshore. I mean, it's not it wasn't necessarily it wasn't necessarily a point of principle, but it was kind of increasingly we have found that we've been able to get much better collaboration. And with Eric being able to, like, just go down to the facility and, you know, walk the line and I mean, it feels like the the spirit of problem solving has been much, much, much better by having someone who's close to the facility.
• 48:16
  Speaker 2: Yeah. And I do think keeping it in the US on point of principle is really important as well from a personal standpoint for some of our, manufacturing processes. It's a unique thing in the text place, which is a key differentiate differentiator from a security standpoint as well. Totally.
• 48:33
  Speaker 1: And it's it's funny because I think that has become more and more important. You know, I mean, at this point, you know, Oxide was founded 3 years ago, and there's been some things that have changed in the last couple of years and some things that were very, you know, very kind of idiosyncratic or iconoclastic beliefs 3 years ago. And now we're people are relieved that all this is on shore, for all the reasons you outlined. So it feels like we we made the right decision there
• 48:58
  Speaker 3: Yeah.
• 48:59
  Speaker 1: In a lot of different dimensions.
• 49:00
  Speaker 2: Yeah. And I I do wanna talk about a couple of our other team members if that's something.
• 49:03
  Speaker 1: Yeah. Yeah. Yeah. Yeah. Sorry.
• 49:04
  Speaker 2: So I mentioned Kirsten. She's been crushing it on the supply chain stuff. So a lot of those examples I gave before, but she comes from kind of different part of the product life cycle, which is helpful. So we all have different strengths on the team and hers, is a long career at Lenovo IBM working in RMA, rework, warranty, and repair side of the house. And so as we think about going to market, that skill set is really, really strong and important to us, how we think about supporting our customer till the end.
• 49:30
  And so while she's been the one doing some of these creative supply chain things like working with Sanjay Denki and the broker industry. She also has a lot of experience later in the life cycle that we are going to leverage. And so she's one of our team members. And then, we also have CJ Mendez Mendez who is in Texas, and he kind of is one of those all around ops generalists who wears many hats, and he also has a sales background. So I talk about we need to kind of sell our suppliers on the oxide mission and the opportunity and what we're building.
• 50:02
  He's been key there. He stepped up as program manager and the lead of the ops team while I was out on maternity leave. So he's, he's pretty awesome. And he's also doing a lot of ops automation work, which has been something CJ and Daniel Suter on our team have been working on together, which is also really unique about what we're doing in ops. And actually the first time I've seen it where the ops team is looking at automation, not as an after the fact, let's go put a band aid on this, but as like a first class citizen, how do we go build our ops processes and tooling from the ground up with automation in mind, like API access being front of mind when we select operations tools.
• 50:44
  Speaker 1: Okay. So I you know, we're we've gotta talk about PLM and so this is one of the other things that you did earlier when you came to Oxide. It's like, we need to go make a PLM company. I'm like, what is POM? And, Adam, did you know anything about POM?
• 50:59
  I assume you
• 51:00
  Speaker 3: didn't know anything about it.
• 51:00
  Speaker 4: No. No. I knew nothing about it. And I think, like, Kate, it must have been very very early wrote wrote this great RFP describing it.
• 51:08
  Speaker 1: And how many of those calls did you have to be on, Adam? How many did
• 51:12
  Speaker 4: you know what? I got both
• 51:13
  Speaker 2: How much pain did you take care?
• 51:15
  Speaker 4: I got both into more than you might imagine because, when people hear API, they think that that's gonna be something I care about, which is often true.
• 51:25
  Speaker 1: People in the sense is me, and I definitely recall, like, if I have to sit through this pain, Adam's gotta sit through this pain with me. Because talking to these vendors is like stepping into a time machine. And we because Kate, you should talk about, like, why API access is so important for us and some of the answers that we heard from the, the PLM industrial complex.
• 51:47
  Speaker 2: Yeah. So I come from a past life where a lot of the job was done in Microsoft Excel copying and pasting or running macros. And a lot of the data including and putting stuff together was just a lot of copy pasting. And I didn't want that. And I didn't think it scaled for a small team of 5 people to be able to do it.
• 52:07
  We're talking, thousands of different parts in our design and needing to check things like on hand inventory, lead time, pricing, and just keep track of all that data, as well as tracking, like, engineering change orders as, design changes happen, and product status. And all these things were important in a PLM solution that we could manipulate the data via API, not just one component at a time, and that we could continue to iterate on that. And we could do forecasting, we could do our MRP, like, material resource planning in terms of how many do we need, we can do our clear to builds for
• 52:46
  Speaker 3: Did we already say product
• 52:47
  Speaker 4: life cycle management? I just wanna make sure we we we actually find it.
• 52:51
  Speaker 1: Yeah. Sorry. Yeah. Okay.
• 52:53
  Speaker 4: Yeah. Sorry.
• 52:54
  Speaker 2: Sorry. For those that don't know what that is, it's basically a tool. There are some clunky old ones, which, or there are some newer ones, which we're using Duro as our PLM tool, which is a cloud based solution and it's much newer to market, a smaller startup. We found that to be the right fit fit because of API access and some of the ways they're thinking about that as a collaborative space in the cloud, which fits with what we're doing. But PLM is basically you host house things like your bill of materials.
• 53:22
  So it's the point of record that will have, these are the parts that make up your sub assembly. This is what revision are on. It even has, this is the price. This is the lead time for that part. And then it can have drawings attached or our CAD files attached.
• 53:39
  Our TTM, like our, our files we send to TTM, who is our PCB manufacturer. We get a tech data package. We build some automation around what gets extracted from Duro and from OrCAD, and then we have a package we can upload to them to go build it. So it's like, how do we know we're pointing to the source of truth? We need to have some system in place that's not just shooting emails back and forth.
• 54:05
  Speaker 1: Yeah. And we it's been really important for us. I mean, Adam, do you wanna talk about it in Postgres?
• 54:10
  Speaker 4: Sure. So, I mean, this is this is sort of a a piece of automation that that is persistently 80% done or something or maybe that's even generous. But I think what as Keith was saying, like, you want a source of truth where you say, whether I'm designing, a PCB or I'm putting together the bomb or investigating alternatives for parts, like, I want a single source of truth, and we want that ideally to be, to be RPLM, in this case, DUro. OrCAD is the design software that we're using. And when I say we, I mean people who know what they're doing, which is not me.
• 54:44
  I've I've only, like, pulled it up enough to, like, validate, this one component of how it works, which is what, OrCAD really wants is like a database, like a relational database somewhere that houses, an inventory of all the parts. And so, the sort of prototype that I did was to build a, like, a ORDB connector, that that pretends to be, that that that points to a Postgres database and then a a layer on top of Duro that makes it look like a Postgres database. So we could point Orca or design tool directly, you know, via this kind of arcane path, you know, through a service, but at Duro to use it as its system of truth. Now, where where it was a prototype that kind of fell apart was that OrCAD needs some other data, like, that it can't pull directly out of there, that needs to live in the file system. And we kinda didn't quite get all the way there, but I still have hopes that, like, that will make this one, you know, one system of truth, one source a source of truth, you know, happen in the future.
• 55:59
  Speaker 1: Well, in the end, just highlights the importance of having an API. Without an API,
• 56:02
  Speaker 4: none of that Impossible. Absolutely. Absolutely.
• 56:05
  Speaker 1: So we feel very, very strongly that I mean, an API was a constraint on that. And, Kate, I feel like the API constraint alone eliminated the vast majority of of contenders.
• 56:16
  Speaker 2: Yeah. I think we probably surveyed maybe 15 companies and maybe 2 or 3 had APIs.
• 56:23
  Speaker 3: APIs. It was
• 56:24
  Speaker 2: it was pretty bad.
• 56:25
  Speaker 1: It was so brutal. Those calls were absolutely brutal.
• 56:28
  Speaker 4: And even the APIs that I mean, the APIs that existed were part of their value. That, like, making you know, a lot of these folks don't wanna give APIs because they don't wanna make it easy to get your data out because getting easy to get your data out means it's easy to get to another
• 56:50
  Speaker 1: system. Yeah. Adam, you keep saying your data. They do it.
• 56:54
  Speaker 2: Their data. Their data that
• 57:01
  Speaker 1: Okay. Alright. I'm getting flashbacks.
• 57:03
  Speaker 6: This is too much. This is
• 57:04
  Speaker 1: it's gotten too vivid. I oh, god. Those calls were were absolute brutal. So sorry, Kate. I think you were going through and so you're talking about, about CJ and Daniel obviously collaborating on that, roping in Adam and other members of the team as well.
• 57:17
  Speaker 2: Yes. And so we've built some great automation using API access. So both API access to Duro and some of our distribution channels, we've been able to get API access to pull the open order report and the open order status. And so we're able to run some automation every Monday that basically is our in house MRP system, which is material resource planning. And that's how many of each part that's on one of these many bombs we have for different parts of the rack do we have on hand?
• 57:47
  How many do we need by a given date, given our forecast? And then when should we go place orders based on lead time? And so that's something we've been able to automate and just run a report every Monday morning. Daniel has crushed that. And that allows Kirsten and I to go receive a signal from him, like, okay, what do we need to go buy?
• 58:08
  What's most urgent? Crestco placed POs for this week. And so that's something that I've seen in the past take an entire expensive tool, a formal MRP system and, like, a whole team surround to run those types of reports.
• 58:25
  Speaker 1: Yeah. And I I mean, I definitely get the sense that I mean, Daniel was very excited to come to Oxide to work on this kind automation, and I get the sense that a lot of ops teams don't get the opportunity to focus on that. Is that a is that a fair statement?
• 58:37
  Speaker 2: Yeah. In my past life, it was always, like, a third, team within the org, like a separate silo where you'd go and put in a request for automation and then you get prioritized or often you don't get any resources for success.
• 58:49
  Speaker 4: Get prioritized. Right.
• 58:50
  Speaker 2: Yeah. And so you you might have great ideas of things that would help your workflow be so much smoother and take something that takes 8 hours down to like 3 minutes, but they wouldn't resource it. And so we've kind of built that in from the beginning with our ops team and how we structure so that we can build the tools the right way upfront. Yeah.
• 59:09
  Speaker 1: And when I think it, you know, it kinda goes into a point, Adam, that I know we talked about on the last phase as well in terms of the importance of tool making and in training that importance. And then we we read that across the company, obviously, in the ops team as well. You got I mean, Daniel's building the tools that Daniel needs, that you need Kate that that yours need, which means that it's very directly connected to to
• 59:29
  Speaker 6: its end use case.
• 59:30
  Speaker 1: And we end up building, you know, the right thing or or or evolving it to be the right thing.
• 59:36
  Speaker 2: Yeah. I think it's really exciting that we're able to do that. And I think will serve us really well and will continue to scale that way as we look look for new ops team members to kind of keep automation as part of our core.
• 59:49
  Speaker 1: So I would ask about that too because in terms of, you know, when we on the ops team, it was kind of the first time that we are hiring, non engineers effective. I mean, engine you are all obviously extremely technical and and are most engineers by education, but it was different than hiring a software engineer. And we were, as I recall, like, changing the materials a bit. What what kinds of things were you looking for in a portfolio for someone coming in on the operation side? Something you had had some really interesting things we're asking people about.
• 1:00:21
  Speaker 2: Yeah. I think a big piece of it was a little bit broader experience with the product in various stages. So not someone who just has cut POs before worked in procurement, but someone who has seen kind of a product go through infancy up through, like, new product introduction and go into market, go into staining and go to end of life. And so a lot of our team members have kind of been through products through those various stages of the life cycle, and not just served in one role, like, not just in procurement, but familiar with logistics, familiar with, like, some of the automation, familiar with ops finance, familiar with working alongside the sales team to make sure we're supporting them and able to kind of have that tight collaboration and kind of think more strategically about the product roadmap. So not quite so siloed as typically you would see these as all different functions within ops.
• 1:01:16
  Speaker 1: Yeah. And I I also I loved the way you so we asked, personally, we've asked engineers for an analysis sample and, to get into their analytic ability. And I love if I recall correctly, like, you can you rephrase that as a process improvement sample.
• 1:01:35
  Speaker 2: Yes. Like, with someone who's not okay with the status quo and wants to improve an existing process, how they've done that. And I think that's been really interesting to see people's replies to that question. Because we like the people that are not they're okay with kind of doing what their company needs them to do, but willing to shake things up a bit and like press the status quo and push for something better and work on improving something that's just because it's been done one way for past 10 years doesn't mean you need to continue to be that way.
• 1:02:07
  Speaker 1: Well, and I love that question too because it kinda captures I mean, there's, like, there there's obviously constraints. You're talking about, you know, your your you've got a process that's often involved with, like, making a thing. It's very important to the business, and you need to find a way to improve that without breaking it. And, you know, we as with all things oxide, for those that were applying to the company, we asked them to write this down and and get a written explanation of this stuff. And I I felt that really allowed us to get into some folks that we wouldn't have otherwise seen in a way that was really interesting.
• 1:02:40
  Speaker 2: Yeah. I agree. It's been a really good, way to kind of learn about people's experiences. Like I mentioned, on the ops team, we all have different backgrounds, different companies, different, kind of stages of the product life cycle that we've worked in different skill sets. And I think that's been really good to kind of get that end to end, scope across the 5 of us.
• 1:03:02
  Something we'll continue to expand upon for sure. Because we don't have everything covered. There's still some ops, ops, niches that we need we need to expand for.
• 1:03:11
  Speaker 1: For sure. So I I guess well, 2 questions then. 1, for for those I mean, obviously, we as a company are a big believer in hardware software co design. We think that there should be more startups that are focusing on this co design. But there's this supply chain issue that kind of terrifies people.
• 1:03:27
  What counsel would you give to those companies that are kind of contemplating this and being scared off by the supply chain?
• 1:03:36
  Speaker 2: I think supply chain is a complex problem, but if you're detail oriented and you continue to seek partnership versus just, like, throw the hammer down, like, I think Steve mentioned, and I've seen that approach as well. You can still be successful even in a constrained market. I think especially if you are working at lower volumes, I think once you become encumbered by some of the bureaucracy and the processes of some of the large companies, it becomes harder to think of those creative solutions and be allowed to execute them. And so I think staying scrappy and flexible, in this small start up space Yeah. That's can still succeed.
• 1:04:15
  Speaker 1: That and that's great advice. I mean, use some of the advantages you have as a smaller company. And we've I've certainly watched you do that over and over and over again at Oxide. Like, what are the advantages we have that we can bring to bear to help crack this problem? And then I guess the the the kind of next question would be, like, as you're talking about, you know, when you look forward and, obviously, we wanna be a great big company.
• 1:04:36
  But what are some of the things that you think about as you're scaling the team and scaling the way we we approach the blockchain? Because these problems are gonna get a I mean, they're obviously gonna get harder as the as the numbers get larger. So what are some of the things that you're thinking about in that regard?
• 1:04:51
  Speaker 2: Yeah. Definitely thinking about how we build out our forecasting process to be a little bit more robust and make sure that we're communicating often with our supply chain partners to get ahead of things that go on allocation, especially as our volumes increase where we're no longer in this, like, here, just give us 500 p sample. It'll be okay space. And we get into these larger orders. Hopefully, some volume pricing and, supplier contract agreements go along with that.
• 1:05:19
  And then also being very strategic about our on hand inventory and which parts we hold inventory on, and where where we hold that inventory as well. Like, whether that's at our contract manufacturer, 3rd party warehouse, at Oxide. I still think as we scale, there's some unique challenges and things that we will sort through in that space.
• 1:05:40
  Speaker 1: Yeah. Absolutely. Those are the 2 yeah. Definitely, challenges. I mean, I feel like there's we we in in operations in the supply chain, there are new challenges every day.
• 1:05:49
  Once the this space is not boring at all. Like, there's always something new that's happening, and it's been real I mean, that one thing that's been fun to just watch is watch you and the team deal with, you know, something that was totally green that we had totally under under control. Now all of a sudden, we've got a new issue that we've all gotta go tackle. And watching the team tackle that as a team is really inspiring. There've been a been a couple of those recently that have been really fun to watch.
• 1:06:19
  Speaker 2: Yeah. I think those that ever evolving, ever changing atmosphere that exists is why I love working in supply chain. It's never the same thing week over week. It's always something new that you didn't see coming.
• 1:06:32
  Speaker 1: Yeah. And then I just think also and, again, you're getting back to kind of part of the reason that Oxide was attracted to you from the beginning, allowing the working across an organization and allowing folks that work in operations supply chain to work very closely with the engineers who are actually designing the thing and working collaboratively and not kind of, you know, thinking of this as, you know, my problem versus your problem. But, thinking about it together, I think that has been really, really important for us as a team.
• 1:07:03
  Speaker 2: Yeah. I think that's really a sharp point I wanna make, Ryan, is that that was really what attracted me was I do have an engineering background, studying industrial and systems engineering. And I had worked in operations for the past 10 years, basically, and kind of gotten deeper and deeper down that silo where I got more and more isolated from the engineering org within the company. And I was a program manager for an NPI team doing new product introductions. So I was in a very cross functional role, but I was not engaged in the technical conversations as much as I enjoyed from a personal side of things.
• 1:07:39
  With that engineering background, I have a curious mind, and I wanna be involved and learn. I feel like I've gotten, like, a mini double e degree over the past few years. But, yeah, I think the ability to work across silos and not just be doing, pricing negotiation and contracts, but be doing that on top of being in the conversations about quality issue or design issue or us needing to add a grounding strap to a power spike because we failed, pre compliance testing or whatever it may be. Like, I like being in those conversations.
• 1:08:11
  Speaker 1: Yeah. Well, it's been great having you in those conversations. I mean, it has been I mean, you've been singular for the the the company. Such is so important. And, obviously, I've I've told you before, but it's been, a highlight of my own career to be able to work with you, learn from you, watch the team you built, learn from that team.
• 1:08:29
  I mean, it's just been it's been a lot of fun. Got got a lot of work ahead of us. But when people ask us about, you know, how but, you know, you're making it harder. How are you possibly doing the supply chain? It's always like, well, let let us tell you about Kate and her team.
• 1:08:42
  So it's been, it's been a lot of fun to have you here and be able to to to have you talk about it here. So thank you so much for being willing to to to make the time to join us.
• 1:08:53
  Speaker 2: Yeah. Thank you. I think we're definitely gonna follow it up with a supply chain blog post on the Oxite site later this week as well because there's a lot, to talk about and a lot that our team is doing. So it's been fun, and I'm excited to continue to scale.
• 1:09:07
  Speaker 1: That's gonna be awesome. And I think you're gonna see us, I mean, just as we have been as a company, but we are gonna be transparent and continue to be transparent. And that blog post, that that really looking forward to to reading it. That's gonna be great as we talk about some of the things we're working on.
• 1:09:20
  Speaker 2: For sure.
• 1:09:21
  Speaker 1: Awesome. Alright. Thank you, Kate. Thank you, everyone. And we'll, we'll see you next time.
• 1:09:26
  Speaker 2: Thanks, Brian. Thanks, Adam.
• 1:09:28
  Speaker 1: Thanks.