1 00:00:00,000 --> 00:00:00,449 [Miles] So, 2 00:00:00,649 --> 00:00:01,550 [Miles] May twentieth, 3 00:00:01,900 --> 00:00:04,199 [Miles] Sam Altman walks into a YC event, 4 00:00:04,399 --> 00:00:05,899 [Miles] and by the time he's done talking, 5 00:00:05,899 --> 00:00:06,349 [Grant] Wow. 6 00:00:06,349 --> 00:00:08,750 [Miles] every startup in the room has an offer on the table. 7 00:00:09,236 --> 00:00:10,086 [Miles] Two million dollars. 8 00:00:10,336 --> 00:00:10,786 [Miles] Except, 9 00:00:11,036 --> 00:00:11,936 [Miles] not cash. 10 00:00:12,286 --> 00:00:14,185 [Speaker 3] Compute, API tokens, 11 00:00:14,386 --> 00:00:15,336 [Speaker 3] and in exchange, 12 00:00:15,435 --> 00:00:18,786 [Speaker 3] each startup signs an uncapped SAFE giving OpenAI future equity. 13 00:00:19,336 --> 00:00:21,036 [Speaker 3] TechCrunch called it a mic drop moment, 14 00:00:21,185 --> 00:00:21,836 [Speaker 3] and honestly, 15 00:00:21,936 --> 00:00:22,636 [Speaker 3] that's about right. 16 00:00:22,786 --> 00:00:23,336 [Miles] Yeah, 17 00:00:23,386 --> 00:00:30,586 [Miles] but mic drop is doing a lot of work there. Because the moment you look at what an uncapped SAFE actually means on a cap table, 18 00:00:30,786 --> 00:00:32,536 [Miles] the math gets complicated fast. 19 00:00:33,085 --> 00:00:35,085 [Speaker 3] That's exactly where we're going today. 20 00:00:35,466 --> 00:00:36,716 [Miles] With pleasure. 21 00:00:36,916 --> 00:00:38,766 [Miles] So here's the shape of the episode. 22 00:00:39,016 --> 00:00:40,716 [Miles] We're going to walk through the deal mechanics, 23 00:00:40,916 --> 00:00:47,216 [Miles] what uncapped actually means when OpenAI is sitting on your cap table alongside YC and your seed investors. 24 00:00:47,516 --> 00:00:50,016 [Speaker 3] Then we talked to two founders in the same batch, 25 00:00:50,116 --> 00:00:50,866 [Speaker 3] same offer, 26 00:00:51,066 --> 00:00:52,466 [Speaker 3] completely different calls. 27 00:00:52,566 --> 00:00:53,366 [Speaker 3] One signed, 28 00:00:53,716 --> 00:00:54,216 [Speaker 3] one walked. 29 00:00:54,416 --> 00:00:57,816 [Speaker 3] And the reasons tell you more about this deal than any term sheet analysis will. 30 00:00:58,216 --> 00:01:01,566 [Miles] Nodding. And there's a YC partner lens on the back half, 31 00:01:01,716 --> 00:01:04,266 [Miles] pattern matching this against the hundreds of deals they've seen. 32 00:01:04,019 --> 00:01:06,219 [Miles] They've seen blow up or pay off at Series A. 33 00:01:06,469 --> 00:01:08,119 [Speaker 3] Because the real cost of this thing is, 34 00:01:08,519 --> 00:01:12,120 [Speaker 3] nobody finds out until the Series A room-that's the whole problem. 35 00:01:12,469 --> 00:01:18,670 [Miles] Founder had a sharp read on that timing risk and American Bazaar Online flagged the founder control questions early. 36 00:01:19,120 --> 00:01:20,069 [Miles] We'll pull both threads. 37 00:01:20,370 --> 00:01:20,969 [Speaker 3] All right, 38 00:01:21,120 --> 00:01:23,769 [Speaker 3] let's start where everyone in that room started-Altman's 39 00:01:23,870 --> 00:01:24,170 [Speaker 3] offer, 40 00:01:24,319 --> 00:01:25,269 [Speaker 3] what it actually said, 41 00:01:25,519 --> 00:01:26,269 [Speaker 3] and why the room went 42 00:01:26,269 --> 00:01:26,370 [Miles] Yes. 43 00:01:26,370 --> 00:01:27,819 [Speaker 3] quiet before it went electric. 44 00:01:28,874 --> 00:01:30,774 [Miles] We all just looked at each other. 45 00:01:31,374 --> 00:01:34,024 [Miles] Nobody said anything for maybe five seconds, 46 00:01:34,573 --> 00:01:37,324 [Miles] and then the room went completely electric. 47 00:01:38,036 --> 00:01:40,986 [Speaker 3] Five seconds of silence in a room full of founders. 48 00:01:41,236 --> 00:01:42,385 [Speaker 3] That's a long time. 49 00:01:42,585 --> 00:01:43,686 [Miles] It really is. 50 00:01:43,936 --> 00:01:46,635 [Miles] So let me set this up for anyone who wasn't there. 51 00:01:47,236 --> 00:01:47,885 [Miles] Tuesday night, 52 00:01:48,186 --> 00:01:51,835 [Miles] May 20th, Sam Altman walks into a YC event. 53 00:01:51,986 --> 00:01:53,086 [Miles] Not a demo day, 54 00:01:53,136 --> 00:01:54,586 [Miles] not a big public stage, 55 00:01:54,786 --> 00:01:56,136 [Miles] just the current batch. 56 00:01:56,236 --> 00:01:57,836 [Miles] And according to TechCrunch, 57 00:01:57,986 --> 00:02:04,486 [Miles] he offers every single startup in the room $2 million worth of OpenAI API tokens. 58 00:02:04,275 --> 00:02:06,325 [Miles] tokens in exchange for equity. 59 00:02:06,676 --> 00:02:09,275 [Speaker 3] All one hundred and sixty nine of 60 00:02:09,275 --> 00:02:09,376 [Miles] Yeah. 61 00:02:09,376 --> 00:02:10,325 [Speaker 3] them at once. 62 00:02:10,804 --> 00:02:11,753 [Miles] At once! 63 00:02:12,403 --> 00:02:17,954 [Miles] YC partner Tyler Bosmeny went on X immediately afterwards and called it a mic drop moment, 64 00:02:18,253 --> 00:02:19,454 [Miles] which honestly is earned. 65 00:02:19,903 --> 00:02:24,503 [Speaker 3] So what does $2 million in tokens actually mean in practice? 66 00:02:24,753 --> 00:02:25,204 [Speaker 3] Because, 67 00:02:25,304 --> 00:02:26,304 [Speaker 3] cash I understand, 68 00:02:26,753 --> 00:02:27,853 [Speaker 3] tokens are different. 69 00:02:28,304 --> 00:02:28,853 [Miles] Right. 70 00:02:28,903 --> 00:02:31,103 [Miles] So the founder piece on this put it well. 71 00:02:31,353 --> 00:02:33,454 [Miles] We're talking about API credits, 72 00:02:33,704 --> 00:02:34,603 [Miles] compute runway. 73 00:02:34,954 --> 00:02:36,403 [Miles] According to their analysis, 74 00:02:36,503 --> 00:02:39,753 [Miles] that's roughly 12 to 24 months of serious... 75 00:02:39,508 --> 00:02:42,758 [Miles] various AI infrastructure for most early-stage teams. 76 00:02:43,058 --> 00:02:44,508 [Miles] You don't hire the servers, 77 00:02:44,707 --> 00:02:46,707 [Miles] you don't negotiate a cloud contract, 78 00:02:46,908 --> 00:02:48,008 [Miles] you just build. 79 00:02:48,308 --> 00:02:49,608 [Speaker 3] No cash out of pocket. 80 00:02:49,908 --> 00:02:50,458 [Miles] Nothing. 81 00:02:50,908 --> 00:02:54,408 [Miles] And that's what made it feel like a gift before it felt like a deal. 82 00:02:55,608 --> 00:02:57,358 [Speaker 3] So the room goes quiet, 83 00:02:57,558 --> 00:02:58,158 [Speaker 3] then electric. 84 00:02:58,408 --> 00:03:02,308 [Speaker 3] What was the first instinct before anyone started running numbers? 85 00:03:02,900 --> 00:03:05,700 [Miles] I think the first instinct was pure relief. 86 00:03:06,000 --> 00:03:07,950 [Miles] If you're six weeks into your batch, 87 00:03:08,099 --> 00:03:16,700 [Miles] you're burning through whatever YC gave you, you're trying to figure out your AI infrastructure costs, and someone just handed you a year's worth of compute? 88 00:03:17,000 --> 00:03:19,150 [Miles] I mean the emotional reaction is obvious. 89 00:03:19,599 --> 00:03:20,150 [Speaker 3] Yeah, 90 00:03:20,300 --> 00:03:22,099 [Speaker 3] the math looks clean on the surface. 91 00:03:22,700 --> 00:03:26,200 [Speaker 3] American Bazaar Online flagged a vendor lock-in angle pretty quickly. 92 00:03:26,500 --> 00:03:28,100 [Speaker 3] Critics raised it within hours, 93 00:03:28,450 --> 00:03:30,350 [Speaker 3] but in the room, at that moment... 94 00:03:30,548 --> 00:03:31,397 [Miles] In the room, 95 00:03:31,548 --> 00:03:33,298 [Miles] nobody's thinking vendor lock-in. 96 00:03:33,747 --> 00:03:34,198 [Miles] They're thinking, 97 00:03:34,397 --> 00:03:37,048 [Miles] I can actually build the thing I came here to build. 98 00:03:37,497 --> 00:03:40,048 [Speaker 3] Which is exactly when you should be most careful. 99 00:03:40,497 --> 00:03:41,747 [Miles] I've seen this play out: 100 00:03:42,147 --> 00:03:44,247 [Miles] the deals that feel frictionless at signing, 101 00:03:44,647 --> 00:03:46,348 [Miles] those are the ones worth turning over, 102 00:03:46,747 --> 00:03:51,147 [Miles] because what you signed and what you thought you signed are sometimes two different things. 103 00:03:51,747 --> 00:03:53,948 [Speaker 3] So what did they actually sign? 104 00:03:54,397 --> 00:03:57,548 [Speaker 3] That's the question nobody in that room had fully answered yet. 105 00:03:57,491 --> 00:04:03,091 [Speaker 3] What does an uncapped SAFE really mean when the person across the table is Sam Altman? 106 00:04:04,211 --> 00:04:10,561 [Miles] So the structure Jared Friedman confirmed to TechCrunch uncapped SAFE that converts at the Series A. 107 00:04:11,111 --> 00:04:13,911 [Miles] Walk me through what that actually means for a founder who signed it. 108 00:04:14,161 --> 00:04:14,711 [Speaker 3] Right, 109 00:04:14,811 --> 00:04:18,211 [Speaker 3] so a SAFE is a promise of future equity – no interest, 110 00:04:18,361 --> 00:04:19,312 [Speaker 3] no maturity date. 111 00:04:19,562 --> 00:04:21,911 [Speaker 3] It just sits there until you raise a priced round. 112 00:04:22,312 --> 00:04:24,911 [Speaker 3] And uncapped means no valuation ceiling. 113 00:04:25,111 --> 00:04:27,262 [Speaker 3] The higher your Series A valuation, 114 00:04:27,562 --> 00:04:29,762 [Speaker 3] the smaller OpenAI's actual bite. 115 00:04:29,939 --> 00:04:31,540 [Miles] Which sounds great, right? 116 00:04:31,639 --> 00:04:32,389 [Miles] On paper, 117 00:04:32,540 --> 00:04:33,740 [Miles] founders want uncapped, 118 00:04:33,839 --> 00:04:35,639 [Miles] the investor bears the valuation risk, 119 00:04:35,689 --> 00:04:36,240 [Miles] not you. 120 00:04:36,389 --> 00:04:37,790 [Speaker 3] And Tom Blomfield, 121 00:04:37,939 --> 00:04:38,839 [Speaker 3] YC partner, 122 00:04:39,089 --> 00:04:40,839 [Speaker 3] basically confirmed that math on X. 123 00:04:40,939 --> 00:04:48,689 [Speaker 3] He framed it as OpenAI exchanging roughly $800 million in compute for about 2% equity across 400 startups. 124 00:04:48,839 --> 00:04:50,439 [Speaker 3] So they're betting on volume. 125 00:04:50,740 --> 00:04:52,490 [Miles] I mean, that's a wild bet, 126 00:04:52,589 --> 00:04:53,689 [Miles] but here's my question. 127 00:04:53,839 --> 00:04:56,339 [Miles] Who in that room was actually modeling that? 128 00:04:56,564 --> 00:04:57,213 [Speaker 3] Nobody. 129 00:04:57,514 --> 00:04:58,164 [Miles] Nobody. 130 00:04:58,414 --> 00:04:59,164 [Miles] You're at a YC event. 131 00:04:59,213 --> 00:04:59,514 [Miles] Altman just did what Tyler Bosmeny called the mic drop moment, 132 00:04:59,713 --> 00:05:02,864 [Miles] and the vibe is free compute for a year. 133 00:05:02,963 --> 00:05:04,864 [Miles] Your brain goes there first. 134 00:05:05,014 --> 00:05:06,314 [Miles] That's exactly what I've seen happen with term sheets that look clean on the surface. 135 00:05:06,713 --> 00:05:10,614 [Speaker 3] You see the number, 136 00:05:11,114 --> 00:05:11,764 [Speaker 3] you feel the relief, 137 00:05:11,914 --> 00:05:13,014 [Speaker 3] and the structural details just don't land until later. 138 00:05:13,414 --> 00:05:16,864 [Speaker 3] Agreed. The number was real. 139 00:05:17,264 --> 00:05:18,764 [Miles] $2 million in API credits, useful compute, 140 00:05:19,014 --> 00:05:21,463 [Miles] $2 million in API credits, useful compute, 141 00:05:21,764 --> 00:05:24,264 [Miles] but the equity is the part nobody priced out that night. 142 00:05:24,531 --> 00:05:24,981 [Speaker 3] Okay, 143 00:05:25,031 --> 00:05:26,281 [Speaker 3] so let's actually run it. 144 00:05:26,632 --> 00:05:31,531 [Speaker 3] Uncapped means OpenAI converts at whatever your Series A values the company at. 145 00:05:31,731 --> 00:05:33,181 [Speaker 3] If you raise at $10 million, 146 00:05:33,431 --> 00:05:34,531 [Speaker 3] they get a bigger slice. 147 00:05:34,731 --> 00:05:36,681 [Speaker 3] If you raise at $100 million... Right, 148 00:05:36,681 --> 00:05:37,931 [Miles] They get a smaller slice. 149 00:05:38,181 --> 00:05:41,981 [Miles] The founder who builds fast and raises high actually dilutes OpenAI down. 150 00:05:42,981 --> 00:05:44,831 [Speaker 3] which is the founder-friendly read. 151 00:05:45,031 --> 00:05:51,382 [Speaker 3] And Founder's analysis noted this is the most aggressive equity-for-compute structure any model lab has tried. 152 00:05:51,699 --> 00:05:54,500 [Speaker 3] Most of the coverage called it founder-friendly, and technically, 153 00:05:54,649 --> 00:05:55,349 [Speaker 3] they're not wrong. 154 00:05:55,732 --> 00:05:58,282 [Miles] I mean, I've pushed back a little on that framing. 155 00:05:58,532 --> 00:05:59,082 [Speaker 3] Go. 156 00:05:59,632 --> 00:06:03,782 [Miles] Founder-friendly on paper assumes you're going to raise a big Series A fast. 157 00:06:03,982 --> 00:06:09,482 [Miles] A lot of these companies won't. And the uncapped structure also means there's no floor on OpenAI stake, 158 00:06:09,782 --> 00:06:11,982 [Miles] no cap protecting anyone in either direction. 159 00:06:12,532 --> 00:06:13,232 [Speaker 3] And notably, 160 00:06:13,482 --> 00:06:17,882 [Speaker 3] Friedman told TechCrunch this version doesn't include a most favored nation clause. 161 00:06:18,082 --> 00:06:21,832 [Speaker 3] So if you later sign a capped SAFE with another investor at better terms, 162 00:06:21,588 --> 00:06:22,138 [Speaker 3] Better terms, 163 00:06:22,588 --> 00:06:25,237 [Speaker 3] OpenAI can't automatically match those terms. 164 00:06:25,438 --> 00:06:28,588 [Miles] That's actually the one thing that is founder-friendly here. 165 00:06:28,938 --> 00:06:31,788 [Miles] YC's own standard deal includes MFN. 166 00:06:31,938 --> 00:06:33,838 [Speaker 3] So they gave a little back, 167 00:06:34,237 --> 00:06:35,237 [Speaker 3] strip the MFN. 168 00:06:35,237 --> 00:06:36,487 [Miles] Which is real, 169 00:06:36,588 --> 00:06:42,088 [Miles] but most first-time founders don't even know what MFN means until their lawyer explains it after the fact. 170 00:06:42,188 --> 00:06:43,438 [Speaker 3] That's the honest answer. 171 00:06:43,638 --> 00:06:46,188 [Speaker 3] The math is founder-friendly if you raise well. 172 00:06:46,288 --> 00:06:50,038 [Speaker 3] The problem is you're signing the deal before you know how you'll raise. 173 00:06:50,451 --> 00:06:54,451 [Miles] Yeah, and that gap between the paper terms and the moment in the room, 174 00:06:54,552 --> 00:06:57,802 [Miles] that's where I think the interesting decisions actually got made. 175 00:06:57,901 --> 00:07:00,002 [Speaker 3] And speaking of decisions made in that room, 176 00:07:00,101 --> 00:07:05,551 [Miles] Two founders who were sitting there walked out with completely different answers to the same offer. 177 00:07:06,601 --> 00:07:07,351 [Grant] Two founders, 178 00:07:07,701 --> 00:07:08,251 [Grant] same room, 179 00:07:08,551 --> 00:07:09,151 [Grant] same offer, 180 00:07:09,651 --> 00:07:10,301 [Grant] different answers. 181 00:07:11,001 --> 00:07:13,851 [Miles] And the gap between those answers is where this gets interesting. 182 00:07:14,051 --> 00:07:15,751 [Miles] So walk me through the one who signed. 183 00:07:15,851 --> 00:07:17,501 [Miles] What was the actual tipping point? 184 00:07:17,945 --> 00:07:20,295 [Grant] So she said something I keep coming back to. 185 00:07:20,545 --> 00:07:22,045 [Grant] It wasn't really about the commute. 186 00:07:22,295 --> 00:07:23,295 [Grant] She's got enough runway. 187 00:07:23,745 --> 00:07:26,495 [Grant] What she wanted, and she was pretty clear about this, 188 00:07:26,695 --> 00:07:31,395 [Grant] was that OpenAI name on her cap table before she walks into a Series A pitch. 189 00:07:31,795 --> 00:07:32,445 [Miles] A signal. 190 00:07:32,645 --> 00:07:33,495 [Grant] A signal. 191 00:07:33,895 --> 00:07:36,745 [Grant] She's raising in a market where every deck looks the same. 192 00:07:36,945 --> 00:07:41,145 [Grant] Having a model lab as an investor says something that two million in cash doesn't. 193 00:07:41,445 --> 00:07:42,095 [Miles] Okay, 194 00:07:42,145 --> 00:07:43,845 [Miles] but did she run the lock-in math? 195 00:07:44,473 --> 00:07:47,623 [Grant] That's where it got quiet—I asked her directly, 196 00:07:48,073 --> 00:07:53,873 [Grant] "What happens if you need to switch providers eighteen months from now?" and she paused for a while before she answered. 197 00:07:54,373 --> 00:07:54,873 [Miles] What did she say? 198 00:07:55,473 --> 00:07:57,873 [Grant] She said she'd cross that bridge when she got there. 199 00:07:58,323 --> 00:07:58,623 [Miles] Yeah, 200 00:07:58,923 --> 00:07:59,773 [Miles] that's not a strategy, 201 00:07:59,873 --> 00:08:00,423 [Miles] That's a hope. 202 00:08:00,761 --> 00:08:05,011 [Grant] Founder published an analysis of the actual switching costs on something like this. 203 00:08:05,511 --> 00:08:05,961 [Grant] Their read: 204 00:08:06,261 --> 00:08:11,111 [Grant] moving from one frontier model to another runs about six weeks of engineering for a small team, 205 00:08:11,461 --> 00:08:14,211 [Grant] plus another couple weeks of quality regression to close. 206 00:08:14,461 --> 00:08:16,661 [Grant] And that's before you touch pricing risk. 207 00:08:17,161 --> 00:08:20,961 [Grant] OpenAI has moved API prices six times in the last two years. 208 00:08:21,461 --> 00:08:22,511 [Miles] Six times. 209 00:08:22,511 --> 00:08:22,811 [Grant] Mm-hmm. 210 00:08:22,811 --> 00:08:26,911 [Miles] So you're not just locked in architecturally, you can't even forecast your own cost. 211 00:08:26,809 --> 00:08:27,209 [Miles] Right—and none of that shows up in the SAFE document; it shows up in your engineering org chart a year later. 212 00:08:27,659 --> 00:08:33,359 [Grant] Okay, 213 00:08:33,809 --> 00:08:34,159 [Miles] so flip to the founder who passed. 214 00:08:34,409 --> 00:08:36,109 [Miles] What did he see? 215 00:08:36,359 --> 00:08:37,109 [Miles] He was blunter: 216 00:08:37,609 --> 00:08:38,459 [Grant] he'd actually thought about the platform risk question before anyone handed him a term sheet. 217 00:08:38,759 --> 00:08:43,009 [Grant] Meaning the Calacanis thing? 218 00:08:43,409 --> 00:08:44,659 [Miles] He knew that argument; Jason Calacanis posted publicly on X warning founders that "there's a non-zero chance that OpenAI will study exactly 219 00:08:45,159 --> 00:08:53,859 [Grant] what you're doing. 220 00:08:53,859 --> 00:08:54,309 [Speaker 4] what you're doing. 221 00:08:53,657 --> 00:08:57,107 [Grant] Exactly what your startup is doing and replicate it into a free offering. 222 00:08:57,657 --> 00:08:59,257 [Grant] Classic platform playbook. 223 00:08:59,407 --> 00:09:00,057 [Miles] And look, 224 00:09:00,107 --> 00:09:01,557 [Miles] TechCrunch pushed back on that. 225 00:09:01,657 --> 00:09:07,057 [Miles] They pointed out OpenAI can watch what you're building through API calls whether or not you've signed equity over. 226 00:09:07,107 --> 00:09:08,757 [Miles] The deal doesn't give them new visibility. 227 00:09:09,107 --> 00:09:09,607 [Grant] True. 228 00:09:09,807 --> 00:09:11,957 [Grant] But this founder's point was different. 229 00:09:12,057 --> 00:09:13,607 [Grant] He said right now, 230 00:09:13,657 --> 00:09:17,107 [Grant] OpenAI has a financial incentive to want me to succeed. 231 00:09:17,357 --> 00:09:20,207 [Grant] If they also start competing with me in my category, 232 00:09:20,307 --> 00:09:21,057 [Grant] which they might. 233 00:09:21,081 --> 00:09:23,881 [Grant] I'm holding equity from a direct competitor. 234 00:09:24,381 --> 00:09:26,231 [Miles] Your investor becomes your competitor. 235 00:09:26,631 --> 00:09:27,981 [Grant] He said it pretty cleanly: 236 00:09:28,331 --> 00:09:30,431 [Grant] I don't want a vendor on my cap table, 237 00:09:30,631 --> 00:09:31,281 [Grant] full stop. 238 00:09:31,731 --> 00:09:33,881 [Miles] Did that cost him anything inside the batch, 239 00:09:33,931 --> 00:09:34,131 [Miles] I mean? 240 00:09:34,581 --> 00:09:35,181 [Grant] Socially, 241 00:09:35,431 --> 00:09:35,881 [Grant] a little. 242 00:09:36,431 --> 00:09:40,931 [Grant] He said there was a period where people in the cohort treated it like he'd turned down free money, 243 00:09:41,231 --> 00:09:42,431 [Grant] like he'd made a mistake. 244 00:09:42,681 --> 00:09:43,881 [Miles] Which is interesting, 245 00:09:44,131 --> 00:09:45,531 [Miles] because the founder who signed, 246 00:09:45,831 --> 00:09:47,381 [Miles] her reasoning was about the room. 247 00:09:47,385 --> 00:09:49,485 [Miles] The signal-the social proof. 248 00:09:49,685 --> 00:09:53,635 [Grant] And his reasoning was also about the room-just a different read of it. 249 00:09:53,985 --> 00:09:57,885 [Grant] He thought the pressure to sign itself was a signal worth paying attention to. 250 00:09:59,449 --> 00:10:01,149 [Miles] I don't think either of them is wrong. 251 00:10:01,449 --> 00:10:01,799 [Grant] No. 252 00:10:02,299 --> 00:10:03,799 [Grant] And neither of them could tell me, 253 00:10:03,899 --> 00:10:05,049 [Grant] on the day they made the call, 254 00:10:05,199 --> 00:10:06,499 [Grant] whether they'd made the right one. 255 00:10:07,149 --> 00:10:13,749 [Grant] She doesn't know what her Series A will look like; he doesn't know if his competitors are already burning through those tokens to outship him. 256 00:10:14,049 --> 00:10:15,249 [Miles] Which is the honest answer? 257 00:10:15,699 --> 00:10:20,599 [Miles] You're making a bet under incomplete information and the outcome isn't scored for another two years. 258 00:10:20,849 --> 00:10:24,699 [Grant] That's the part neither founder could fully answer-and honestly, 259 00:10:24,799 --> 00:10:26,099 [Grant] I'm not sure anyone could. 260 00:10:26,073 --> 00:10:33,073 [Grant] We put both of those conversations to a YC partner who's watched this exact fork across hundreds of batches. 261 00:10:33,373 --> 00:10:37,273 [Grant] What they said about the pattern changes how you hear both of those founders. 262 00:10:37,523 --> 00:10:38,523 [Grant] Let's get into them. 263 00:10:39,353 --> 00:10:42,353 [Grant] So the two founders gave us their reasoning in the moment. 264 00:10:42,603 --> 00:10:47,853 [Grant] What I want to know is whether that reasoning holds up against what YC partners actually see happen next. 265 00:10:48,003 --> 00:10:49,903 [Miles] And that's the pattern math question, 266 00:10:50,203 --> 00:10:52,703 [Miles] because any one founder thinks they're deciding on compute. 267 00:10:53,103 --> 00:10:56,603 [Miles] A partner who's watched hundreds of companies go through this sees something different. 268 00:10:56,803 --> 00:11:00,653 [Grant] Walk me through the cap table math first, because I don't think everyone's visualizing it. 269 00:11:00,803 --> 00:11:01,203 [Miles] Okay, 270 00:11:01,303 --> 00:11:03,053 [Miles] so you come into YC, 271 00:11:03,103 --> 00:11:05,153 [Miles] you've already taken the YC standard terms. 272 00:11:05,081 --> 00:11:19,681 [Miles] Terms you may have a pre-seed SAFE sitting unconverted now you add the OpenAI uncapped SAFE on top founder ran the scenario where a stacked unconverted instruments reach Series A representing 35 to 45 percent of the cap table before 273 00:11:19,681 --> 00:11:22,181 [Grant] Before a single Series A dollar comes in, 274 00:11:22,281 --> 00:11:22,681 [Grant] does 275 00:11:22,681 --> 00:11:27,281 [Miles] a single dollar and the OpenAI instrument is one more line in that math dramatically 276 00:11:27,281 --> 00:11:29,581 [Grant] the Series A timeline change the calculus at all? 277 00:11:30,531 --> 00:11:32,531 [Miles] founder flagged this specifically 278 00:11:32,761 --> 00:11:42,861 [Miles] Only 15% of seed-funded startups now reach Series A within two years, down from 51%. The uncapped SAFE converts at a price round that, statistically, 279 00:11:43,011 --> 00:11:44,961 [Miles] most founders won't see on that timeline. 280 00:11:45,361 --> 00:11:47,261 [Grant] So the instrument may just sit there. 281 00:11:47,461 --> 00:11:48,561 [Miles] It sits there, 282 00:11:48,711 --> 00:11:50,811 [Miles] it accumulates, and the longer it sits, 283 00:11:50,961 --> 00:11:53,761 [Miles] the worse the conversion math gets if the company is growing. 284 00:11:54,061 --> 00:11:58,511 [Grant] That's the thing that token maxing framing obscures. You're not just deciding on compute. 285 00:11:58,521 --> 00:12:04,371 [Grant] Cute. You're deciding on a permanent relationship with your infrastructure provider before you know what your product actually needs. 286 00:12:04,571 --> 00:12:07,471 [Miles] Which brings me to the question I kept coming back to: 287 00:12:07,871 --> 00:12:12,321 [Miles] is token maxing a real operating strategy or is it a rationalization? 288 00:12:12,621 --> 00:12:18,021 [Miles] Like, did these founders sign because the economics made sense or because Sam Altman was in the room? 289 00:12:18,617 --> 00:12:19,117 [Grant] I mean, 290 00:12:19,167 --> 00:12:20,517 [Grant] those are not mutually exclusive. 291 00:12:20,867 --> 00:12:21,417 [Miles] Right, 292 00:12:21,467 --> 00:12:21,667 [Miles] right. 293 00:12:21,817 --> 00:12:23,267 [Miles] But when you strip out the room, 294 00:12:23,367 --> 00:12:23,917 [Miles] the moment, 295 00:12:24,067 --> 00:12:25,017 [Miles] the social pressure, 296 00:12:25,267 --> 00:12:29,567 [Miles] the deal still needs to work on paper and the conflict of interest question is real. 297 00:12:29,867 --> 00:12:33,467 [Miles] YC was advising on a deal that came from its own alumni network. 298 00:12:33,617 --> 00:12:35,817 [Grant] One partner I spoke to put it this way. 299 00:12:35,867 --> 00:12:38,067 [Grant] The signal value the founder mentioned, 300 00:12:38,117 --> 00:12:40,267 [Grant] the OpenAI name on the cop table, 301 00:12:40,367 --> 00:12:44,367 [Grant] That's only worth something if the Series A investor reads it as validation. 302 00:12:44,761 --> 00:12:48,061 [Grant] Some will; some will read it as vendor dependency. 303 00:12:48,861 --> 00:12:50,911 [Miles] And you won't know until you're in the room. 304 00:12:51,289 --> 00:12:53,089 [Grant] You won't know until you're in the room. 305 00:12:53,339 --> 00:12:56,739 [Grant] Which means the S-26 founders who signed three weeks ago, 306 00:12:57,189 --> 00:12:58,389 [Grant] the verdict isn't in yet. 307 00:12:58,639 --> 00:12:59,489 [Grant] Tokens are live, 308 00:12:59,589 --> 00:13:00,639 [Grant] products are being built, 309 00:13:00,839 --> 00:13:02,289 [Grant] but the test hasn't arrived. 310 00:13:02,889 --> 00:13:09,639 [Miles] The real question every founder in his batch is sitting with is what does your cap table look like when that room happens? 311 00:13:10,139 --> 00:13:14,139 [Grant] And whether OpenAI being in it helps you or complicates the conversation. 312 00:13:14,839 --> 00:13:16,839 [Miles] That's where this lands for now. 313 00:13:17,817 --> 00:13:21,067 [Grant] So right now, somewhere in a co-working space, 314 00:13:21,267 --> 00:13:24,817 [Grant] an S-26 founder is three weeks into building on those tokens. 315 00:13:25,067 --> 00:13:26,217 [Grant] The product's moving, 316 00:13:26,317 --> 00:13:28,217 [Grant] but the Series A hasn't happened. 317 00:13:28,417 --> 00:13:30,067 [Miles] Right, and that's the thing. 318 00:13:30,267 --> 00:13:33,067 [Miles] The real price of this deal hasn't been paid yet. 319 00:13:33,367 --> 00:13:37,517 [Miles] No one knows what that equity actually costs until they're in a Series A room. 320 00:13:37,667 --> 00:13:40,117 [Grant] And it's not just S-26 anymore. 321 00:13:40,767 --> 00:13:42,417 [Grant] According to American Bazaar Online, 322 00:13:42,767 --> 00:13:44,967 [Grant] the deal reportedly extends to the summer 2025. 323 00:13:44,729 --> 00:13:46,529 [Grant] From our S26 Batch 2, 324 00:13:47,079 --> 00:13:47,429 [Grant] YC 325 00:13:47,479 --> 00:13:51,579 [Grant] even extended the application deadline specifically for founders who wanted in, 326 00:13:51,629 --> 00:13:55,229 [Miles] So we're talking hundreds of founders who still haven't faced this decision. 327 00:13:55,229 --> 00:13:55,379 [Grant] which 328 00:13:55,379 --> 00:13:56,929 [Miles] The question is live. 329 00:13:56,929 --> 00:14:00,079 [Grant] means the real question every founder has. 330 00:14:00,202 --> 00:14:08,052 [Miles] To answer not just the ones who signed is what's your compute plan for the next 24 months and is it built on a single point of failure? 331 00:14:08,202 --> 00:14:12,652 [Grant] And the honest answer for most early stage companies right now, probably yes, 332 00:14:12,852 --> 00:14:16,152 [Grant] because cheap and available compute wins in the short term. 333 00:14:16,302 --> 00:14:21,452 [Miles] I've watched founders make this call before, not with tokens, but with AWS credits, 334 00:14:21,652 --> 00:14:26,952 [Miles] with Azure deals. The infrastructure relationship always outlasts the original terms. 335 00:14:27,262 --> 00:14:29,762 [Grant] The tokens run out; the SAFE doesn't. 336 00:14:30,062 --> 00:14:35,712 [Miles] The founders who signed weeks ago are building real things; the cost is still theoretical, 337 00:14:36,312 --> 00:14:43,012 [Miles] and that gap between what you signed and what you'll eventually owe-that is where this story actually lives right now. 338 00:14:43,806 --> 00:14:45,506 [Miles] So, two founders, 339 00:14:45,706 --> 00:14:46,606 [Miles] same offer, 340 00:14:46,906 --> 00:14:48,206 [Miles] completely different calls. 341 00:14:48,356 --> 00:14:49,506 [Miles] That's what stuck with me. 342 00:14:49,706 --> 00:14:50,306 [Grant] Yeah, 343 00:14:50,356 --> 00:14:52,956 [Grant] and the honest answer is neither of them was wrong. 344 00:14:53,006 --> 00:14:57,206 [Grant] It really does come down to how fast you build and how high you raise. 345 00:14:57,406 --> 00:15:00,306 [Miles] That's the thing about the uncapped SAFE structure. 346 00:15:00,506 --> 00:15:02,906 [Miles] The math only resolves at Series A. 347 00:15:03,206 --> 00:15:07,606 [Miles] Right now, every founder who signed is carrying equity they haven't priced yet. 348 00:15:07,906 --> 00:15:09,206 [Grant] According to Foundra, 349 00:15:09,256 --> 00:15:12,256 [Grant] the offer extends to the Summer 2026 batch. 350 00:15:12,190 --> 00:15:16,140 [Grant] So hundreds more founders are about to sit in the same room, 351 00:15:16,490 --> 00:15:17,340 [Grant] same silence, 352 00:15:17,740 --> 00:15:18,990 [Grant] same five seconds. 353 00:15:19,518 --> 00:15:22,118 [Miles] Five seconds before the room went electric. 354 00:15:22,568 --> 00:15:23,168 [Grant] Right. 355 00:15:23,318 --> 00:15:25,118 [Grant] And now they'll have this episode. 356 00:15:25,668 --> 00:15:26,118 [Miles] Wormley, 357 00:15:26,418 --> 00:15:30,518 [Miles] if you know a YC founder in their first year who'd actually tell their story, 358 00:15:30,768 --> 00:15:31,418 [Miles] send them our way. 359 00:15:31,868 --> 00:15:33,768 [Miles] yearone at hey matto dot com. 360 00:15:34,018 --> 00:15:36,918 [Grant] And if this episode helped you see the deal differently, 361 00:15:37,318 --> 00:15:38,018 [Grant] leave a review. 362 00:15:38,168 --> 00:15:39,618 [Grant] It matters more than you think. 363 00:15:39,918 --> 00:15:41,018 [Miles] Thanks for being here. 364 00:15:41,268 --> 00:15:42,118 [Miles] We'll see you next week. 365 00:15:42,318 --> 00:15:43,018 [Grant] See you then.