Learning Library

← Back to Library

Stripe's Bridge Deal Signals AI‑Driven Payments

Key Points

  • A new Sequoia paper reframes the AI opportunity as a multi‑trillion‑dollar market, expanding the addressable “software and services” pie from roughly $0.5 trillion to potentially $10 trillion when AI’s impact is accounted for.
  • Stripe’s $1.1 billion acquisition of Bridge brings stable‑coin payment APIs into its ecosystem, a “boring” but financially strategic move aimed at cutting the 1‑3 % fees Stripe pays to Visa and Mastercard on its trillion‑dollar‑a‑year transaction volume.
  • By enabling low‑cost, frictionless stable‑coin transactions, Bridge opens the door to micro‑payments that were previously uneconomical with cards, a capability that can power a range of emerging AI use cases and services.
  • The deal illustrates a broader contrast in the current tech landscape: solid, ROI‑driven crypto infrastructure developments versus more speculative AI investments, highlighting where investors may see tangible returns in the short term.

Full Transcript

# Stripe's Bridge Deal Signals AI‑Driven Payments **Source:** [https://www.youtube.com/watch?v=WOPFh7CdSX0](https://www.youtube.com/watch?v=WOPFh7CdSX0) **Duration:** 00:10:55 ## Summary - A new Sequoia paper reframes the AI opportunity as a multi‑trillion‑dollar market, expanding the addressable “software and services” pie from roughly $0.5 trillion to potentially $10 trillion when AI’s impact is accounted for. - Stripe’s $1.1 billion acquisition of Bridge brings stable‑coin payment APIs into its ecosystem, a “boring” but financially strategic move aimed at cutting the 1‑3 % fees Stripe pays to Visa and Mastercard on its trillion‑dollar‑a‑year transaction volume. - By enabling low‑cost, frictionless stable‑coin transactions, Bridge opens the door to micro‑payments that were previously uneconomical with cards, a capability that can power a range of emerging AI use cases and services. - The deal illustrates a broader contrast in the current tech landscape: solid, ROI‑driven crypto infrastructure developments versus more speculative AI investments, highlighting where investors may see tangible returns in the short term. ## Sections - [00:00:00](https://www.youtube.com/watch?v=WOPFh7CdSX0&t=0s) **AI Market Boom Frames Stripe Deal** - The speaker links recent AI headlines to a Sequoia paper that inflates the software‑and‑services market from half a trillion to potentially ten trillion dollars, and highlights Stripe’s $1.1 billion acquisition of crypto payment platform Bridge as a non‑AI move that will become AI‑relevant within the next year. ## Full Transcript
0:00I want to take two or three pieces of 0:01news that all broken the last 24 hours 0:04in the tech space in Ai and I want to 0:07put them together in the context of a 0:09larger frame around where AI is going so 0:12the frame here is the Sequoia paper that 0:15came out that talks about the fact that 0:17we're moving into a world where we're no 0:19longer looking just at the pure software 0:24Market which could be valued at maybe 0:27half a trillion dollars but we're 0:29actually looking at A1 trillion do 0:31software and services Market as 0:33addressable if we factor in AI because 0:35AI is so effective at services so 0:38basically if you look at like where do 0:39people get the money back on these 0:41Investments it's their them thinking 0:43they have this pie that's grown much 0:45much larger and it's now a 10 trillion 0:47dollar pie to play with rather than a 0:49half a trillion dollar piie to play with 0:51all of these are very large numbers but 0:53when you're looking at investing as much 0:55as people are investing in AI the larger 0:58number does matter because you do have 1:00to get that Roi back okay with that 1:03context in that frame what developments 1:06matter over the last couple of days 1:08number one you will be surprised by this 1:11but number one is a boring business so 1:13so to speak that is not necessarily 1:15directly AI related but will tie into AI 1:18over the next 12 months that is Stripes 1:21acquisition of bridge so stripe bought 1:24bridge for a rumored $1.1 1:27billion and stripe of course is in the 1:30payment space bridge is also in the 1:31payment space but it's in crypto and 1:33it's in stable coins and providing apis 1:35for stable coins specifically the reason 1:38why this is interesting is because this 1:40is a classic example of what I would 1:42call a stripe type acquisition where it 1:45makes a lot of financial sense there's a 1:47very hard use case in Roi and that's not 1:49typical for crypto and it's really 1:52interesting to be in a world where if 1:55you're looking at the range of news 1:57crypto is the thing with the stable use 1:59case the boring business case and it's 2:01the AI stuff that's more 2:03speculative and in this situation what 2:06stripe is looking to do is essentially 2:08to undercut the power that the card 2:10monopolies have so Visa Mastercard 2:13they're charging stripe for handling 2:16payments and it is non-trivial at 2:18strip's volume stripe handles over a 2:20trillion dollars in transactions every 2:22year and so if you're paying 1 to 3% on 2:25that that really adds up and so what 2:27bridge does is it provides you the 2:30option to start to move some of that 2:32payment volume to stable coins which in 2:34turn gives you a frictionless payment 2:37method that unlocks a lot of really 2:39interesting use cases yes including 2:41potentially AI use cases so for 2:44example the underlying unit economics of 2:47card volume make it hard to do micro 2:51payments and one of the real challenges 2:53on the web is that we have sort of a 2:55lower limit around $5 where it doesn't 2:58make sense to charge less than that 3:01except in really extraordinary 3:02circumstances and that's why games will 3:04often charge you in smaller increments 3:06but not really charge your wallet and 3:07they you'll just recharge your wallet 3:09periodically and so we don't really have 3:12a good you know answer for micro 3:14payments Bridge or something like Bridge 3:17is a way to handle that because it could 3:19be potentially really frictionless to do 3:21micro payments another example uh just 3:25fundamentally is that you could do 3:28something like a wallet experience where 3:30you move your currency into stable coins 3:33with stripe you could pay on stripe as 3:37much as you want with zero transaction 3:39fees and stripe could pass those savings 3:42to you and stripe could then know you 3:45had the cash because like it's in the 3:47wallet in the account and the entire 3:51system would be sort of a closed loop 3:53with stripe driving it so there's 3:54there's like more margin for stripe to 3:56capture there there's savings to pass on 3:57to Merchants savings to pass on to 3:59customers Al 4:00Ely that was the really hard-headed use 4:02case for bridge uh I noticed they have 4:05API documents 4:06Etc and that's very strip like I'm 4:09really curious to see where this goes 4:11because stripe has definitely been very 4:13hard-headed when it comes to how they 4:14handle payments and so if they're buying 4:16this they clearly see a use case that 4:17they're going to go somewhere with 4:18that's the boring news here's the the 4:22more speculative news and how this 4:24relates to this larger 10 trillion doll 4:26software frame so we have two examples 4:30of agentic reasoning and agentic search 4:33and agentic use cases for AI and just to 4:36remind you like agentic means like 4:38essentially the AI is driving the 4:40computer for you the AI will do a task 4:42independently and come back and you 4:45don't have to tell it what to do every 4:47single time right you don't have to give 4:48it micro tasks and micro micro work 4:51along the way to accomplish a larger 4:53task release one is anthropic actually 4:57releasing a larger 5:00SEO search agent experience and by 5:02larger I don't mean that it's like 5:03bigger on the screen I mean that it's 5:05more in-depth I mean that you can give 5:08the agent more to do and that is not 5:11coincidentally tied to the release of 5:14Claude 3.5 which is the other major 5:16release that I'm going to talk about so 5:18Claud 3.5 is specifically trained on 5:22agentic use cases and they actually 5:23released that as a bench and they talked 5:25about how much better the new 3.5 Sonet 5:28is than the existing one at agentic use 5:31cases and when you think about 5:34it what Claud is trying to do from a 5:37sort of strategic perspective is unlock 5:39agentic use cases for developers and so 5:41what they highlighted was you know Reet 5:43using uh the new claw agent for computer 5:47use to do uat 5:50Etc but what people's imagination got 5:52caught by it wasn't necessarily 5:54developers it was actually consumer use 5:57cases that I saw I saw someone order 5:58lunch with their Claude 6:01agent just yesterday now it was hard 6:03because Claude doesn't really want you 6:04to do that they don't want the agent 6:06making purchases there's obviously big 6:07liabilities there imagine the agent 6:09racking up thousands at like a luxury 6:12store without your consent Etc it could 6:13be a real problem but the idea of an 6:17agent that drives your screen and uses 6:20your screen as an API to see the world 6:24that's something that has consumer 6:26implications as well now this is very 6:28token heavy so like it's like 2 billion 6:30and change tokens for 15 20 minutes of 6:33of screen time you do have to bail it 6:35out it get stuck in a ditch a lot I 6:37joked yesterday about how distractable 6:39it is it is definitely 6:41distractable um but the point is like 6:44this is one of those situations where if 6:46you walk back to stripe and Bridge 6:48stripe and Bridge like the crypto use 6:49case it's actually very hard-headed in 6:51strip's case whereas with anthropic it's 6:54clearly in experimental mode they said 6:55it's in experimental mode it is a 6:58technology looking for use cases and we 7:00have guesses at what those use cases 7:02would be but we're going to have to 7:04discover them and perplexity is taking 7:07that same underlying technology and 7:09working it into the perplexity search 7:12and actually making it useful for 7:14extended search query so it can 7:16independent the agent independently sefs 7:17the web for you and answers your query 7:20and it's already wrapped into perplexity 7:22Pro so if you're perplexity Pro user you 7:23already have 7:25it and what I find interesting is is 7:28that in this situation 7:30the technology is finding use cases 7:34within existing pieces of AI software 7:36faster because they're more narrowly 7:39defined and so perplexity just has to 7:41handle search they don't have to handle 7:43a mouse moving around your screen they 7:45don't have to handle keyboard inputs 7:47it's just a search agent and that 7:50limitation that like lockdown of what 7:53the scope of the agent can do makes it 7:55more useful and I am wondering as we 7:58look back at this like a 10 trillion 8:00play if essentially what we're going to 8:03see is three different kinds of plays 8:05we're going to see infr plays where you 8:08have folks like stripe or bridge that 8:10are innovating on infrastructure to 8:12enable the kind of underlying payments 8:16software infrastructure that you need to 8:18have an AI economy and I think the 8:21bridge fits into that like you will need 8:23micro payments to make AI work well 8:26because you have a lot of micro 8:27interactions 8:29and then on top of that you're going to 8:32have very defined and narrow use case 8:35Innovation and that's really the 8:36perplexity space like we do search that 8:38is what we do and I think that is also 8:41coming soon because I think you have 8:43immediate applications for that core 8:45Innovation and then at the very end of 8:47the of the value chain you have this 8:49sort of core speculative Innovation 8:51that's coming from the core big four or 8:54five language model shops and anthropic 8:57is one of them and when they do things 8:59like release a general purpose computer 9:02user 9:04experiment it's going to be a bit before 9:07we figure out exactly how to use that I 9:08think it's coming I think it's probably 9:10coming sometime we we'll see some 9:11usefulness uh and some scaled 9:13application for it sometime next year in 9:162025 it is not something that I would 9:18say is close to ready for prime time 9:23today in October 2024 and I realize that 9:25saying that over a period of months 9:27makes me seem crazy because this is all 9:31moving so fast and like you fast forward 9:336 months and normally you're you're used 9:34to seeing like gradual progress because 9:36that's what we've had it will feel much 9:38faster than that but still it is very 9:41speculative today it is very speculative 9:45today so if I wrap all of that back to 9:48the AOA frame and kind of where we 9:50started out 9:52ultimately what will capture the 9:55services Market that 10 trillion Doll 9:57Market that SEO is painting a picture of 9:59is specific use cases that enable people 10:03to solve their problems dramatically 10:05faster I actually think perplexity is 10:07positioned really well here I think 10:09stripe is positioned well just to 10:11capture underlying uh margin from 10:13payments volume they're just moving into 10:15into a place where they can solidify 10:17that position and solidify their scale 10:19over time that's what the bridge 10:20acquisition was about 10:22um and ultimately anthropic is 10:25positioned as a utility anthropic is 10:28positioned as a utility that that 10:29enables that scale by charging per token 10:32and so really the key underlying factor 10:34for them to get more adoption of some of 10:36these Technologies is to bring down the 10:38cost per token which is exactly what 10:39we've seen them doing I know this has 10:41been in-depth at length and it's been a 10:44little bit of a strategic frame but I 10:46like to step back and put the news in 10:47that perspective sometimes because I 10:49think it's really important so we don't 10:50lose our way hope you enjoyed it cheers