WEBVTT
1
00:00:11.160 --> 00:00:15.740
Hello, everyone. Welcome.
2
00:00:16.560 --> 00:00:18.960
Happy Sessions. I'm Maia,
3
00:00:19.440 --> 00:00:23.280
chief revenue Officer of AI at Stripe. Twenty years ago,
4
00:00:24.000 --> 00:00:25.100
I founded Urban Escapes,
5
00:00:25.700 --> 00:00:29.500
an ecommerce travel and experience company-kayaking,
6
00:00:30.500 --> 00:00:35.120
whitewater rafting, skydiving. That's me in white, 2007.
7
00:00:36.260 --> 00:00:41.020
And man, things were different back then. Now, this is before WeWork.
8
00:00:41.640 --> 00:00:45.320
So we thought we were being really clever because we rented out a conference
9
00:00:45.360 --> 00:00:49.760
room in a Midtown corporate building to save on rent. Now, back then,
10
00:00:50.320 --> 00:00:54.760
going from idea to a website where customers could pay online took
11
00:00:55.140 --> 00:00:55.973
months.
12
00:00:57.440 --> 00:01:01.580
I think about that experience a lot because if I was starting Urban Escapes
13
00:01:01.640 --> 00:01:05.680
today, I would basically do everything differently.
14
00:01:07.000 --> 00:01:11.120
And I know that because of you all. At Stripe,
15
00:01:11.580 --> 00:01:14.300
we work with the fastest growing AI companies in the world,
16
00:01:15.120 --> 00:01:19.660
many of them in this room. And we have a front row seat to what's working,
17
00:01:20.240 --> 00:01:24.560
what's not, and how the best companies scale. Today,
18
00:01:25.160 --> 00:01:28.700
I want to share some of those insights because the playbook for how to run a
19
00:01:28.740 --> 00:01:33.340
company is being rewritten faster than any of us can wrap our heads around.
20
00:01:34.380 --> 00:01:38.020
So what are the four patterns we're seeing? One:
21
00:01:38.620 --> 00:01:42.780
the top AI companies build faster. I think we're all feeling that.
22
00:01:44.060 --> 00:01:46.620
Two: they sell globally by default.
23
00:01:47.580 --> 00:01:52.000
Your day one market is now the whole world. And three:
24
00:01:52.700 --> 00:01:55.520
pricing is evolving faster than most of us realize.
25
00:01:56.360 --> 00:02:00.200
Subscriptions were the breakthrough model 10 years ago, but not anymore.
26
00:02:01.500 --> 00:02:05.400
And four: they adopt to new go-to-market motions and fast.
27
00:02:06.240 --> 00:02:10.940
Building enterprise sales used to take a decade. Now it happens in year one.
28
00:02:12.200 --> 00:02:15.060
We'll dig into each of these, but first, the big picture.
29
00:02:16.420 --> 00:02:20.980
We saw the top AI companies grow by 120% in
30
00:02:21.270 --> 00:02:24.360
2025. So far in 26,
31
00:02:25.200 --> 00:02:29.540
by 175%. That's nearly tripling in one year.
32
00:02:31.000 --> 00:02:33.400
The growth is not slowing down.
It's accelerating.
33
00:02:34.640 --> 00:02:36.860
And the top ones are even more mind-blowing.
34
00:02:37.720 --> 00:02:40.800
Lovable reported $100 million in revenue in eight months.
35
00:02:41.740 --> 00:02:44.220
And then eight months later, $400 million.
36
00:02:45.360 --> 00:02:49.520
Cursor announced they hit a $1 billion dollar run rate in under two years.
37
00:02:50.740 --> 00:02:52.480
Three months later, $2 billion.
38
00:02:53.960 --> 00:02:58.760
Anthropic shared going from $0 in January 23 to $1 billion just two years
39
00:02:58.800 --> 00:03:02.420
later. And now they're at a $30 billion run rate.
40
00:03:03.580 --> 00:03:06.920
It is wild. And it's not just B2B.
41
00:03:08.100 --> 00:03:12.460
Our Link data shows consumer adoption of AI has doubled from just under six
42
00:03:12.600 --> 00:03:16.220
million to over 14 million consumers in just one year.
43
00:03:17.620 --> 00:03:18.880
And they're spending more.
44
00:03:19.920 --> 00:03:24.420
Top Link buyers now spend $371 a month on AI.
45
00:03:24.980 --> 00:03:29.680
That's up from $140 just one year ago. To put that into perspective,
46
00:03:30.360 --> 00:03:34.840
that's more than the average American spends on internet, streaming,
47
00:03:35.280 --> 00:03:36.600
and phone service combined.
48
00:03:37.880 --> 00:03:41.160
People aren't treating AI like a streaming service they might cancel.
49
00:03:42.020 --> 00:03:46.260
They're treating it like a critical utility they can't live without.
All right,
50
00:03:46.320 --> 00:03:50.280
so we have 175% revenue growth, adoption doubling,
51
00:03:51.060 --> 00:03:54.180
billion dollar run rates in months. What's behind all this?
52
00:03:55.400 --> 00:03:58.120
Let's dig into the four patterns, starting with speed.
53
00:03:59.640 --> 00:04:03.680
When I first started Urban Escapes, getting to a prototype took months.
54
00:04:04.520 --> 00:04:05.720
And, okay, I'm not kidding here.
55
00:04:06.180 --> 00:04:09.880
Building a shopping cart was so complex that in the first iteration,
56
00:04:10.240 --> 00:04:14.600
when you hit "buy," we would tell the customers to mail a check to my apartment.
57
00:04:15.600 --> 00:04:19.060
And they did-like a lot of them-which is really unsafe when you think back to
58
00:04:19.560 --> 00:04:21.160
it. No,
59
00:04:21.460 --> 00:04:24.920
just like you no longer need months of engineering to accept a payment.
60
00:04:25.740 --> 00:04:28.220
You don't need months of engineering to ship software anymore.
61
00:04:29.380 --> 00:04:33.160
And this chart really shows when that clicked. So as late as 2024,
62
00:04:33.860 --> 00:04:35.300
iOS app releases were declining.
63
00:04:36.280 --> 00:04:41.160
Then agentic coding tools hit and app launches jumped 24%
64
00:04:41.940 --> 00:04:46.580
month-over-month.
And Delaware incorporations followed the same pattern.
65
00:04:47.500 --> 00:04:51.920
You can see the huge spike when these tools went mainstream. Now,
66
00:04:52.260 --> 00:04:55.180
everyone's talking about nontechnical founders vibe coding.
67
00:04:56.120 --> 00:05:00.760
But here's something that's really surprising. As startup creation has grown,
68
00:05:01.660 --> 00:05:06.320
the share of technical founders has increased by seven points in just one year.
69
00:05:07.700 --> 00:05:12.540
That means that a technical founder with AI tools can now do in days what
70
00:05:12.640 --> 00:05:16.200
used to take a team months. These are not side projects.
71
00:05:16.680 --> 00:05:19.660
These are real businesses. In fact,
72
00:05:20.280 --> 00:05:24.260
since embedding payments into developer platforms, like Replit or Vercel,
73
00:05:25.200 --> 00:05:29.360
we've seen each monthly cohort-which you see in these lines-go from idea to
74
00:05:29.460 --> 00:05:31.300
first charge faster and faster.
75
00:05:32.420 --> 00:05:36.660
And you can see it now takes builders less than six weeks to get their first
76
00:05:36.680 --> 00:05:39.200
paying customer. And I can promise you,
77
00:05:39.460 --> 00:05:42.300
none of these founders are asking people to send checks to their apartment.
78
00:05:43.860 --> 00:05:46.220
All right, so what does this mean for you? Now,
79
00:05:46.340 --> 00:05:49.380
many of you are already doing these things, but for those that aren't,
80
00:05:49.780 --> 00:05:52.080
here's the playbook that's emerging. First,
81
00:05:52.640 --> 00:05:55.080
obsess over developer productivity.
82
00:05:55.840 --> 00:05:58.740
Collapsing build times is a competitive advantage.
83
00:05:59.680 --> 00:06:04.140
The companies moving fastest, treat it as a priority, not as an afterthought.
84
00:06:05.920 --> 00:06:09.480
Second, build, sell, and iterate at the same time.
85
00:06:10.240 --> 00:06:13.680
That linear playbook where you could finish the product and then go find your
86
00:06:13.700 --> 00:06:15.280
customers-too slow.
87
00:06:16.500 --> 00:06:20.980
The best teams do all three in parallel. And then third:
88
00:06:21.340 --> 00:06:23.700
when you're posed with the "build versus buy" question,
89
00:06:24.100 --> 00:06:26.800
the fastest moving teams, build and buy.
90
00:06:27.680 --> 00:06:32.440
So focus your resources on what makes you most differentiated and buy the
91
00:06:32.500 --> 00:06:36.300
infrastructure and building blocks to enable that. Now,
92
00:06:36.640 --> 00:06:40.960
getting to market fast is one thing, but where you sell matters just as much.
93
00:06:41.840 --> 00:06:46.540
With Urban Escapes, expansion was city by city, New York first, then Philly,
94
00:06:46.700 --> 00:06:50.020
Boston, DC.
And SaaS had the same logic, right?
95
00:06:50.340 --> 00:06:52.220
You'd explore selling in other markets,
96
00:06:52.560 --> 00:06:55.720
but the focus was really on nailing your home market. And then,
97
00:06:56.340 --> 00:06:57.600
once you had the playbook figured out,
98
00:06:57.900 --> 00:07:02.220
you'd hire a GM in either London or Dublin, you'd open an office,
99
00:07:02.540 --> 00:07:04.320
and then congrats, you were international.
100
00:07:05.740 --> 00:07:09.900
That playbook just doesn't work anymore, and the data confirms it.
101
00:07:10.760 --> 00:07:11.593
A few years ago,
102
00:07:12.120 --> 00:07:16.580
the fastest growing SaaS companies would reach 25 countries in Year 1
103
00:07:17.320 --> 00:07:21.840
and 50 by Year 3. But AI companies, in 2025,
104
00:07:22.680 --> 00:07:25.820
we saw 42 countries in Year 1. And get this,
105
00:07:27.100 --> 00:07:29.340
120 countries by Year 3.
106
00:07:29.980 --> 00:07:34.640
That means Kazakhstan is now showing up on the list for many AI
107
00:07:34.680 --> 00:07:37.300
companies. That was not on my bingo card.
108
00:07:38.940 --> 00:07:41.840
And this isn't about selling to one person in these countries.
109
00:07:42.360 --> 00:07:46.000
There's real revenue coming from them. Gamma is a great example.
110
00:07:46.680 --> 00:07:50.800
It's an AI-powered slide and website builder.
They reported $100 hundred million
111
00:07:50.920 --> 00:07:54.460
in revenue in their first year. And although they're based in SF,
112
00:07:55.120 --> 00:07:59.640
the majority of their revenue comes from outside the US. Now,
113
00:07:59.700 --> 00:08:02.180
that's not the exception anymore. It's actually the rule.
114
00:08:03.360 --> 00:08:04.640
Across top AI companies,
115
00:08:05.540 --> 00:08:08.520
48% of revenue comes from outside their home market.
116
00:08:09.680 --> 00:08:13.180
That's nearly half of every dollar. Three years ago,
117
00:08:13.320 --> 00:08:17.980
that number was only 33%. So global revenue isn't a bonus anymore.
118
00:08:19.520 --> 00:08:21.560
It's just the baseline. Now,
119
00:08:21.640 --> 00:08:25.240
let's take a look at which markets are spending the most on AI tools on Stripe.
120
00:08:26.260 --> 00:08:31.040
Not surprisingly, we're seeing strong spend in markets with high GDP,
121
00:08:31.520 --> 00:08:33.440
like the US, Japan, Germany.
122
00:08:34.320 --> 00:08:38.680
And then there are some emerging countries that are coming up as hotspots like
123
00:08:38.960 --> 00:08:41.280
South Korea, Brazil, and India.
124
00:08:42.480 --> 00:08:46.040
The demand is definitely there. So how do you actually capture it?
125
00:08:46.980 --> 00:08:48.920
By meeting your customers where they are,
126
00:08:49.440 --> 00:08:54.320
which means offering local currencies and local payment methods.
Okay, get this.
127
00:08:55.070 --> 00:09:00.000
Localized pricing drives 18% higher cross-border
128
00:09:00.080 --> 00:09:04.000
revenue. And when you add at least one local payment method,
129
00:09:04.400 --> 00:09:09.160
we see more than a 7% conversion uplift. With these stats,
130
00:09:09.600 --> 00:09:13.640
it's kind of crazy not to be doing this. A good check:
131
00:09:14.180 --> 00:09:16.760
put yourself in the shoes of a customer in Brazil.
132
00:09:18.080 --> 00:09:21.680
Can they pay in reais with Pix? If your answer isn't,
133
00:09:21.680 --> 00:09:24.400
"Of course," you're leaving money on the table.
134
00:09:25.440 --> 00:09:29.280
The playbook for this one: three ways to pressure test your global readiness.
135
00:09:29.600 --> 00:09:30.433
Number one,
136
00:09:30.680 --> 00:09:34.360
localize prices and make sure you have local payment methods in key markets.
137
00:09:35.280 --> 00:09:39.600
Second: automate tax collection because none of us want to get in trouble.
138
00:09:40.200 --> 00:09:41.033
And third:
139
00:09:41.040 --> 00:09:44.560
track revenue and conversion by country and obsess over optimizing it.
140
00:09:46.440 --> 00:09:48.400
All right. Onto pricing, one of my favorite topics.
141
00:09:49.840 --> 00:09:54.280
Because the old models just aren't working, and AI has upended what we know.
142
00:09:55.160 --> 00:09:59.800
So urban escapes-pricing, way more straightforward.
Skydiving: $300.
143
00:10:00.320 --> 00:10:03.740
Rafting: $100. Same experience, same price-done.
144
00:10:04.940 --> 00:10:08.620
And the SaaS corollary was similar with seats-same price for everyone.
145
00:10:10.320 --> 00:10:13.160
But as we all know, AI pricing is way more complicated.
146
00:10:14.160 --> 00:10:17.960
Value is elastic, and so is cost. Now,
147
00:10:18.100 --> 00:10:19.880
everyone might use the same AI tool,
148
00:10:20.480 --> 00:10:24.820
but they're having vastly different experiences. Let's take two users:
149
00:10:25.700 --> 00:10:29.980
one is an engineer who sets up a bunch of agents before bed and wakes up with
150
00:10:30.020 --> 00:10:33.160
code ready for a review. And the other is my mom,
151
00:10:33.760 --> 00:10:37.680
who last week told me very proudly-and I'm very proud of her too-that she
152
00:10:37.740 --> 00:10:39.820
replaced Safari with ChatGPT on her phone.
153
00:10:40.960 --> 00:10:43.720
One is creating tremendous engineering output,
154
00:10:44.420 --> 00:10:47.500
and the other has a very upgraded search experience.
155
00:10:48.420 --> 00:10:52.940
The value and the cost are different, and so the pricing should be too.
156
00:10:54.520 --> 00:10:57.200
Now, this change in value isn't new.
157
00:10:58.140 --> 00:11:02.520
Every major technology shift has repriced software.
For on-prem,
158
00:11:03.080 --> 00:11:06.220
you paid once for all the code written up until installation.
159
00:11:06.960 --> 00:11:11.380
So a one-time license made sense. Then cloud changed everything.
160
00:11:12.300 --> 00:11:14.380
Software became continuous and autoupdating,
161
00:11:15.160 --> 00:11:18.560
so recurring subscriptions made sense. Okay,
162
00:11:18.620 --> 00:11:22.420
so now is when you'd expect me to tell you the exact right pricing model for AI.
163
00:11:23.580 --> 00:11:27.660
Good news, bad news. Bad news first. We're still in the early innings,
164
00:11:28.240 --> 00:11:30.080
and the gold standard hasn't been stamped yet.
165
00:11:31.240 --> 00:11:35.480
But the good news is a clear pattern is emerging. So we'll take Replit,
166
00:11:35.880 --> 00:11:38.100
a developer tools company that's been around for almost a decade.
167
00:11:38.980 --> 00:11:41.560
When agentic coding took off, they pivoted their product,
168
00:11:42.120 --> 00:11:44.780
rethought their pricing, and hockey-sticked their growth.
169
00:11:45.700 --> 00:11:49.620
They started with flat subscriptions and then layered in credits as AI became
170
00:11:49.640 --> 00:11:50.473
core to the product.
171
00:11:51.100 --> 00:11:55.620
Subscriptions for predictability and then usage for value capture. The result?
172
00:11:55.680 --> 00:11:59.280
They're targeting a billion dollars in ARR in Year 1 this year,
173
00:11:59.920 --> 00:12:04.520
and they're not alone.
Usage-based pricing is fundamental to how AI delivers
174
00:12:04.540 --> 00:12:05.373
value.
175
00:12:05.740 --> 00:12:09.860
Two in three companies on the Forbes AI 50 have some form of usage-based
176
00:12:09.880 --> 00:12:13.800
pricing, and a majority of those companies are running a hybrid model,
177
00:12:14.500 --> 00:12:18.340
a subscription that anchors the relationship layered with credits that scale
178
00:12:18.360 --> 00:12:22.980
with value. Now, knowing you need to be usage-based is one thing.
179
00:12:23.520 --> 00:12:27.360
Getting it right is another. There's three steps to add to our playbook here.
180
00:12:28.400 --> 00:12:32.440
First: price in your customer's language. Developers think in tokens.
181
00:12:33.120 --> 00:12:35.100
Enterprises, historically in seats,
182
00:12:35.420 --> 00:12:38.120
but they're getting increasingly much more comfortable with consumption.
183
00:12:39.180 --> 00:12:43.760
Match your pricing to the way your customers experience value. Second:
184
00:12:44.280 --> 00:12:47.480
show customers what they're consuming before they get the bill.
185
00:12:48.340 --> 00:12:51.900
Who has ever opened their electricity bill and said, "Well,
186
00:12:52.080 --> 00:12:55.980
that's exactly what I was expecting"? It's always a surprise,
187
00:12:56.500 --> 00:12:57.840
and that's not a great experience.
188
00:12:58.880 --> 00:13:02.040
Real-time visibility isn't just nice to have.
189
00:13:02.720 --> 00:13:05.860
It's a critical part of your product experience and it's how you prevent churn.
190
00:13:06.960 --> 00:13:07.680
So practically,
191
00:13:07.680 --> 00:13:11.000
this means finding ways to tell your customers how much they're using while
192
00:13:11.020 --> 00:13:14.640
they're using. And then third: sell credits, not cost.
193
00:13:15.160 --> 00:13:20.040
Have your customers exchange money for credits once so they're thinking about
194
00:13:20.080 --> 00:13:24.740
value, not dollars and cents every time they use your product. Okay,
195
00:13:24.820 --> 00:13:25.653
onto the last pattern.
196
00:13:27.060 --> 00:13:32.040
Companies are adapting their go-to-market motions much faster. At Urban Escapes,
197
00:13:32.580 --> 00:13:35.180
we started with consumers first, and then believe it or not,
198
00:13:35.360 --> 00:13:38.700
built out an enterprise business because the Googles and Facebooks of the world
199
00:13:39.180 --> 00:13:43.140
loved taking their clients on adventures. And that's been the norm.
200
00:13:43.660 --> 00:13:46.140
You start product-led, startups find you organically,
201
00:13:46.600 --> 00:13:50.260
and then you layer in enterprise sales years later-once you prove a product
202
00:13:50.280 --> 00:13:54.180
market fit at scale. I mean, Stripe followed the same arc.
203
00:13:54.840 --> 00:13:58.520
I was our first head of enterprise product, and that was just three years ago.
204
00:14:00.160 --> 00:14:03.840
Now, every AI founder I talk to is hiring a CRO.
205
00:14:05.460 --> 00:14:09.280
Building out an enterprise motion isn't a "three years from now" problem.
206
00:14:10.280 --> 00:14:14.080
People are doing it now. Cursor, the AI code editor,
207
00:14:14.400 --> 00:14:16.640
launched in '23 as self-serve for developers.
208
00:14:17.540 --> 00:14:20.180
Then they layered in sales-led to land enterprise contracts.
209
00:14:21.140 --> 00:14:22.160
Just a couple years later,
210
00:14:22.380 --> 00:14:25.960
they now have an incredibly impressive enterprise business that would have taken
211
00:14:26.000 --> 00:14:30.940
other companies decades to build. But that's not the only shift. No,
212
00:14:31.360 --> 00:14:35.260
that would be way too easy. Entirely new motions are emerging too.
213
00:14:36.340 --> 00:14:40.620
For one, channel sales have become a huge part of the AI go-to-market motion.
214
00:14:41.800 --> 00:14:45.680
A lot of people buy OpenAI or Anthropic through their cloud provider today.
215
00:14:46.820 --> 00:14:51.700
And now the AI companies are also building their own marketplaces so others can
216
00:14:51.740 --> 00:14:56.730
do the same. Channel sales is getting to be a much bigger part of the equation.
217
00:14:58.230 --> 00:15:02.050
And the billion or really trillion dollar question on everyone's mind:
218
00:15:02.050 --> 00:15:04.790
"How do we all sell to our newest buyer,
219
00:15:04.890 --> 00:15:09.590
the agent?"
We've spent decades designing pricing
220
00:15:09.710 --> 00:15:12.490
around human psychology, anchoring on
221
00:15:14.710 --> 00:15:15.960
"good," "better," "best,"-$9.99 instead of $10.
222
00:15:17.510 --> 00:15:20.270
Agents don't care about any of that.
223
00:15:21.830 --> 00:15:24.640
This chart is one of the clearest signals. In 2025,
224
00:15:25.710 --> 00:15:28.730
agent traffic to Stripe Docs 10x'ed in a single year.
225
00:15:29.730 --> 00:15:33.410
The purple line is humans. The pink line is agents.
226
00:15:34.510 --> 00:15:37.390
And you can see they're converging. By the end of this year,
227
00:15:38.110 --> 00:15:42.230
agents will read more Stripe Docs than humans will. Now,
228
00:15:42.690 --> 00:15:44.970
here's where a lot of companies underestimate the work.
229
00:15:46.210 --> 00:15:48.090
Go-to-market is the tip of the iceberg.
230
00:15:48.970 --> 00:15:51.010
Everything underneath it has to move with it.
231
00:15:51.970 --> 00:15:55.050
Adding a new go-to-market motion touches everything.
232
00:15:55.110 --> 00:15:59.590
Your product looks different for each motion because the onboarding flow is
233
00:15:59.630 --> 00:16:03.030
different. Pricing will have different nuances.
234
00:16:03.550 --> 00:16:06.670
You have commits for enterprises, sticker pricing for product-led.
235
00:16:06.800 --> 00:16:09.850
And then your org has to be built for all this.
236
00:16:10.610 --> 00:16:14.870
We all know that an enterprise sale is very different than how you do customer
237
00:16:14.910 --> 00:16:18.690
support for the long tail.
It's not just a go-to-market change.
238
00:16:19.230 --> 00:16:23.130
We're really talking about your whole entire business model. Now,
239
00:16:23.630 --> 00:16:28.150
most companies build revenue infrastructure the same way: one piece at a time.
240
00:16:28.890 --> 00:16:33.270
A billing tool here, a tax vendor there, payments bolted on when you need to.
241
00:16:34.530 --> 00:16:38.490
And when you're growing slowly, a Frankenstein stack is much more manageable.
242
00:16:39.130 --> 00:16:42.690
But when you're compressing a decade of go-to-market evolution into months,
243
00:16:43.430 --> 00:16:47.890
every single seam becomes liability. Here's how I think about it:
244
00:16:48.550 --> 00:16:50.890
your customers, they're channel-agnostic.
245
00:16:51.350 --> 00:16:53.810
The same customer might find you self-serve,
246
00:16:54.150 --> 00:16:57.870
but then graduate to enterprise and then have an agent implement a new product.
247
00:16:58.650 --> 00:17:00.850
They move across motions without thinking about it,
248
00:17:01.610 --> 00:17:04.950
and they certainly don't think of themselves as a channel. Now,
249
00:17:05.170 --> 00:17:09.010
that means your systems have to move with them. The playbook for this one?
250
00:17:09.770 --> 00:17:11.830
First, have a clear graduation process.
251
00:17:11.890 --> 00:17:15.870
Know when self-serve becomes enterprise and how pricing changes with it.
252
00:17:16.930 --> 00:17:21.390
Second: build unified systems. That means one customer object,
253
00:17:21.690 --> 00:17:25.430
one product catalog, and one data model. And third,
254
00:17:25.690 --> 00:17:29.710
design for agent readiness, where an agent could discover, evaluate,
255
00:17:29.910 --> 00:17:34.110
and activate your product without a human-in-the-loop. All right.
256
00:17:34.510 --> 00:17:38.390
We've covered a lot. And one thing is for sure: the playbook is being rewritten,
257
00:17:38.980 --> 00:17:42.270
from how you build on price to where you sell and go to market.
258
00:17:43.130 --> 00:17:44.930
The question we all have to ask ourselves is:
259
00:17:45.790 --> 00:17:50.790
"Are you driving this change or are you reacting to it?" And now I'm really
260
00:17:50.830 --> 00:17:54.290
excited to introduce someone who is very much in the driver's seat,
261
00:17:54.710 --> 00:17:58.390
both as a founder of scaling his own company and as a platform enabling
262
00:17:58.530 --> 00:18:01.830
thousands of builders to do the same. Let's welcome on stage Guillermo,
263
00:18:02.410 --> 00:18:04.010
CEO and founder of Vercel.
264
00:18:12.760 --> 00:18:14.640
Hello. Thanks for being here.
265
00:18:14.720 --> 00:18:15.553
Thanks for having me.
266
00:18:15.640 --> 00:18:16.390
Great to have you.
267
00:18:16.390 --> 00:18:17.223
Great presentation.
268
00:18:18.050 --> 00:18:19.810
Thank you. You inspired a lot of it.
269
00:18:20.010 --> 00:18:20.843
Thank you.
270
00:18:20.930 --> 00:18:24.170
Okay. So we talked about a lot of things, but one thing's for sure:
271
00:18:24.650 --> 00:18:28.730
the speed and the new types of builders that are emerging because of AI is
272
00:18:28.750 --> 00:18:31.650
pretty wild. And I think you have just been at the forefront of that.
273
00:18:31.730 --> 00:18:34.630
So share a little bit about how you think about it and what you're seeing.
274
00:18:35.150 --> 00:18:39.730
Well, you talked about your idea started out with you wanted to create a website
275
00:18:40.210 --> 00:18:41.043
and make money.
276
00:18:41.610 --> 00:18:46.290
And one of the obsessions when I started Vercel was how quickly can you go from
277
00:18:46.470 --> 00:18:50.750
idea to software that's global, that's deployed,
278
00:18:51.010 --> 00:18:54.210
that's secure, that's performing everywhere on the planet.
279
00:18:54.890 --> 00:18:56.450
And the mission has continued, right?
280
00:18:56.530 --> 00:18:59.430
It started out with the developer experience. In fact,
281
00:18:59.850 --> 00:19:03.330
one of the early metaphors that one of my first angel investors used for Vercel
282
00:19:03.470 --> 00:19:05.750
was it's a Stripe for deployments.
283
00:19:06.470 --> 00:19:11.450
And the thing that's changed really that you kind of spoke to is nowadays I find
284
00:19:11.470 --> 00:19:16.450
myself thinking more about the agent as my customer and the agentic
285
00:19:16.510 --> 00:19:19.830
developer experience. So the mission of Vercel has not really changed.
286
00:19:20.070 --> 00:19:23.430
It's evolved. And instead of building just developer infrastructure,
287
00:19:23.510 --> 00:19:27.690
we're building agentic infrastructure. And I think of the agent as my customer.
288
00:19:28.330 --> 00:19:32.890
I'm sweating the details of the error messages for the agent as my customer.
289
00:19:33.790 --> 00:19:37.670
I think about the agent getting stuck. You were talking about Stripe,
290
00:19:37.790 --> 00:19:40.470
seeing all of this traffic coming from agents.
291
00:19:40.890 --> 00:19:45.770
It's our responsibility at Vercel and as developers to upgrade the web for
292
00:19:45.870 --> 00:19:48.810
agents, which is a really exciting, I think, once-in-a-lifetime opportunity.
293
00:19:49.830 --> 00:19:52.650
In some ways it's an upgrade. It's going to get better for agents.
294
00:19:53.110 --> 00:19:56.530
But even technically sometimes it can be a "downgrade." It's just less HTML,
295
00:19:56.590 --> 00:19:59.230
more markdown. And as you said, it's the early innings. So,
296
00:19:59.810 --> 00:20:03.910
we are figuring it out as much as you all are and sharing it in the open,
297
00:20:04.370 --> 00:20:05.390
and learning as we go, as well.
298
00:20:06.630 --> 00:20:09.550
When did you make that shift that the agent was the customer?
299
00:20:12.050 --> 00:20:15.890
As a founder, I spend a lot of time talking to customers as much as I can.
300
00:20:16.290 --> 00:20:18.270
Sometimes on X, DMs, all these things.
301
00:20:18.690 --> 00:20:23.150
I started noticing there's the data, of course-nowadays,
302
00:20:23.580 --> 00:20:28.090
70% of page views on Vercel's developer systems and documentation
303
00:20:28.530 --> 00:20:30.880
are coming from agents, which is nuts, 70%.
304
00:20:32.050 --> 00:20:34.290
And 90% of that is served as markdown.
305
00:20:35.150 --> 00:20:38.490
So obviously the data is telling us all the time. At one point,
306
00:20:38.550 --> 00:20:43.190
I remember I tweeted/Xeeted that "10%"-this is a while
307
00:20:43.210 --> 00:20:44.770
ago-like, "Oh my God, I woke up one day,
308
00:20:44.810 --> 00:20:48.610
and 10% of my sign-ups were coming from ChatGPT." This is before agents.
309
00:20:49.430 --> 00:20:53.110
And those sign-ups, the data told us were more intentful.
310
00:20:54.510 --> 00:20:56.190
They tended to convert better,
311
00:20:56.710 --> 00:20:59.350
meaning that when you've been cooking with an assistant,
312
00:20:59.850 --> 00:21:03.030
the assistant understands your goals better.
313
00:21:03.430 --> 00:21:04.470
And so the data was telling us that,
314
00:21:04.550 --> 00:21:09.130
but the anecdata was people were coming to Vercel that were not
315
00:21:09.730 --> 00:21:14.290
as professional with software engineering. They were not as experienced.
316
00:21:14.910 --> 00:21:18.190
And they would speak to me in terms of what the agent told them.
317
00:21:18.870 --> 00:21:20.970
So they were carrying the message of the agent.
318
00:21:21.290 --> 00:21:22.910
I would find myself in conversation saying,
319
00:21:22.970 --> 00:21:25.030
"Can you just share the transcript?" It's almost like,
320
00:21:25.350 --> 00:21:26.510
"Can I speak with your agent?".
321
00:21:26.590 --> 00:21:27.110
Absolutely.
322
00:21:27.110 --> 00:21:28.130
It's actually built in the software.
323
00:21:28.810 --> 00:21:31.330
And so I started noticing that the persona was changing, evolving,
324
00:21:31.390 --> 00:21:34.150
and true to our mission,
325
00:21:34.610 --> 00:21:38.410
more people are being able to build. And that honestly got me really excited.
326
00:21:38.670 --> 00:21:39.890
Yeah, that makes so much sense.
327
00:21:39.950 --> 00:21:41.730
And I think so many people are experiencing something similar.
328
00:21:42.210 --> 00:21:42.710
Yeah.
329
00:21:42.710 --> 00:21:44.530
Okay. So we talked a little bit about pricing,
330
00:21:44.630 --> 00:21:47.450
and I know that's top of mind for everybody. How do you think... I mean,
331
00:21:47.510 --> 00:21:49.570
you guys have been at the forefront of usage-based pricing-.
332
00:21:50.150 --> 00:21:50.250
Yeah.
333
00:21:50.250 --> 00:21:52.670
And with agents, you really have to think about pricing differently.
334
00:21:53.010 --> 00:21:54.770
So what can you share with the audience about that?
335
00:21:54.930 --> 00:21:57.730
Yeah. Pricing is the journey of every entrepreneur, I guess.
336
00:21:58.050 --> 00:21:58.883
Such a journey.
337
00:21:59.330 --> 00:22:02.230
One of the things that's always been tough about pricing is that you get,
338
00:22:02.230 --> 00:22:04.790
with this platforms, you get so many different kinds of customers.
339
00:22:05.330 --> 00:22:08.030
We have the customers that come in and they're building their first prototype or
340
00:22:08.050 --> 00:22:09.330
the inkling of their first idea.
341
00:22:09.830 --> 00:22:13.510
And then we have the largest enterprises in the world that have mastered and
342
00:22:13.610 --> 00:22:15.630
studied the blade of cloud consumption.
343
00:22:16.010 --> 00:22:18.790
So they're used to buying stuff on the go.
344
00:22:19.110 --> 00:22:21.510
They're used to forecasting their cloud spend.
345
00:22:21.570 --> 00:22:25.030
They're used to buying through marketplaces of hyperscalers.
346
00:22:25.370 --> 00:22:29.390
And so we're always trying to find, is there a one size fits all?
347
00:22:29.870 --> 00:22:31.910
And to your point, I don't think there is. I think-.
348
00:22:32.550 --> 00:22:33.650
I wish. That'd be so much easier.
349
00:22:33.730 --> 00:22:36.110
I really wish it was just like, "Hey, one plan, fixed costs.
350
00:22:36.170 --> 00:22:39.650
Everybody go nuts." In reality, it's really nuanced.
351
00:22:40.110 --> 00:22:43.330
And so what we've tried to do over the last few years is,
352
00:22:44.050 --> 00:22:47.570
can we make full consumption models more digestible,
353
00:22:48.210 --> 00:22:49.510
more understandable?
354
00:22:50.050 --> 00:22:53.890
I think there is a huge edge if you're building a business to really embracing
355
00:22:53.950 --> 00:22:57.870
the versatility of consumption because I love the slide where you were showing
356
00:22:58.230 --> 00:23:02.890
there is the mythical 20-agent-in-parallel engineer.
357
00:23:03.350 --> 00:23:04.590
Is that not true? I feel like that's what you do.
358
00:23:04.760 --> 00:23:05.770
I have a few, but-.
359
00:23:05.770 --> 00:23:06.430
I feel like that's what you do.
360
00:23:06.510 --> 00:23:08.330
It's true that you have the power users, right?
361
00:23:08.750 --> 00:23:12.790
And then you have the P75 that is using AI more occasionally.
362
00:23:13.610 --> 00:23:16.190
And I think over indexing on one or the other is really a mistake.
363
00:23:16.670 --> 00:23:19.550
And so what we found is, for example, real-time usage visibility.
364
00:23:20.430 --> 00:23:21.450
This is before agents.
365
00:23:21.550 --> 00:23:26.170
We realized if we want to make consumption scale for small businesses,
366
00:23:26.630 --> 00:23:27.930
we can't just yeet, "Hey,
367
00:23:28.010 --> 00:23:32.270
here's a CSV of everything that's happened on the Vercel platform for the last
368
00:23:32.630 --> 00:23:33.270
three months.
369
00:23:33.270 --> 00:23:37.970
You figure it out." So much of the user experience became the
370
00:23:38.050 --> 00:23:41.510
experience of understanding your costs, monitoring them,
371
00:23:42.170 --> 00:23:47.070
soft caps and hard caps in real time, beautiful graphs, forecasts,
372
00:23:47.410 --> 00:23:48.490
all of these things. By the way,
373
00:23:48.730 --> 00:23:53.090
I'm glad Stripe exists because it's so much work, and, ultimately,
374
00:23:53.530 --> 00:23:54.750
it's undifferentiated work.
375
00:23:54.830 --> 00:23:58.570
I think all of us-I think I speak for most of us in this room-we want to build
376
00:23:58.610 --> 00:24:01.610
great products, we want to build great agents, we want to ship.
377
00:24:02.590 --> 00:24:06.210
But I'll tell you that, all of that work, for us, it really paid off.
378
00:24:06.870 --> 00:24:08.990
We went from customers that would get frustrated, like: "Okay,
379
00:24:09.270 --> 00:24:13.710
how can I understand this consumption thing?" To it really becomes a power and
380
00:24:13.730 --> 00:24:16.590
it enables basically unlimited growth for their business.
381
00:24:17.390 --> 00:24:18.390
That makes a lot of sense.
382
00:24:18.450 --> 00:24:21.490
And I think the real-time visibility is something we talk a lot about is how do
383
00:24:21.530 --> 00:24:23.830
we create that for people because I think that makes a big difference.
384
00:24:24.550 --> 00:24:24.550
Yeah.
385
00:24:24.550 --> 00:24:28.610
Okay. Switching gears: global. You guys have such a global customer base.
386
00:24:28.810 --> 00:24:32.370
Tell us more. Was that intentional? Has that just happened?
387
00:24:32.670 --> 00:24:34.770
Yeah. The first tagline for the company was,
388
00:24:34.950 --> 00:24:38.750
"Real-time global deployments." I ditched "real-time" because no one really
389
00:24:38.770 --> 00:24:40.890
understood what it means for our deployment to be "real-time",
390
00:24:40.950 --> 00:24:44.150
but the idea was you're thinking in real-time,
391
00:24:44.510 --> 00:24:46.190
the software is getting deployed in real-time.
392
00:24:46.850 --> 00:24:49.270
And global was very important to me-I'm from Argentina originally,
393
00:24:49.810 --> 00:24:53.170
and I would travel from the US back to Argentina, back and forth a lot.
394
00:24:53.830 --> 00:24:58.150
The experience of the internet is not even across the world. I always said,
395
00:24:58.150 --> 00:25:00.970
"If I were to live in one place, it would be US-East-1."
396
00:25:02.470 --> 00:25:04.510
Because when you go to Virginia, you go to the East Coast,
397
00:25:04.710 --> 00:25:08.090
they have a better internet there, especially if you sweat the milliseconds.
398
00:25:08.710 --> 00:25:11.330
So much of the internet lives in one region of the world,
399
00:25:11.710 --> 00:25:15.690
and I wanted to decentralize that. So Vercel today operates in 20 regions.
400
00:25:16.350 --> 00:25:19.190
By the way, for the nerdier people in the room,
401
00:25:19.310 --> 00:25:24.010
it's still hard to pull data gravity and compute gravity out of US-East-1,
402
00:25:24.330 --> 00:25:27.610
but I think we've made tremendous progress.
We've made tremendous progress with
403
00:25:27.730 --> 00:25:28.530
static,
404
00:25:28.530 --> 00:25:33.250
dynamic hybrid deployments where we make some of the pages be more automatically
405
00:25:33.310 --> 00:25:36.730
cached near the visitor. We made CDN a default,
406
00:25:37.450 --> 00:25:40.010
and now there's a huge opportunity to do the same for tokens.
407
00:25:40.790 --> 00:25:42.050
So Vercel has this product called the "AI Gateway."
408
00:25:43.590 --> 00:25:47.090
You can think of it as one API key, access to every model.
409
00:25:47.530 --> 00:25:51.850
And one of our aspirations there is to decentralize token access.
410
00:25:52.510 --> 00:25:54.170
We have huge customers in Brazil,
411
00:25:55.210 --> 00:25:59.390
huge customers in Europe that are serving their customers by getting their
412
00:25:59.450 --> 00:26:03.010
tokens from Portland and Virginia, which is pretty nuts.
413
00:26:03.350 --> 00:26:04.890
And the reason for that is that, today,
414
00:26:05.390 --> 00:26:07.010
we're like in the early days of the internet.
415
00:26:07.630 --> 00:26:11.970
A lot of the premium compute that's running the best models on the planet,
416
00:26:12.410 --> 00:26:14.370
we're basically out of capacity all the time.
417
00:26:14.470 --> 00:26:18.950
We're dealing with quotas and getting more capacity from the AI labs.
418
00:26:19.310 --> 00:26:20.770
But I expect that to change over time.
419
00:26:21.410 --> 00:26:24.090
Much like compute today is all over the planet,
420
00:26:24.590 --> 00:26:28.710
one of our goals is to create this token delivery network and make token access
421
00:26:29.230 --> 00:26:31.510
as close as possible to the end visitor.
422
00:26:31.570 --> 00:26:35.670
But the point remains that by designing to be global first,
423
00:26:36.410 --> 00:26:39.570
you basically create this incredible upside for your business.
424
00:26:39.710 --> 00:26:43.250
And I think using infrastructure like Vercel and Stripe makes that a default.
425
00:26:43.310 --> 00:26:46.410
Well, you guys are doing a great job with it. Okay.
426
00:26:46.710 --> 00:26:48.290
Let's try to hit on all the trends we talked about.
427
00:26:48.570 --> 00:26:50.050
Let's talk about all the go-to-market motions.
428
00:26:50.210 --> 00:26:53.270
So you guys started very product-led as well,
429
00:26:53.430 --> 00:26:56.490
but you now have a wildly impressive enterprise business too.
430
00:26:57.390 --> 00:26:58.390
What's that journey been like?
431
00:26:59.290 --> 00:27:02.650
Well, as I mentioned, it involved changing everything about the company,
432
00:27:02.710 --> 00:27:05.790
like adapting our pricing, adapting our product, our marketing,
433
00:27:05.850 --> 00:27:07.210
investing a lot more in security.
434
00:27:07.990 --> 00:27:11.090
But one of the things that we're talking about a lot these days is agent-led
435
00:27:11.110 --> 00:27:14.970
growth. It's not fully PLG, it's not SLG, it's agent-led.
436
00:27:15.450 --> 00:27:19.830
So I have incredible stories now of enterprises where there is a
437
00:27:20.610 --> 00:27:25.490
move by the CTO to start and sort of rethink the company from
438
00:27:25.530 --> 00:27:26.363
its bones.
439
00:27:26.850 --> 00:27:30.770
They create new initiatives to rethink existing codebases with AI.
440
00:27:31.370 --> 00:27:33.730
You might have heard rumors of Meta saying, "Look,
441
00:27:34.210 --> 00:27:37.530
our existing codebase is great, but agents are not so
442
00:27:39.110 --> 00:27:43.050
apt at using our old styles of writing code." So we're seeing this motion within
443
00:27:43.110 --> 00:27:47.930
enterprises where there's a reinvigoration and rebirth of a lot of
444
00:27:48.030 --> 00:27:52.210
infrastructure thanks to AI. A lot of SaaS companies are thinking, "Okay,
445
00:27:52.290 --> 00:27:56.030
what's the future of our business? Is it just seat-based?
446
00:27:56.450 --> 00:28:01.070
Is it doubling down on our UIs?" You might have heard of Salesforce
447
00:28:01.090 --> 00:28:04.110
announcing-this is kind of an important moment, I think,
448
00:28:04.290 --> 00:28:08.430
in the history of our industry. Benioff announcing the shipping of a CLI.
449
00:28:09.210 --> 00:28:11.040
You didn't have that in your bingo card for 2026.
450
00:28:11.790 --> 00:28:12.310
Definitely not.
451
00:28:12.310 --> 00:28:16.270
What we're seeing is this enterprise reinvention more towards primitives and
452
00:28:16.390 --> 00:28:20.750
building blocks, where enterprises are exposing themselves through APIs,
453
00:28:21.090 --> 00:28:23.430
SDKs, CLIs, MCPs.
454
00:28:23.650 --> 00:28:28.310
And what we're seeing is there is tremendous appetite for AI to help accelerate
455
00:28:28.350 --> 00:28:31.870
this. So products like v0, which you have a bunch of data,
456
00:28:32.050 --> 00:28:35.830
a lot of enterprise data, but now instead of procuring new software,
457
00:28:35.890 --> 00:28:39.990
you can just generate it. The go-to-market motion is accelerating.
458
00:28:40.070 --> 00:28:43.350
I think a lot of the C-suites that are making these decisions for the future of
459
00:28:43.410 --> 00:28:47.310
companies with AI, they are using the AI products themselves.
460
00:28:47.830 --> 00:28:49.990
I have champions within enterprises that tell me, "Well,
461
00:28:50.770 --> 00:28:52.570
I was using Claude Code over the holiday break.
462
00:28:53.250 --> 00:28:57.190
I now want to bring the power of Claude plus Vercel to my enterprise." And so I
463
00:28:57.230 --> 00:29:01.950
think this world is at continuing to converge so much so that I think if your
464
00:29:02.050 --> 00:29:02.350
idea was,
465
00:29:02.350 --> 00:29:07.010
"I'm going to have a super differentiated feature set only for enterprises," and
466
00:29:07.030 --> 00:29:10.730
you have to contact sales to even just imagine what the product is going to be
467
00:29:10.810 --> 00:29:11.590
like,
468
00:29:11.590 --> 00:29:14.830
I think you're going to have a hard time because that agent wants to try out
469
00:29:14.850 --> 00:29:16.370
that feature right away.
470
00:29:16.690 --> 00:29:20.250
Totally. Okay. Last rapid fire question.
471
00:29:20.890 --> 00:29:23.250
We've gone through a bunch of trends. We shared some checklists.
472
00:29:23.670 --> 00:29:26.650
What's one thing on your checklist that you think about all the time?
473
00:29:28.250 --> 00:29:31.930
I'm continuing to sort of obsess about the experience of the agent.
474
00:29:31.990 --> 00:29:35.550
Just this morning, I saw my agent getting stuck on a problem,
475
00:29:36.590 --> 00:29:39.370
and I really think about the cultural change that it's going to take.
476
00:29:40.830 --> 00:29:45.690
All of us in this room, we were born into an internet that didn't have AI,
477
00:29:46.110 --> 00:29:47.070
didn't have agents.
478
00:29:47.870 --> 00:29:51.810
And so I think a lot of what's going to make us successful in the future is this
479
00:29:52.030 --> 00:29:56.150
reinvention of every part of the stack, reinvention of the internet,
480
00:29:56.210 --> 00:29:57.043
reinvention of the web.
481
00:29:57.310 --> 00:29:59.870
My biggest piece of advice that I give to people is that like:
482
00:30:00.190 --> 00:30:05.150
"The faster you can get rid of your preconceptions-in some ways,
483
00:30:05.210 --> 00:30:08.730
a little bit of the ego of the things that the tools that you knew or the things
484
00:30:08.750 --> 00:30:12.110
that you thought you were good at-I think the faster we can actualize ourselves,
485
00:30:12.850 --> 00:30:15.890
continuing to ship, encourage people to build in public.
486
00:30:15.990 --> 00:30:19.190
I think a lot of the great things that have made companies like Stripe
487
00:30:19.550 --> 00:30:20.710
successful come from that,
488
00:30:22.110 --> 00:30:26.430
betting on the builders and the indie makers and building in public." So
489
00:30:27.710 --> 00:30:30.750
yeah, it was great to be here and thanks for this for hosting me.
490
00:30:30.910 --> 00:30:32.890
Yeah. Obsess over the agent experience.
491
00:30:32.950 --> 00:30:36.050
I think that is a very clear takeaway and one for all of us.
Guillermo,
492
00:30:36.190 --> 00:30:37.710
thank you. That was awesome.
493
00:30:41.550 --> 00:30:43.590
Before we wrap, I want to leave you all with this.
494
00:30:44.910 --> 00:30:48.930
I remember sitting around a campfire and a friend telling me to pursue Urban
495
00:30:48.970 --> 00:30:53.150
Escapes and that it was okay to quit my job to do that. I did.
496
00:30:53.830 --> 00:30:56.090
And it took months to build a website,
497
00:30:56.790 --> 00:31:00.410
two and a half years to get to four cities, and a lot of sweat along the way.
498
00:31:01.510 --> 00:31:03.960
And if I were that same person today-same campfire,
499
00:31:04.440 --> 00:31:09.340
same idea-I could have had my product live using Stripe and Vercel by the
500
00:31:09.400 --> 00:31:13.660
time that fire burned out. That's what's changed. It's not just the tools.
501
00:31:14.120 --> 00:31:17.760
The entire starting line has moved. You can build in hours.
502
00:31:18.280 --> 00:31:20.620
You can sell in 42 countries on day one,
503
00:31:21.320 --> 00:31:25.780
and you can go from PLG to enterprise to fully agent-led before
504
00:31:26.280 --> 00:31:27.113
your Series A.
505
00:31:28.180 --> 00:31:31.190
So the question isn't whether the opportunity is real.
You've seen the data,
506
00:31:31.510 --> 00:31:32.350
you just heard from G.
507
00:31:33.150 --> 00:31:35.670
The question is whether you're moving fast enough to capture it.
508
00:31:36.830 --> 00:31:40.670
This is the most fun time to start a company,
509
00:31:41.150 --> 00:31:45.070
whether you're a startup or one of the fastest-growing AI companies like Vercel
510
00:31:45.630 --> 00:31:47.950
or an enterprise trying to navigate all of it.
511
00:31:48.750 --> 00:31:52.350
And our job at Stripe is to be here with you every step of the way.
512
00:31:53.070 --> 00:31:56.390
So go build and know that Stripe will be here to help you as you scale.
513
00:31:56.950 --> 00:31:57.270
Thank you.