You.com raises $100M at $1.5B valuation: Richard Socher on pivoting from consumer search to enterprise AI infrastructure
Sep 3, 2025 · Full transcript · This transcript is auto-generated and may contain errors.
Featuring Richard Socher
guest, Ilia or Richard from you. com. Oh, he did. No. Oh, I thought that was news. Anyway, we got our next guest in the reream waiting room coming into the TVPN Ultradome. Welcome. We got Richard from you. com. How you doing? Very well. How are you guys doing? We're doing fantastic. It's a wonderful day.
We're having a great time. Uh, what's new in your world? Give us the news. Any big numbers? Yeah, we just raised $100 million at a$1. 5 billion. Boom. Uh, when did you start the company again? uh in 2020 overnight success. Nothing like a 5year overnight success. Love it. Um give us give us the update.
You were on the show was it uh I'm forgetting was it maybe three months ago at this point? Mhm. Yeah. Yeah. Main main thing since then is we kept growing more and more aggressively. Uh like we have over a billion API calls a month now. There are not many other startups.
In fact, I can't think of any other startup at that scale now. And these API calls give search results for both people directly, but also for LMS like the way we search is going to change, right? So, you have this whole Google thing that's unfolding now.
But that's kind of the the last uh you know, sort of past um of of searching. Now, people are going to search more and more through the LMS and LMS can search hundreds of times and then summarize all of that content for you, right?
So it's a whole new world and we're building composable infrastructure layers for companies and startups massive consumer companies like Dr. Go uh and also enterprise companies like Harvey and Windsurf uh the NIH is a customer view.
com telegraph like lots of different organizations uh both for the APIs but then also we built endto-end solutions cuz not every company is ready to just like take in an API and may have an epic AI transformation. So, we're doing both of those together. Yeah.
When when did you realize that that it made sense to combine those efforts?
I'm I'm sure early on companies like may maybe that maybe that you know actually building out these uh that this infrastructure for companies was not something that you were selling and then you realized like hey we in in order to actually unlock the value of our product we need to we need to kind of like handd deliver some of these features.
That's right. Yeah. For the last couple of years, we were the most accurate answer engine uh in the world. But normal consumers, they care more about marketing and and fun ads and stuff like that and designs, but companies kept coming to us and be like, "We did a benchmark. You are the most accurate. We need that.
" You know, there are hedge funds, the NIH, like journalists like that really care about accuracy. And so they were like, "We want the technology, but we don't necessarily want it on. com. We wanted inside our own products on our own websites in our own apps and so on.
And so at some point there's so much market pull there for the technology that we said let's build out a proper sales team. Let's really go to market with those APIs in an enterprise context. Give us your reaction to the Google news today. They beat the charges almost.
I mean they were found guilty but then the remediation battles. They won the war. Uh you're obviously an expert in this category. Uh walk us through how you process the news.
Yeah, I mean I think overall it goes in the right direction um in terms of giving uh some more access to certain companies um for their index and some of their data. Uh at the same time, you know, we're a startup still and by the time this gets really implemented, you know, they're probably going to try to appeal it.
Really, we were really struggling to understand what giving the search index one time and then on ongoing basis means. It's only like some percentage of the quer data based they're giving you data based on some percentage of your query. Yeah. Can you explain like what exactly is available? Can I just go get this data?
Can you. com go get this data? Can anyone get this data? Like what does it actually take? Is this just like Microsoft Bing and the other like duck. go like the big competitors? Are they named? Like how does this actually work? Yeah. So no one fully knows yet, right?
there's like recommendations from judge and then like so there's like a ton of meta within eh not company meta um but insider dealing um but like my hunch is uh the one the most extreme scenario here would be like they have a full-on API the way you. com right now sells it, right? Which could be good uh for some people.
My hunch is they're going to try to drag their feet as much as possible on that for as long as possible because that is very much core IP for them. Um, and then they're also going to realize that the index, the way it used to be built was for people to decide which of 10 blue links to click on.
Our index now is looking very different. Uh, it can go deeper. Uh, it can crawl more. It has these decomposible components and it actually gives you much more content per link back rather than just a short snippet. the way like the Google search results would.
And so by the time this all comes out and is might become available maybe as an API, um it'll probably be somewhat outdated for the use cases LM need. Do we need an API here? I feel like if I have an AI agent that has a browser, the browser can go to Google and search for me and then scrape the results.
Like is not the the front end the API?
Like we does Google have a like a blocker on this like can openai not just Google things like they certain their employees can Google things I imagine um but can they not like resell that to me in their chat app like what's the dynamic there yeah it's a really interesting world actually you know there are not many uh real web indices out there and you have a lot of startups including some of our competitors who essentially just do what you described uh through so-called SER APIs and proxy networks where you have your phone, you have your smart TV, uh, and you install some app and that app then has some SDK installed somewhere where if you're not using your phone or your TV, it becomes an IP address.
Interesting. And then pretends to be a user, goes to Google, scrapes the content from a search query in that, sends it back to centralized IP, and then folks sell search results, but they're really just Google results.
Like literally biggest competitors in the consumer space are just really fast at scraping Google results and then connecting it to an LLM and doing all of that. It's it's hilarious.
Now the problem with that is that a if you're a bigger company like OpenAI or or Facebook Meta, like they wouldn't want to send all of their data and all their queries to Google, right? They're literally trying to compete with them. So that's not an ideal setting. Um two, it doesn't really scale.
um like cuz Google can and does shut down these proxy networks all the time and they try to block it and have captures and all of that and then three as it you know as you go through someone's random phone and maybe they did click on something now it's really slow and there slow response times and so where we actually have benchmarks that we can you know back up we can go on.
com and find them like we are super fast and we're more accurate than Google serve APIs and that is really unique in the market. What about this news? It just broke that um Apple will be potentially partnering with Google to have uh Siri powered by some sort of version of Gemini. Was that on your bingo card?
Was that expected? Does that seem like a reasonable uh kind of end state of the market uh dynamic? I was I was talking about this earlier. Feels very Yeah. something that feels like it'd be on your on like your forecast given that you guys are helping a lot of companies build these sort of like knowledge. Yeah.
Like if you're if you're in like business 101 going to Silicon Valley, you would say that like the ability to search and get great results is valuable. Therefore, Apple would be paying Google for search results.
But in fact, the money flows the other direction because the search because Google's making $60 billion a year or whatever off of uh off of Safari search results. And so how do you see the relationship and the and the economic flow d uh changing there? I think the news report was that Anthropic wanted 1.
5 billion and Apple said that's too much. We're going with Google and Google will be getting less than 1. 5 billion maybe from Apple for the rights to use the LLM. But what yeah what's your take on all that? Yeah. So a couple different angles here. So one, it's surprising that Apple cannot do it themselves still.
Like it's kind of crazy. uh and surprising they're like you know trillions of dollars worth and and somehow are not able to execute on this themselves. Um the second thing and you know they have great privacy at Apple but then you know they partner with companies uh that have a different stance on privacy.
So there's like interesting complexities there. The second thing is that uh indeed you know one of the biggest changes also uh of the um uh of the ruling today was that Apple is will continue to be allowed to take the 20 billion to not build their own search index. Right.
Um so that's that's a really interesting part of the ruling too. Google is allowed to still pay Apple $20 billion a year to stay the default. Um, so that is is a really important piece of this whole situation because defaults are very strong.
I don't know where I read this, but like 80% of all iPhone users never change a single setting. Maybe not your viewers, but many normal iPhone users, they don't change any setting ever. So, if you're the default, you'll be on there forever.
I wonder how I I mean, I imagine for me don't need I don't need to adjust the volume. I'll never adjust the brightness of my screen. Not a single setting. No, I imagine that it's like deeper in the settings, of course. That is hilarious. Uh anyway, do you have anything else?
We're We have a couple other people in the re very happy to to hear about your your progress. Yeah. Congratulations. This is awesome. I I It's my favorite new pun. Unicorn. Yeah. There we go. There we go. Unicorn. Unicorn. Fantastic. Well, yeah. Tremendous. Thanks for joining. We'll talk to you soon.
Have a great rest of your day. Our next guest is already in the reream