Keith Rabois on why most defense startups will fail, Ramp's pitch to Congress, and OpenAI saving San Francisco

Okay. 10 minutes. Yeah. Let's do 10 with Keith and then 10 with 10 with Keith and then we'll have you on. Is that great? Yeah. Keith, come on. Sweet. Welcome. Hey, how are you? What's going on? I love this live studio. Yes. Yes. Yes. Uh we are live.

Um but uh we're very fortunate to be graced by someone who's not going for confirmation. Maybe can give us a little more hot takes, you know. Oh yeah, I can see. First question, uh do we need a department of government efficiency for confirmation hearings? They seem like they're going slow. Absolutely.

Um you know, I think this should be simplified at some point, which is basically the president should nominate you for the constitution and and the senate should provide advice and consent and we don't need too many other steps in the process.

But that said, um you know there's lots of room improvements in the government. First saving money is probably more important. Yeah. Making it more attractive for talented people as we talked about um to serve and do setting a good example for this is being a magnet for talent.

So the best people in America want to improve society through the government would be great. Those are much more fundamental. So there's details that can be improved at the margin, but I wouldn't confuse the details with the strategy. Yeah. Are you doing more uh investing in companies that touch the government now?

Because the this feels like Hillen Valley is like the floodgates are open. Uh DC is kind of open for business in some ways willing to do business with startups.

We've seen success with the big defense tech companies, but now we're seeing energy and agriculture and a lot of uh kind of like hard tech companies or hard tech adjacent just Yeah.

I mean the big issue is it feels you know being here nothing about investing in these categories feels contrarian anymore it was for a long time well if you look at like the hot sectors AI still is the most you know there's some founders here I would say and defense tech is the second and so I think that that is a structural problem I don't think I agree more with Delian um's comments about there's not going to be many successful defense companies I think the one thing that is a positive change and is structural and you know uh can improve the odds of success for many founders and entrepreneurs is moving into federal revenue earlier than usual.

So many companies have historically had a federal revenue arm. They may not talk about as much publicly for a variety of reasons but usually that comes after significant commercial success. You might see the sequencing change. I mean the government runs on Microsoft scale.

ai had federal revenue has federal revenue that's meaning VA significant federal revenue early in the trajectory so that may be more normal and more common but I don't think government only revenue is going to lead to a lot of successful companies except for very very small sliver so we as a fund are not really investing very actively in government oriented solutions we do have I just did make a defenseoriented technology investment but that's like an outlier but we have lots of lots of companies that have meaningful federal revenue and continue to scale.

And this feels like a category. Last time we had you on the show or maybe it was the first time you talked about how most like a lot of people just shouldn't be like, "Oh, I want to be a founder. I'm going to start a company. " They should go work for a great company.

And if you're trying to uh if you're excited about spaces that are going to require federal funding early, you probably shouldn't start that right, you know, when you're 20 years old maybe, unless you're, you know, truly well, you need to understand the sales process in the federal government, which is different than a sales process in a commercial environment.

And there's probably a limited scare there's probably more scarcity on quality salespeople that can sell successfully to the federal government. Yeah. And maybe there's 40 of them. And so unless you have an angle to be able to attract one of those 40.

So do you have a better product that the sales people think, "Oh my god, I can sell this. " Like the best way to attract sales people is make their life easy. I've got something cool differentiated and I know the buyer. Second, you have a missionary zeal that's off the charts. Yeah.

um without without the ability to attract one of the top 40 people who can sell into let's say defenseoriented stuff threeletter agency stuff I think it's really dangerous um so I think that's more critical sometimes than the technical breakthrough um you know the government is decades behind as talking about like you know Oracle started serving CIA and NSA there are examples like that but there's very few and so you know one every decade so that seems to me to be a bad strategy Y unless you have a reason to believe you're the once every decade, you know, company and that should be part of your initial seed deck.

Yeah. Uh talk about what you're talking about today. You're you're doing a talk with Eric Lyman, correct? Yeah, we just finished uh Eric and I were on stage with Senator Erns from uh Iowa.

um talking about how we can make the federal government's expenses, you know, more streamlined using programmatic expense reporting like like that RAM offers commercial sector.

Commercial products, commercial customers of ramp save between 5 and 15% on expenses and the federal government probably would be able to save more. So there's four 4. 5 million card holders in the federal government. There's only 2 something million employees. So there's clearly a lot of waste, fraud, and abuse.

Um the senator cited a bunch of examples of 80,000 charges on federal credit cards at casinos. Clearly most of those 80,000 charges are not doing official business on behalf of the federal government. So hopefully you know that's an easy savings. It doesn't require cutting programs that people care about.

Doesn't require this big controversial debate. But every dollar we save goes back to the American family and they can spend it on raising a family successfully.

Secondly, if we can cut 1% from the federal good budget, just 1%, the Federal Reserve can lower interest rates by 1% safely and allow people to afford mortgages, allow the economy to grow without triggering inflation. So, it's a very big deal to save 1 or 2%. I think a lot of people miss that. Yeah.

Do you have any good uh stories from uh visits to Washington during the PayPal days? Was it It was different time. It was different. I was actually just mentioning um I don't know if I'd been back in the capital before today since 2003. Wow. Um when I was lobbying, that means we're officially back. America is back.

Um but yeah, like we we spent a lot of time and energy.

I spent a lot of personal time and energy um uh using the federal government uh to help protect payball from a lot of people that didn't like us on Mastercard, eBay, and then later the Treasury Department post 911, the Treasury Department wanted to issue these rags that would have imposed significant burdens on consumers and small businesses.

And we were able uh using a lot of uh connections on the hill and messaging about not interfering with e-commerce uh to put put a lot of pressure back on the Treasury to to draft much more reasonable regulations.

They really wanted everybody to put in their social security number on every like every transaction which is you know crazy intense. Last question. Jensen was saying SF is back because of AI. What's your take? Well, he stole my line.

Yeah, I mentioned that I my my version of the line was slight slightly narrower but you obviously can expand it. Mine was open AI saved SF without open AI success SF would be dead and still dying.

Um you get another power law company and that drives a ton of it does and you know the people who are favorable to SF would say oh that well that happens every 10 or 20 years like the internet saved SF too.

You know, when I graduated Stanford before the internet, um SF was like a a 12th to 15th ranked city in the United States. Nobody there was no business going on. Oakland was actually a better commercial city at the time. The only people who stayed in SF to work at like Bane or BCG were like for lifestyle reasons.

It was a dead city. Then the internet revolution happened and although a lot of the best internet companies and best technology companies were in Palo Alto and Mountain View and Sunny there was enough spill over at SF to sort of save the city. It was just cheaper in some ways, right? Yep. No.

And well cheaper and you need large floor plan like floor plan off uh floor floorprint uh offices like uh and those those didn't exist in the city. The city's culture was never really never great for building companies.

Like when we considered moving the PayPal office to SF in 2001, Peter asked this question, Peter Tale asked this question of name a successful SF company and the largest SF company at the time was like CNC or something incredibly mediocre. So we didn't we stay we moved to Mountain View. Interesting. Wow.

Well, thanks so much for joining. This was fantastic day and we'll see you out. Enjoy the rest of the conference. Uh Ben, do we need to shift Thank you. Uh do we need to shift the microphone a little bit? Have Zach maybe sit at the end of the table? point towards us. Okay, we're gonna shoot around. Yeah.

Um, but come on in. Welcome. Uh, repeat guest. Yeah, we're