Loading summary
Transcript38 lines
- [00:00]
Panasonic Narrator
For over a century, Panasonic has harnessed the power of innovation to advance industry and improve lives. Today, we are continuing that legacy by investing in US Based manufacturing, reinforcing resilient and secure supply chains and creating high quality future ready jobs. Learn more about how Panasonic's efforts are directly supporting economic competitiveness and long term energy security.
- [00:26]
Aaron Levy
Foreign.
- [00:31]
Stephen Overlea
Hey, welcome back to Politico Tech. I'm your host Stephen Overle and on this show I break down tech politics and policy with the people shaping our digital future. Let's talk about the art of the tech deal. In recent weeks, the federal government has been buying AI technology from private companies like Anthropic and OpenAI at very cheap rates. I'm talking $1 per federal agency. It's a program called OneGov and the general Services Administration says the goal is to finally bring the government into modern times. And for companies, it's a chance to get the government hooked on their technology. One of the companies to cut a deal is Box, the seller of cloud computing services. Box CEO Aaron Levy has been on the show a few times before and I called him up again to discuss why these deals are happening now under President Donald Trump and what he thinks the AI powered government of the future might actually look like. Here's our conversation. Aaron, welcome back on Politico Tech.
- [01:50]
Aaron Levy
Thanks for having me.
- [01:51]
Stephen Overlea
Of course, you know, every tech company right now it seems is cutting a deal with the federal government. Box is among them. Why now? Like, what is this administration doing differently?
- [02:03]
Aaron Levy
Yeah, so there's, there's, I think at least two major dimensions where I think the federal government has been very proactive in thinking through how to work with the tech ecosystem in a strategic way. The first is really on the procurement front. The government obviously buys lots of technology and it works with lots of technology companies. But when they do that in kind of separate work streams, they're not often able to connect enough dots to get the kind of scale that could be advantageous economically if they were to kind of route those through, you know, one centralized buying power. So there's been a kind of a procurement efficiency, cost savings orientation of this administration that I think has been fairly effective. And really kind of tied with that is a modernization push which then gets you to the second point, which is really the power of AI and the government using more AI across really all of the various kind of government functions to just drive more efficiency, better productivity, get more output. And I think that has been a unique effort under this administration, partly just because the technology is now ready for that. So you probably could have done more of these procurement efforts previously. But the AI push is sort of uniquely timed for the moment when AI is actually ready to be used really in federal agencies for the first time.
- [03:31]
Stephen Overlea
Got it. You mentioned driving down prices and obviously, as you know, the headline on some of these deals has been OpenAI charging a dollar for chatgpt. Right. These nominal kind of fees. What are you charging?
- [03:45]
Aaron Levy
Yeah, so what we provided and kind of collaborated on with the GSA was to effectively provide a discounted rate for our two highest tiers of our enterprise plans, which so 75% and 65% respectively, discounts from our kind of normal typical list rate. And so that again, allows for an agency that might only have a few hundred people using the product or deploying the product to get the same volume pricing discounts that maybe an agency who had 10 or 20,000 users on the product would have gotten previously. So it's a way of really unifying the volume pricing structure for agencies of all sizes and for any kind of use case in the government.
- [04:34]
Stephen Overlea
Got it. So not exactly a dollar, but a discounted rate nevertheless.
- [04:39]
Aaron Levy
We're a little bit beyond, unfortunately, the dollar kind of program, but maybe 10 years ago we would have been able to pull that off.
- [04:45]
Stephen Overlea
What is the business case there? I mean, CEOs are not in the business of offering their products usually at a discounted rate. So what is the real business play.
- [04:55]
Aaron Levy
For us for ChatGPT? I'd say two very different situations.
- [04:58]
Stephen Overlea
Let me, let me hear your take on both.
- [05:00]
Aaron Levy
Yeah, well, for us it's, I think, actually just a logical way to support, let's say, smaller agencies where again, they get the benefit of the full buying power of the government, where previously that would have been kind of harder to pull off before this program. You know, overall at box, we have hundreds of deployments across the federal government. And so to us it's mission critical work that we want to support. And it's also very kind of well timed because, you know, again, there's a need to drive not just more efficiency, but better security of data, better protection of very important information. These are use cases that align to our platform, I think quite nicely. So that's sort of the benefit to us. But our federal effort and focus has been a core focus of the company for really kind of the past decade.
- [05:49]
Stephen Overlea
What do you think is the motivation for an OpenAI to offer their stuff for a dollar?
- [05:54]
Aaron Levy
Well, yeah, I think that the dollar thing is probably a little bit of a kind of seeding strategy to Ensure that more people to the technology and can start to use these tools and build more fluency with AI. Obviously ChatGPT would like that fluency to be within the OpenAI kind of universe and paradigm. I'm sure Anthropic and Claude would like that to be in their universe and so on. So I think the idea that our leading private sector AI companies are making their AI much more easily accessible by federal agencies, I think is fantastic. We'll have to obviously see how that pricing evolves over time. And again, that's all kind of just consistent with this idea of how do you start to modernize how, how the government works.
- [06:40]
Stephen Overlea
Well, that was sort of my read on this was part of the motivation for some of these companies breaking into the government was get in the door, get people using it and then, you know, you can charge more later once it's proven it's worth.
- [06:52]
Aaron Levy
Clearly there's like these, these people are business people at the end of the day and they have, they have like, you know, kind of creative growth strategies that, that are, that are likely t. And I think there's clearly a race toward who can drive the greatest adoption possible of any kind of various AI systems. And so I think you should assume this is happening in some form in all sectors. So this is not federal specific in any way. These are the strategies of these AI companies. And the federal vertical just offers one other unique approach where, you know, I think One Gov, the one Gov program is doing a great job at getting again, these vendors to lower their prices in a way that makes it easy to adopt this technology. And then of course, again, we'll see how the pricing evolves over time.
- [07:41]
Stephen Overlea
Now, you're not a CEO who shies away from politics. This. I know we've talked before, so let me just kind of ask you straight up.
- [07:48]
Aaron Levy
Although I am a CEO who probably regrets not shying away from politics. So that's a different category.
- [07:56]
Stephen Overlea
That is a different category. I mean, is part of the motivation here getting on Trump's good side?
- [08:04]
Aaron Levy
Not really. We don't have that much kind of.
- [08:08]
Stephen Overlea
Political.
- [08:10]
Aaron Levy
Kind of needs per se. Most of our policy issues are already well covered by the tech companies, much larger than ours. So we care about things like obviously the kind of core economic policies. We care about AI regulation being kind of done in a thoughtful, pragmatic way. I care about being able to make sure that we have an environment for great talent, whether that's homegrown talent that goes through our education systems or high skilled immigration programs. I don't really see it as any kind of political move. This is purely. We've worked with the federal government for well over a decade now and we want to continue to support its broad mission because there's obviously, you know, federal government is doing incredibly important mission critical work every day and we want to continue to help power that.
- [09:14]
Panasonic Narrator
For over a century, Panasonic has harnessed the power of innovation to advance industry and improve lives. Today we are continuing that legacy by investing in US based manufacturing, reinforcing resilient and secure supply chains and creating high quality future ready jobs. Learn more about how Panasonic's efforts are directly supporting economic competitiveness and long term energy security.
- [09:42]
Stephen Overlea
You mentioned AI technology now being in a place where it can actually be useful. The real question on my mind though is like how confident are you that the government is actually ready and in a position to use the technology? Because that's a different question.
- [09:58]
Aaron Levy
Yeah, yeah, yeah, totally. So technology readiness, obviously kind of one factor kind of workflow readiness, IT readiness, employee readiness, totally different kind of challenges there. The reality is there's only one way to kind of get that ball rolling and it's to start with access to the technology. There's only so much kind of theory you can have around how much will AI impact that workflow, how much will AI make that project more efficient? You ultimately just have to start using these tools. So in the private sector there's a concept of going AI first. We have an AI first push. Companies like Shopify, Duolingo, some of the bigger tech companies like Google and whatnot are using AI in kind of every part of its work and every, every element of how we operate. So you can obviously expect that the same thing should happen in the government. Imagine if you're a 21 year old that wants to go in and make an incredible impact on America and you join the government. And yet the tech stack looks entirely different from what you've been using for the past four years in college. And the tools aren't modern. You don't have access to AI. Think about how much energy and productivity we will have just squeezed out of that person that we're not able to then tap into because we don't have the right technology for them to use. So we need to make sure that our government is as fluent as possible in using these tools. And I think that again, one Gov is a fantastic catalyst for doing that. I think what David Sachs and the overall team out of the science and tech kind of policy dimension they've been pushing on incredible kind of AI I Think regulatory work. All of this is in, I think in service of how do you make sure that the US government is at the forefront of AI and do you.
- [11:51]
Stephen Overlea
Think they can get there? Because I mean you have like the workforce questions of the workforce being ready for the technology. You also have I think some security questions given the sensitivity of government data and you know, the reputation, the legacy of not necessarily being the most cutting edge enterprise, the federal government. Right. And I guess I just see a lot of hurdles that have to be overcome.
- [12:12]
Aaron Levy
I think everybody would agree with you. Those hurdles totally exist. And, and even if you could not move as fast as like the brand new Silicon Valley startup that just emerged, you know, tomorrow with an all new, you know, kind of a blank sheet of paper as their tech stack, the question is sort of what, what percentage delta would you want the government to have versus the private sector? Do you want it to be 1 to 1? Do you want it to be 0.9? Do you want it to be zero point five? Three? Do you want to be zero point one? That's a continuum. And I'd say actually there's many areas where the federal government is ahead of the private sector because of the vast data troves that they have access to. And again I kind of think about it as if you were, let's just take something that is sort of, we're unrelated to. So I can be even more, let's say unbiased. Would I as a taxpayer or as a citizen want an engineer going into the federal government and not having access to cursor or Claude code or cognition from Devin or repl it like what I want every, every line of code in the federal government has to be written by hand and you can only then move as quickly as you can type out the code on your own. Do I want you to do the slow path or do I want you to do the accelerated path? And I want you as a government employee to have access to the best technology possible to do the fast path. And again, the only way then you can get there is if you have a top down mandate that says we do want the government to be as AI first as possible.
- [13:41]
Stephen Overlea
Do you have a vision for 5 years out, 10 years out what an AI powered government looks like to you?
- [13:49]
Aaron Levy
It's different between state or local and federal level for the citizens daily experiences because the federal level is usually from a federal agency standpoint, one or two steps removed from your daily life. I think the impact is going to be more often one click indirect. But the way that it will show up in your life is imagine an FDA that can streamline again the drug review process. Imagine the small business loans that will get processed much faster. Imagine the disaster response where data gets shared much more effectively. Imagine the, you know, the breakthrough research in material science or basic research that can get done because of better collaboration between universities and the federal government. All of those things will be aided by AI. That's not going to change my daily life as I drive down the road or go to some of the ways that I experience kind of more local related government issues. But it will absolutely continue to push the economy and push the country forward in all of these indirect ways. And that's, I think, the power in the first place of the federal government. And, and now I think with the power of AI that just gets further accelerated.
- [15:09]
Stephen Overlea
You know, all of the embrace of AI and the governments looking to be more efficient using technology in many ways. A lot people see that as kind of Doge 2.0, like this next phase of reforming the government. And I think last time you were here, we talked about Doge, we talked about Elon Musk and his impact. I'm curious what you think now that obviously he's no longer involved directly with that effort. What you make of how that's gone and where this is all headed in terms of government efficiency and the streamlining we're talking about.
- [15:43]
Aaron Levy
Yeah, I mean, the thing that I've kind of been pretty focused on over again well over a decade and actually Obama had the first push on this idea of kind of a cloud first strategy. And through that I think there was clearly a push to how do we modernize technology, how do we get more leverage from our technologies? And that's the power of cloud in the first place. I think over probably a five to eight year period, for a variety of reasons, Covid other political drama and stuff, I think that kind of took a backseat. And then I think at the very start of the Trump administration, I'm sure credit is owed to many different people. But let's say Elon driving the Doge side that came into the forefront again, which said, okay, let's actually modernize the government. And so I think that became a component of Doge, which clearly is the thing that will probably transcend any particular format of Doge or the leadership of doge. It's bipartisan. Everybody can get around the idea that let's spend less in technology, let's get more efficient in our government technology operations, let's get more out of our data, let's drive more Automation. So actually we don't have to have people kind of spending all of their time manually reviewing every single process in the government, which is insane. We can use that time in much more effective, much more creative, much more productive ways. And so that's, I think, the part that will persist. And maybe DOGE was an extra kind of catalyst for that in this new era.
- [17:13]
Stephen Overlea
Last question for you. And that's just what do you think is next? I mean, what's the next step between where we are now and sort of this utopia Techtopia, AI driven federal government that is efficient and wonderful and everything it's not today.
- [17:31]
Aaron Levy
Well, maybe this is just because we've been doing this a while. I have a lot of gray hairs. I'm not, you know, I think the road to some Techtopia is like, you know, it is, it has many ups and downs. The journey will be very dynamic. I'm not pitching any kind of instant panacea kind of outcome that AI will deliver. What it delivers is when you can accelerate a workflow that you're doing over and over again that you can define with AI or an AI agent that can help make your process more productive. Review that contract and pull out the key details from that contract, help generate this communication that needs to go out to some ecosystem process data that we have to share with researchers in the private sector. Those are the kind of workflows that we should be bringing automation to. And that can make again, make our workflows just much more efficient, but also again, not impair the work that happens in the government relative to what we're going to see in the private sector. In the private sector, this is the direction that everything is going in. So let's make sure that our government is as aided as possible in using these tools to be as productive as possible. And again, in many areas actually probably innovate in ways that the private sector wouldn't have because there's use cases that exceed the mission criticality of the private sector. Clearly what we're going to see in security and defense will exceed what the private sector was previously doing. And the private sector will actually benefit from those advancements being able to use AI agents for either finding security vulnerabilities or protecting the government. I would put money on the government actually being in the best position to be on the advanced end of that kind of work or those kinds of technologies. And then we'll pull that into the private sector. So I think just in general, we always want the government to be modernizing and staying current with with any kind of important tech trend. It was important in the early PC era. It was important when we went online in the 90s. It was important as we went to the cloud and mobile in the mid 2000s. And now AI is just the next version of that.
- [19:44]
Stephen Overlea
Well, listen, Aaron, thanks for being back here on Politico Tech.
- [19:47]
Aaron Levy
Thanks. Thanks for having me. Appreciate it.
- [19:50]
Stephen Overlea
That's all for this week's Politico Tech. If you like Politico Tech, please subscribe and recommend the show to a friend or colleague. And for more tech news, subscribe to our newsletters, Digital Future Daily and Morning Tech. Our producer is Normal, Maliko Pram Bandy made our theme music. I'm Stephen Overlea. See you back here next week.