Palantir Technologies Inc. (PLTR) Earnings Call Transcript & Summary

Good afternoon. I'm Ana Saro from Palantir's finance team, and I'd like to welcome you to our first quarter 2023 earnings call. We'll be discussing the results announced in our press release issued after the market close and posted on our Investor Relations website. During the call, we will make statements regarding our business that may be considered forward-looking within applicable securities laws, including statements regarding our second quarter and fiscal 2023 results, management's expectations for our future financial and operational performance and other statements regarding our plans, prospects and expectations. These statements are not promises or guarantees and are subject to risks and uncertainties, which could cause them to differ materially from actual results. Information concerning those risks is available in our earnings press release distributed after the market closed today and in our SEC filings. We undertake no obligation to update forward-looking statements, except as required by law. Further, during the course of today's call, we will refer to certain adjusted financial measures. These non-GAAP financial measures should be considered in addition to, not as a substitute for or in isolation from, GAAP measures. Additional information about these non-GAAP measures, including reconciliation of non-GAAP to comparable GAAP measures, is included in our press release and investor presentation provided today. Our press release, investor presentation and SEC filings are available on our Investor Relations website at investors.palantir.com. Over the course of the call, we will refer to various growth rates when discussing our business. These rates reflect year-over-year comparisons, unless otherwise stated. Joining me on today's call are Alex Karp, Chief Executive Officer; Shyam Sankar, Chief Technology Officer; Dave Glazer, Chief Financial Officer; and Ryan Taylor, Chief Revenue Officer and Chief Legal Officer. I'll now turn it over to Alex for opening remarks.

On any normal earnings call, where we have just been profitable and where we're going to be profitable for the next couple of quarters, which opens up the possibility to be on the S&P, where the U.S. market grew by 28% and where we made $187 million in free cash flow, that would be the thing that was the most exciting. But in fact, what's most exciting about Palantir is we have our ability to launch products that are literally the only products on the market and that will, in fact, change your life and will determine who succeeds and who fails across enterprise, both government and commercial. The large language model revolution is one that will raise ships and sink ships. Of course, the profitability is important. Of course, the fact that we make free cash flow is important, but it's not nearly as exciting as our ability to invest our resources and our founder-led energy in using our things we have built to supply enterprises with, and otherwise not available, and in crazy high-demand product. This is just one of the most exciting times to be at Palantir because you have the resources of a large profitable enterprise company. So there is an issue with nonprofitable tech if you're powering the most important enterprises in the world. They want to know, can you provide the resources, the product we need tomorrow purely on a financial basis? But more importantly, to us and to them long term is, can you help us disrupt our adversaries? Can you just help us disrupt our competitors? And for me, personally, can you have a product that will help the West win, especially our government, but also our commercial clients, so that we, in fact, are stronger than our adversaries? And in the last 20 years, there's never been a development like this. You have a technology that will allow you to outproduce, change the margin of your company, understand your business, react on the battlefield quicker, predict things on the battlefield in a way, collapse your enterprise so that the top and the bottom actually work together, preempt attacks, create a software that is so obviously dominant that adversaries quiver and scurry away instead of attacking us or our allies. And there's 1 company in the world that is positioned. The issue of how do you have security, a data model or knowledge and wisdom that's proprietary, interact with an external large language model or with generative AI is not new to Palantir, and that's why we're able to launch our platform AIP so quickly, the demand for of which is nothing I've ever seen in 20 years of being involved in Palantir. And the reason the demand is high is people suspect that this will -- if you wield these technologies correctly, safely and securely, meaning extract the value in the context of your own enterprise, whether that's sensitive or nonsensitive or regulated or moderately regulated, you have a weapon that will allow you to win, that will scare your competitors and your adversaries. And we are in a unique position to supply that platform. And we have the resources, both because of our profitability, our $2.9 billion in the bank, our lack of debt and, quite frankly, our entrepreneurial founder-led spirit at this company. Welcome to our earnings day. Thank you.

As Alex highlighted, our company once again achieved GAAP profitability last quarter, including GAAP operating income for the first time. This marks another milestone in our company's sustained growth, ensuring that we will continue delivering results and impact for our partners for years to come while also investing deeply in the transformational AI opportunities before us. We have been taking steps across the company in recent months to refocus our efforts and optimize on the parts of our business that will drive even further growth alongside sustained profitability. In Q1 2023, we generated $525 million in revenue. Due to the seasonality of our business, Q1 tends to be our slowest quarter. But despite that headwind and the difficult macroeconomic environment that the technology industry continues to face, our commercial business generated $236 million of revenue last quarter and achieved $176 million in TCV, a 70% year-over-year TCV increase. These strong results were driven primarily by the reacceleration of our U.S. commercial business, which surpassed the $100 million revenue threshold for the first time with 26% year-over-year growth. We continue to see robust pilot starts and promising conversions, and we're also beginning to see the realization of our expansion strategy, meaning we're beginning to see meaningful growth and upsell opportunities with our newer customer base. Some notable examples include the expansion of our work with Hertz, who is using Foundry to more efficiently manage and operate its fleet of nearly 500,000 vehicles; and Jacobs Engineering, who is doubling down on our partnership to reduce costs and improve performance across plants. We also signed significant expansion agreements with the largest health system in the country for continued acceleration of our hospital operations efforts and with one of the world's largest paper and packaging companies. Our U.S. commercial customer base, which stood at 155 at the end of Q1 2023, a sevenfold increase in customer count over just 2 years, presents an immense opportunity for continued expansion. While aspects of our international commercial business have been challenging in today's climate, we continue to focus on delivering transformational results to long-term customers. For example, BP, our customer of over a decade, recently shared that Foundry helped them reduce production costs by approximately 60%, from $14 a barrel to less than $6 a barrel. We're also investing for growth in targeted industries and geographies, such as Korea, where I recently met with customers and saw our momentum there firsthand. This includes several strong pilot starts and the expansion of our work with Korea shipbuilding and offshore engineering focused on using Foundry to enhance safety and operational efficiency. Our government business generated $289 million in revenue, driven in large part by our U.S. government business, which grew 22% year-over-year. With strong conviction in our work on the ground and the critical missions we're delivering against, we continue to focus on building our U.S. government business for the long term while acknowledging timing uncertainty in the short term. We're investing in delivering digital deterrence and AI-driven efforts around the world from the Middle Eastern Pacific and European theaters. And with pressing global events, we continue to lean into our support for the U.S. and its allies across Eastern Europe. More broadly, as we look at opportunities across our company, we are strongly investing in AI efforts, with a focus on delivering the foundational systems and software architecture that will enable enterprises to leverage the power of the latest large language models and other machine learning technologies. We're already seeing unprecedented demand for AIP, and we are reorganizing our efforts aggressively to capitalize on the interest. We anticipate that these technologies will be transformational, both for ourselves and for our customers, and we are positioned to meet the moment. I'll now turn it over to Shyam.

Thanks, Ryan. This past February, overlapping with our last earnings call, I had the opportunity to visit Ukraine and witness the incredible speed with which the Ukrainian forces were able to employ AI on the battlefield. It was clear that the future has already arrived, and that future requires us broadly to rewrite road maps. It changes everything to some degree and some things completely. We all can either choose to join the disruptors driving this change or we can be disruptive. From diffusion to large language models, the accelerating pace of AI development is awe-inspiring and exhilarating. But there are many challenges customers will encounter as they attempt to leverage this technology operationally at any scale, from managing the mismatch between the ever-growing big data scale of the enterprise and the bottlenecks and choke points of smaller context windows of LLMs, to product design challenges of constructively embedding these models and workflows and safety and trust challenges in governing the AI in an operational and decision-making context. We are well ahead of the curve because we've built the frameworks, the infrastructure and the software needed. We have already had to solve so many of the hard problems, including the development of deep expertise with the necessary intermediary scale data challenges required to effectively span the gap from big data to small LLM context windows. And there will be many surprises and who ultimately comes up as the winners and losers of all of this disruption. The AI models themselves within months have gone from cutting edge to quickly being commoditized. Developing GPT-2 and GPT-3 class models at this point, table stakes. Anyone can build them in a few days with a few hundred dollars. As the memo purportedly leaked from Google titled, "We Have No Moat, And Neither Does OpenAI" makes clear. Things that big tech companies have considered major open areas of development have actually been solved by a handful of people in the open source community. While some of the proprietary models hold a slight quality edge, the lead is vanishing quickly. The speed of iteration on these models, biasing to a goldilocks size of power iteration pace ratio, will dominate. We believe larger, enduring value is more likely to emerge from the application and workflow layer by the players who know how to navigate the challenges of data scale mismatch that LLMs present, and we are uniquely positioned to continue to be a leader here. And because that future is already here, we must act with speed and conviction. Customers must completely rethink what they are building and how they build it. We are moving fast to ensure existing customers can quickly deploy AIP beside Gotham and Foundry to transform their operations with intelligent, contextual decision-making and sophisticated automation and coordination from the battlefield to the boardroom. AIP enables customers to operate not only Foundry and Gotham, but also their businesses from a greater strategic vantage point. We provide powerful interfaces to rapidly integrate your data, build your ontology, forge AI-driven applications and workflows and build agents to orchestrate and automate enterprise actions, all in an environment with guardrails, safe handoff functions and military-grade security. AIP not only knows how to speak to you, but more importantly, it knows how to speak to GAIA, MetaConstellation, every Foundry service, your ontology and your enterprise. At a major insurance company, we deployed a prerelease version of AIP and, in a few days, built a collaborative AI agent to automate claims processing. The Chief Data Officer told us that AIP was years ahead of anything they had seen. The pace of innovation around AI and LLMs presents a unique opportunity for leveraging our Apollo and FedStart infrastructure. There is tremendous demand for these models in FedRAMP, IL5 and IL6. Via our FedStart offering, we can bring these start-ups to U.S. government markets in weeks. In Q1, we also closed our first $1 million deal for Apollo with a major tech company. Our Foundry Ontology SDK has met the market to a great reception. These SDKs make it easy for customers to build new enterprise applications and integrate ontology into their existing ones. We plan to support AIP directly from these SDKs, extending the AI application forged deep into your enterprise ecosystem. The Gotham ecosystem continues to expand with enhanced capabilities across the AI-enabled kill chain. Our most recent developments include substantial new capabilities called Maverick and fires in the broader air battle management process. This capability enables efficient translation of approved targets to weapon systems. On a final note, we will be sharing our insights on the latest cutting-edge AI, who will be the victors and the vanquished and, of course, showing the latest across AIP and Foundry as well as hearing from our customers on how our products are powering their transformations at our June 1 customer event in Palo Alto. With that, I'll turn it over to Dave.

Thanks, Shyam. The first quarter of 2023 was a record-setting quarter for us. We achieved our second consecutive quarter of GAAP profitability and also achieved GAAP operating income for the first time. This accomplishment was the result of a strong top line growth, driven by the reacceleration of our U.S. commercial business, coupled with continued disciplined spend management. These achievements demonstrate our commitment to driving profitable growth, and we continue to expect 2023 to be our first full year of GAAP profitability. On the back of these exceptional results and this trajectory, we now expect to be GAAP profitable in each quarter this year. Turning to our global top line results. We generated $525 million in revenue in the first quarter of 2023, up 18% year-over-year, $20 million ahead of the midpoint of our prior guidance. Our U.S. business generated $337 million in total revenue in the first quarter, up 23% year-over-year and 12% sequentially, demonstrating our continued momentum in the U.S. Revenue from our largest customers continues to expand. First quarter trailing 12-month revenue per customer from our top 20 customers increased 14% year-over-year to $51 million per customer. New customer acquisition also remained strong as we saw customer count grow 41% year-over-year and 7% sequentially. Now moving to our commercial segment. In the first quarter of 2023, commercial revenue grew 15% year-over-year and 10% sequentially to $236 million. Our first quarter U.S. commercial revenue grew 26% year-over-year and 39% sequentially to $107 million. Excluding revenue from strategic commercial contracts, U.S. commercial revenue grew 46% year-over-year and 24% sequentially. In addition to the strength of our U.S. commercial revenue growth, our U.S. commercial customer count grew 50% year-over-year and 8% sequentially, marking the ninth consecutive quarter of sequential growth. Commercial revenue from our strategic commercial contracts surpassed first quarter expectations at $33 million due to the unanticipated acceleration of revenue from certain of these contracts. We expect second quarter revenue from these customers to be between $17 million and $19 million and to drop off significantly in the third and fourth quarter. Our expectations of full year revenue from these customers remains unchanged at approximately 3% of total full year revenue. In the first quarter, our international commercial business grew 8% year-over-year and declined 7% sequentially. The results were impacted in part due to a tough comparison to the prior quarter as a result of the timing of revenue recognition on certain contracts in the fourth quarter and the seasonal headwinds we often see in the first quarter with some of our largest international enterprise customers. Turning to our government segment. First quarter government revenue grew 20% year-over-year and declined 1% sequentially to $289 million. First quarter U.S. government revenue grew 22% year-over-year and 2% sequentially to $230 million. While we have a strong pipeline of opportunities in our U.S. government business, we acknowledge that there are uncertainties associated with the timing of these contract expansions and renewals. Nonetheless, we remain confident in the growth of our U.S. government business. First quarter international government revenue grew 11% year-over-year and declined 13% sequentially to $59 million. The sequential decline was driven by nonrecurring revenue catch-up that we saw in the fourth quarter as well as some challenges with the timing of contract awards in the first quarter. Turning to bookings. TCV booked in the first quarter was $397 million, up 60% year-over-year. U.S. commercial TCV grew 170% year-over-year to $124 million, again demonstrating the strength of our U.S. commercial business. We ended the first quarter with $3.4 billion in total remaining deal value and $936 million in remaining performance obligations. As a reminder, RPO is primarily comprised of our commercial business as it does not take into account contracts with the initial term of less than 12 months and contractual obligations that fall beyond termination for convenience clauses, both of which are common in most of our government business. Both remaining deal value and remaining performance obligations have continued to face headwinds from the macroeconomic impact on customers from the strategic investment program. As a result, the total remaining deal value and the total value of our commercial contracts from our strategic investment program decreased by $102 million since last quarter as we continue to review and assess the financial condition of these businesses. Turning to margin and expense. Adjusted gross margin, which excludes stock-based compensation expense, was 81% for the quarter. First quarter adjusted income from operations, which excludes stock-based compensation expense, was $125 million, representing an adjusted operating margin of 24%, 600 basis points ahead of our prior guidance and marking the third consecutive quarter of expanding adjusted operating margins. These results demonstrate our ability to drive strong revenue growth while managing costs effectively, with first quarter adjusted expense up only 1% sequentially to $400 million. We continue to manage expense growth by optimizing operations and G&A, capturing cloud efficiencies and focusing our headcount investments in key strategic areas. We remain committed to sustained GAAP profitability while, at the same time, increasing investment in both the U.S. to capture the momentum we are seeing and in the development of advanced software capabilities, particularly our AI-driven offerings. We believe that artificial intelligence, including large language models, will prove transformational for our business and for enterprises in the government and commercial context. To that end, we are rebalancing our efforts and investments to capitalize on these developments. In the first quarter, we generated income from operations of $4 million, our first-ever quarter of GAAP operating income. This reflects our laser focus on profitable growth and continued management of our stock-based compensation. While we expect to see stock-based compensation expense trend up through the remainder of the year, we remain focused on GAAP net income and operating profitability. Turning to net income. First quarter GAAP net income was $17 million, our second consecutive quarter of GAAP profitability. This was the result of our strong GAAP operating income as well as interest income from our balance sheet and the narrowing of losses from investments. First quarter adjusted earnings per share was $0.05, and GAAP earnings per share was $0.01, our second consecutive quarter of positive GAAP EPS. We are extremely proud that we were able to continue to deliver GAAP profitability on a consistent basis. Additionally, our combined revenue growth and adjusted operating margin was 42% in the first quarter. We will continue to strive to achieve the Rule of 40 throughout 2023 and beyond. Turning to our strong cash flow in the first quarter. We generated $189 million and $187 million in adjusted free cash flow and cash from operations, respectively, each representing a margin of 36%. We ended the quarter with $2.9 billion in cash, cash equivalents and short-term U.S. treasury bills. We retain access to additional liquidity of up to $950 million through our $500 million revolving credit facility and $450 million delayed-draw term loan facility, both of which remain entirely undrawn. Now turning to our outlook. For Q2 2023, we expect revenue of between $528 million and $532 million, adjusted income from operations of between $118 million and $122 million and GAAP net income. For full year 2023, we are raising our revenue guidance to between $2.185 billion and $2.235 billion. We are raising our adjusted income from operations guidance to between $506 million and $556 million, and we now expect GAAP net income in each quarter of this year. With that, I'll turn it over to Ana to start the Q&A.

Thanks, Dave. We'll begin with a few questions from our shareholders before we open up the call. We received a lot of questions on AI and AIP. Alex, would you like to share some thoughts?

Well, thank you for being here. I've been at the Palantir building since the beginning for 20 years. We -- our first product was PG, which changed the course of history, especially in Europe, by preventing terror attacks in a constitutional order. And that product basically is the only product for what -- for that use case in the world. And then we built a number of other products, products on the battlefield for special operators, Foundry, which needs no introduction. And in the process of building these products, we were always thinking about human mind, human-computer symbiotic relationships, and then later, algorithmatic relationships to data, especially what's called generative AI. And so we had built these precursor technologies over the -- since inception and also biases. So how would you interact with an algorithm where you expose only the part of the enterprise that you want to expose, need to expose? How do you do a handoff when you're doing -- deciding to make a lethal decision? Who makes the lethal decision? What data sources are used? How do you map the knowledge and wisdom of your enterprise onto the external database? And then with the rise of large language models, how do you impose that on the large language model? How do you take what's useful from AI, whether -- especially in a large language model context, without hallucinating your way into a disaster? How do you expose sensitive parts of your enterprise that need to be exposed while protecting parts of your enterprise that should never be exposed, not just in the classified environment, but if you're working in a hospital context, not exposing health care, personal health care data to a broader LLM run by a third party? How do you do this in a context where some people are technical and nontechnical? How do you expand our offering -- and this is one of the big advantages of the current revolution to people who have very technical needs, but don't know how to wield them. And in general, our business has been a business driven by moral imperatives. One of my -- I don't want to make this overly philosophical, although, in some ways, I prefer it. But when you look at the history of Silicon Valley over the last 20 years, there's, in a weird way, a tacit alliance between the way in which our maybe adversaries think of the world and the way in which many people in Silicon Valley have thought about it. Essentially, we will use data to serve our interest, not inherently the interest of the people, meaning, in our context, should be monetized. And we have always believed in the moral imperative where the human and our governments would define the context in which data and algorithmatic and now large language models implicitly would be controlled. And so that looks like a weird philosophical bias. But in fact, that philosophical bias allowed us to develop branching, allowed us to develop ways of doing handoff technologies, that allowed us to think broadly about what it means to be in lethal situations and behave ethically. It now positioned us to be able to launch ways in which to deal with large language models that are -- is a once-in-a-generational opportunity with technologies that are otherwise not available by other people. Also, even our bias of understanding, as Shyam was saying in his introduction, that it is the person that wields the technology that makes it valuable, that makes it moral, not the technology itself. And these biases have just put us very, very far ahead of the market. And especially, we're seeing this reception to this in the United States, which makes me obviously particularly happy. And so we're going to be very involved in rolling this out in a kind of pretty disciplined way. We see this unprecedented once-in-a-generation opportunity. We have odd resources in terms of IP, software platforms that are uniquely suited for this. We have an ability to run hard, both because we are entrepreneurial. Many people in the space, in fact, don't really build products, they buy products. It's just too late to buy a product to catch up with what has to be done now and where the winners will really accelerate. So we're -- we feel like we're in a -- obviously, our profitability is really important to us, particularly because as we begin to empower more and more of the most important institutions in the world, it basically takes off the table the legitimate question, well, if your technology is so powerful, why can't you make any money? So I believe, it's a unique time for us, a unique place for us, and we're very, very busily at work.

Thanks, Alex. We also received a few questions on GAAP profitability. Nishal and Rick asked, do you still expect to be GAAP profitable this year?

Nishal, Rick, thanks for the question. In short, absolutely, we do. And really, on the strength of Q1, we now know -- we expect to be profitable in each quarter of this year, not just for the full year, which is obviously a big milestone for us. And to recap Q1, it was our second consecutive quarter of GAAP profitability. We did $17 million of net income. It's our first-ever GAAP operating income profitable quarter. We did $4 million of operating income. It's our strongest cash flow quarter ever, $189 million of free cash flow. And we did this all with the U.S. commercial business growing 39% sequentially. And so profitability, it will remain core to our growth, and that's as we mobilize our business and our resources around the AI opportunities that are in front of us.

Thanks, Dave. We received questions about expanding our customer base and the growth and momentum of our commercial business. Ryan, do you want to share some thoughts on the commercial business for 2023?

Yes, absolutely. So in our results, you see the reacceleration of our U.S. commercial business, 26% year-over-year growth, U.S. commercial, as Dave said, 39% sequential. Our customer count, which is now at 155 customers in our U.S. commercial business, grew 50% year-over-year, and the TCV closed in Q1 grew 170% year-over-year. That's on the back of pilot conversions, new pilots starting and expansions at customers, as I highlighted earlier, such as Hertz, Jacobs Engineering, the largest health system in the country and one of the largest paper and packaging companies in the world, among other examples. We expect that momentum to carry forward. And as the U.S. commercial continues to become a larger and larger part of our overall business, that will obviously have larger and larger impacts on our overall results and momentum as well. On top of that, we've highlighted AI presented an opportunity that is creating unprecedented demand for us. We can deliver against it in ways that no one else can. And we're running at that full speed, and we expect that momentum to continue in 2023 and beyond.

Thanks, Ryan. Our next question is from Michael, who asks, with potential S&P 500 inclusion, how does that correlate with the company's prospects on profitability moving forward?

I'm sure there's a prepared answer here. But I believe -- I know that we are outsiders, and this outsider status is one of the main reasons we built all these products because we never, ever thought anyone would buy our products, except that they were disruptive and unique. And that's -- in many -- if you look at the products we built, but also in somehow our uncanny preparedness for this moment, it's because we approach this as we have to succeed under adverse conditions. And my interest in profitability is for obvious reasons, but it's also, I think, we'll just be in a much stronger position as we -- it becomes clear that we qualify for participation in S&P. And a lot of people look at us afresh, anew and will begin to look closer at our strengths. But I think this is -- the outsider strategy is, we are going to be profitable. We're going to -- within the context of being profitable, we are going to take all resources and disrupt, especially in the USA, and we are going to use the profitability to scale our business and to scale the attractiveness of our business to new investors.

Thanks, Alex. Our next question is from Brent with Jefferies. [Operator Instructions]

There were a lot of questions about in the quarter. The backlog growth significantly decelerated, yet the comments about the overall business reaccelerating, you guys were clear about. And so there were questions around were there any onetime milestone [ GV ], rev rec or SPAC contribution in the quarter that may have explained this? And can you just comment about the backlog growth versus the comments you said about reacceleration because they don't really match up when we look at the numbers?

Yes. So happy to take that. And I think just to start off, when you're looking at reacceleration, I'd point to the U.S. commercial business is really driving our business forward and sort of where our focus is, and it did reaccelerate significantly in Q1, right? It was 26% year-over-year growth, 39% sequential. If you want to exclude SPAC revenue, it grew over 45% and grew 24% sequentially. So like it's exceptional reacceleration. And then if you want to look at bookings, you can look at U.S. TCV, up 170% -- excuse me, U.S. commercial TCV is up 170% year-over-year. So like all the measures are there.

I think -- look, I didn't really give you a somewhat orthogonal answer your question. We have a tale of 2 cities here. We have America, which is growing around 28%. It is now 64% of our business. Four years ago, it was 37% of our business. We are absolutely disrupting in the US of A. International is growing around 10%, and that is becoming obviously a smaller part of our business. And so what you -- what I tend to see when I look at the numbers is, yes, if we believe the U.S. is -- we see the U.S. is growing, we see the -- that U.S. is really accelerating, although reaccelerating is the wrong word because, in fact, U.S. commercial has been accelerating and accelerating, and it's just become a much bigger part of our business. And last, not least, what's not in the numbers is, I spend a lot of my time on the road. I'm talking to the U.S. customers. I'm talking, and they're not -- the quality of the questions has shifted from why would we need this to how would I use it? And that's just purely on the Foundry front. And then the inbound on anything related to AI, both from existing customers -- by the way, on that score, we sit on the world's most important private sensitive networks, both in commercial and government. Every single one of those clients needs an AI strategy, and we have unique technology and software for that. So -- but I think, if you wanted to look at the positive or negative of our business, like, yes, sure. We -- I do not expect international to grow much more than it's growing now. I do expect the U.S. to continue to grow, and we're very optimistic about what will happen there.

Thanks, Alex. Our next question is from Mariana with Bank of America. [Operator Instructions] Mariana, can you hear us?

Hello?

Can You hear me?

Yes.

Perfect. I am [ powerless] with technology, still. There, I am muted. Okay. My question is about the...

We can't hear you.

Sorry, Mariana, you're cutting out.

I'm so sorry.

No problem. We can just move on. Our next question is from Keith with Morgan Stanley. [Operator Instructions]

Can you guys hear me?

Yes.

Excellent. So we talked a lot about accelerating U.S. commercial business, reorienting the company to a certain extent for this commercial opportunity with the AI platform. Two questions. One, can you talk to us a little bit about changes that you're looking to make in the go-to-market for this AI platform? Any sense you could give us in terms of how you're looking to price the platform? What's sort of like the initial conversations been like? And what do you think the pricing modality is going to be like around that AI platform? And then the last question on GAAP profitability. This quarter, you did have a pull-forward of revenues that kind of helped the commercial business, helped the overall profitability. In some sense, you got to make that up a little bit in Q2 and Q3. Are there any tactical things that you're doing or cost reduction initiatives that you're putting in place to ensure that GAAP profitability to offset that pull forward, that upfront revenue recognition we saw in Q1?

I'll take the AI thing, and Shyam, maybe, you want to add something and then you can talk about all of our various strategies to be more and more efficient. No, our basic -- it's counterintuitive, but we're going to -- within the context of remaining profitable, our strategy on AI is to just to take the whole market. We have no pricing strategy. We're going to create a lot of value. We're going to get hundreds of customers, and we will price it as we go. One of the things we've seen over and over again is when you're ahead of the market, you need to take territory. We want to take territory for 2 reasons. One, we believe we have the only product currently on the market that can solve some of these intricate technical needs. Two, we've spent years building certain things like how do you manage knowledge systems across algorithms, across security interest. We know how hard just to build this. We believe that people will take years for people to build these things. And three, because as we take territory, we educate the market. And the market, quite frankly, many analysts believe this stuff is really easy. You educate the market and to see how hard it is, and in doing that, you then have pricing parity because they will try other things that, for technical reasons, are likely to fail. So our -- especially in the U.S., we have some non-U.S. customers. There are some exceptions to this. We can't do this for free for governments for legal reasons and because we have so many already AI-driven clients. But in the U.S. commercial context, our strategy is going to be to take territory, educate the market and make it very difficult for other people entering the market because they'll have to compete with a product that we've already been working on for de facto years with solving technical issues that the market doesn't yet realize exist. Then I'll leave...

Yes, I was just going to say, from a product perspective with AIP, we're just razor focused on creating value as quickly as possible. I think it's not actually clear from the outside how far ahead we are. If you think about the intermediary scale data challenges that you face and getting value out of LLMs, and they're extremely small contact windows, like how are you going to bridge the gap between big data in smaller context windows? You need an ontology. And all of these -- like how did we build an insurance GPT agent in 2 days? Because we stood on the shoulders of 2 years of a very robust ontology that enabled us to unlock huge value.

Maybe you should [ explain ] what an ontology is because for most people -- yes.

Yes. I could take the rest of the call with that. But I simply...

You'd love that. It's more interesting than that, but -- yes.

You have the semantic representation of the business. Sometimes people call it digital twin. I hate that because it cheapens it in how it's been diluted. But like what is the business? How do we think about it? How does the data map to the operations that happen? And then you have the kinetic layer of like, well, how do we make decisions? How do I adjudicate a claim? What does that mean? How does it flow through the system? These are prerequisites. You can think about that second piece, the kinetics is the tools you give the LLMs to actually drive and automate your business processes. And you think about the former part of like how do I actually reason about what's going on and I need that, it's almost like a form of compression to get everything into that small context window, right? Like that's a physical limitation of the LLMs. So I think the amount of ground we can take means that the only reasonable strategy is to focus on...

And implicit in what Shyam is saying is, without essentially other people called a knowledge gap, in fact, de facto, you could view it as like a dynamic semantic layer where the assumptions of your world are built in it and can map across your -- you will just not be able to use this technology. Now that's something we know. No one else knows. No one believes it, but they're about to find out. And so we don't have to have this debate with people. We're just going to show clients, "Hey, this will work, you will get value. We've done this before. It's also crazy hard to build this." So if you know how to do it, you have a 3-year build.

And we didn't even talk about the guardrails, the military-grade security, the safe handoff functions that are extent, right? Like we can go running at this so hard.

And again, in a highly regulated environment, whether people listening dislike it or not, America is a highly regular environment. You will not be able to use this technology without what Shyam is calling guardrails. What he really means is with a branching technology where [ ackles ] can be applied to the internal systems in a way that you can still get the work -- be able to do the work because you have a semantic layer. And that's all gobbledygook to people, except for it will not be gobbledygook to everybody who wants to use this technology in a productive way. And we just -- we spent 20 years building these things, and we're going to litigate that on the front and figure out how we get paid later.