C3.ai, Inc. (NYSE:AI) Q3 2021 Earnings Conference Call March 1, 2021 5:00 PM ET
Paul Phillips - Vice President of Investor Relations
Tom Siebel - Chairman and Chief Executive Officer
David Barter - Chief Financial Officer
Conference Call Participants
Brad Sills - BofA Securities
Jack Andrews - Needham
Michael Turits - Keybanc
Dan Ives - Wedbush
Mark Murphy - JP Morgan
Patrick Colville - Deutsche Bank
DJ Hynes - Canaccord
Arvind Ramnani - Piper Sandler
Pat Walravens - JMP
Ladies and gentlemen, thank you for standing by and welcome to the C3.ai Third Quarter Fiscal Year 2021 earnings call. At this time, all participants are in a listen only mode. After the speaker's presentation, there will be a question-and-answer session [Operator instructions].
I would now like to hand the conference over to your speakers today. Paul Phillips, Vice President of Investor Relations at C3 AI. Thank you. Please go ahead, sir.
Good afternoon, everybody. And welcome to C3.ai's earnings call for the third quarter of fiscal year 2021, which ended January 31, 2021. This is Paul Phillips, Vice President of Investor Relations for C3.ai. With me on the call today are Tom Siebel, Chairman and Chief Executive Officer; and David Barter, Chief Financial Officer. After the markets closed today, we issued a press release with details regarding our third quarter results, as well as a supplement to our results, both of which can be accessed on the investor relations section of our Web site at ir.c3.ai. This call is being webcast and a replay will be available on our IR Web site following the conclusion of the call.
During today's call, we will make statements related to our business that may be considered forward looking under federal securities laws. These statements reflect our views only as of today, and should not be considered representative of our views as of any subsequent date. We disclaim any obligation to update any forward looking statements or outlook. These statements are subject to a variety of risks and uncertainties that could cause actual results to differ materially from expectations. These risks are summarized in the press release that we issued today. For a further discussion of the material risks and other important factors that could affect our financial results, please refer to our filings with the SEC, including our prospectus.
Also during the course of today's call, we will refer to certain non-GAAP financial measures. A reconciliation of GAAP to non-GAAP measures is included in our press release. Finally, at times in our prepared comments or responses to your questions, we may discuss metrics that are incremental to our usual presentation to provide greater insight into the dynamics of our business or our quarterly results. Please be advised that we may or may not continue to provide this additional detail in the future.
With that, let me turn the call over to Tom for his prepared remarks, Tom.
Thank you, Paul, and good afternoon, everyone. It's my great pleasure to have the opportunity to spend some time with you this afternoon. So let's talk about the company and talk about the market. Overall, we're really quite pleased with the progress that we continue to make in the market globally. We continue to break ground as the only enterprise AI software pure play. This is a large and rapidly growing market. We continue to innovate. We continue to expand our market partner ecosystem and the associated increased distribution capacity associated with that. We continue to demonstrate technology leadership. I believe that we are increasingly well positioned to establish a global market leadership position in enterprise AI software. So let's talk about the financial highlights. All in all, it was a strong third quarter. Revenue in the third quarter was $49.1 million. $42.7 million of that was subscription revenue, an increase of 23% from a year earlier. Subscriptions have increased to 87% of our revenue mix. Services revenue for the quarter was $6.4 million or 13% of total revenue, representing a decrease of 4% from a year ago. Non-GAAP operating expenses for the quarter increased to $49.2 million, up 27% from a year ago and non-GAAP operating loss for the quarter was $11.9 million compared to $8.4 million year ago.
Let's talk about the overall state of the market. When we think about enterprise AI and digital transformation, we are focused on an extraordinarily large addressable market that according to analysts was $174 billion in 2020, increasing to $271 billion by 2024. This is a significant opportunity by any standard and the largest software market opportunity that I have seen in my professional career. Digital transformation enabled by enterprise AI remains at the top of the agenda of virtually every CEO and board globally. We see increasingly robust interest and demand for enterprise AI solutions, and that our pipeline continues to grow substantially across all industries in all regions.
Let me address a few of the significant customer wins in the third quarter. Firstly, Infor. We have formed a strategic wide ranging relationship with Infor, a multibillion dollar ERP software provider. Under the terms of the agreement, Infor will be integrating many existing Infor applications with the C3.ai suite and the C3.ai applications to market, sell and deploy AI based solutions to their customers under the Infor brand. This will evolve into a broader partnership with Infor to market C3.ai solutions to Infor customers, including C3.ai reliability, C3.ai CRM, and C3.ai Ex Machina, a next generation predictive analytics application that empowers anyone to develop, scale and produce AI based insights without writing code. Secondly, I want to talk about Johnson Controls in Milwaukee. Johnson Controls recently selected C3.ai for their inventory and supply chain optimization. Johnson Controls is a $23.5 billion company with leading manufacturer of fire, HVAC security equipments for building and energy services.
Finally, I want to talk about an opportunity at the United States Army, where we expanded our work with US Army Aviation to improve aircraft readiness. In addition, with our partner Raytheon, we recently began work on the US Army's next generation tactical ground station that will enable automated processing and analytics of massive amounts of incoming sensor data for improved situational battlefield awareness. This is very, very high tech, think Star Wars level technology. This AI based system will provide decision support to commanders based on information coming from a multitude of sensors and sources within the C3.ai suite, synthesizing risk into a single unified data image.
I want to talk about industry diversification. We continue to diversify across industries. We're seeing increasing contributions from the life sciences, financial services and the high tech sectors, as well as increasingly healthy growth in international sales. High tech accounted for 8% of our business in fiscal year '21 year to date compared to less than 1% for the same period a year ago. Life sciences accounted for 7% of our business in fiscal year '21 year to date, compared to less than 1% for the same period year ago. Manufacturing sector accounted for 11% of our business in fiscal year '21 to date compared to approximately 1% over the same period a year ago. At the same time, oil and gas accounted for 38% of our business in fiscal year 2021 year to date compared to approximately 59% a year ago.
Addressing strategic partnerships. Core to our market growth strategy is; one, the evolution of our global direct geographic and vertical market sales organizations to include major accounts, enterprise sales, mid-market sales and mass market sales channels; and two, the expansion of our market partner ecosystem in both horizontal and vertical markets. We realized significant benefits from these efforts in the third quarter with increased geographical expansion, increased vertical diversification, significant growth in software subscriptions and importantly, increased volume and diversity in the average contract in our sales pipeline. By way of example, our average total subscription contract value decreased from $12.1 million in fiscal year '20 to $5.2 million in Q3 fiscal '21. This is substantial progress. We continue to build and increase our engagement with our partner ecosystem as a key growth strategy and enabling us to significantly expand our market reach and serve customers across industries globally. In the third quarter, we established for expanded such partnerships with Microsoft, Baker Hughes, ENGIE, Shell, Infor, FIS and Raytheon, and you can expect to see continued activity, accelerated activity in this area going forward.
In the financial services sector, our industry partner, FIS AML compliance hub powered by C3.ai. The first joint product release under a broad alliance between FIS and C3.ai. This solution leverages C3.ai’s advanced machine learning technology, combined with the deep financial industry domain expertise of FIS, to dramatically improve the effectiveness of financial crime detection. C3.ai and Microsoft closed our first AI CRM deal with a Fortune 500 customer. We believe AI will represent a significant and growing part of the $60 billion plus dollar CRM market globally. In the defense sector, we engaged with our defense industry partner Raytheon to provide technology for the US Army’s intelligent ground station initiative. As I referenced earlier, we formed a comprehensive alliance with Infor to integrate many existing Infor applications using the C3.ai suite and C3.ai applications, and enabling Infor to market these applications to its customers through all its channels. You can expect that in the coming quarters, we will make increased efforts to expand our market partner ecosystem across both horizontal and vertical markets as we see this as a key point of leverage for market growth.
We focused a lot on strengthening company leadership. And we expanded the company's leadership with the addition of Jim Snabe, formerly co-CEO of SAP and Chairman of Zeeman's, also Chairman of Mayers, to the C3.ai Board of Directors. I've come to rely on Jim as a trusted advisor as he has been an advisor to the company for some times, he’s been advisor to me for some time. Okay. As we work to establish C3.ai as a leading global enterprise software company, he brings a unique set of leadership skills and expertise to the company and is a great addition to our world class Board.
In addition, we continue to expand our global advisory board that has been really, really important in establishing customer relationships, okay, and building presence in country and in any industries. The advisory board now includes Jacques Attali, who is founder and first President of the European Bank for Reconstruction and Development and former special advisor to the President of France; Sajid Javid, who is a Member of Parliament and former Home Secretary and former Chancellor of Exchequer in the UK; Admiral Danny McGinn, formal US Assistant Secretary of the Navy in the Obama Administration; Rick Ledgett, former deputy director of the NSA in the Obama Administration; Franck Cohen, former President of SAP in Europe and George Matthew, former President and COO of Alteryx.
Now, those of you who have been tuned into financial channels or watching the wall street journal or reading the financial times or Bloomberg, you know I have noticed that we have been investing in branding globally. And we expanded our enterprise AI branding campaign to include significant cable TV and radio presence with spots running on virtually all the major business and financial networks in the United States and Europe. We also expanded our advertising to include the UK, EMEA and APAC, as well as United States. We are becoming increasingly known as the enterprise AI market leader. Leading indicators of brand equity and brand recognition have been substantially increasing, including PRV, public relations volume, news sediment, social media sediment and internet search frequency.
Let’s talk about energy sustainability. Well, the energy and sustainability market is back and it is on fire and so we are seeing increased interest for AI enabled solutions in the energy and sustainability market. This market is coming fast strong. Over the last year alone, the number of companies and governments that have committed to reach net zero emissions has doubled. No doubt, you've seen the flood of announcements from companies, such as ENGIE, Shell, Baker Hughes, Microsoft, Amazon, Ford, BP, JetBlue, American Airlines, and others announcing their zero carbon commitments. Sustainability is also clearly a top priority for the new US administration targeting $2 trillion investment in climate security. C3.ai is a leader in this space. Energy and climate sustainability is the core market many of you will recall where we started with our first product offering in 2010, C3 Energy Management, which we have deployed in production use with many large organizations globally, including ENGIE and the New York Power Authority.
One iconic Fortune 100 company has been a customer since 2012. They have set ambitious sustainability goals and use our energy management application to identify and prioritize their energy efficiency and carbon reduction investments globally to meet their climate goals. Many utilities use C3.ai technology and C3 Energy Management not only to optimize their own energy usage and optimize the grid infrastructure, but also to help their customers meet their energy efficiency and greenhouse gas goals. A great example is New York Power Authority that deployed C3.ai Energy Management as a service, enabling its large commercial and industrial customers to achieve their energy efficiency and greenhouse gas goals. And with our energy partner ENGIE, we're delivering innovative solutions built with C3 Energy Management. Ohio State University, for example, has deployed ENGIE Smart Institutions built on C3.ai technology to reduce and manage energy use and carbon footprint across its entire 485 building campus in Columbus. We're also working with ENGIE on a novel cutting edge solution addressing the very difficult challenges of managing Scope 3 emissions, a problem that requires an entirely new technology approach.
In the third quarter, we significantly expanded our effort -- our investment in the C3.ai Digital Transformation Institute, issuing a new call for papers to find innovative research and applying AI and digital transformation to energy and climate security. We're very excited to support this next wave of research by providing [enabler], safer, cleaner, lower cost and more reliable energy and to help lead the way to a lower carbon future. We are extraordinarily well positioned for this new opportunity. Many of you will know that C3.ai was recognized by Glassdoor as the best place to work during the pandemic. We were ranked among the top 25 cloud computing companies with stellar employee satisfaction through the COVID crisis. This ranking is based on employee feedback provided through Glassdoor during the first six months of the pandemic. I encourage those of you who are interested to take a look at Glassdoor, to get a feel for the high levels of employee engagement and enthusiasm in C3.ai.
As you know, we placed an exceptionally strong focus on human capital and C3.ai, and are aggressively expanding the company. In a few short months since the company went public in December, we've increased our headcount by approximately 10% to roughly 560 today. We continue to attract massive interest from the global data science and data engineering software pool. In the third quarter, we received over 17,000 job applicants, we conducted 30,966 interviews with over 1,700 candidates and administered almost 700 assessments. Of that, we hired 60 full time employees and we've processed an additional 57 job offers that have been accepted. One of the unique aspects of C3.ai is our clear technology and product differentiation. And we continue to protect the company's intellectual property with a combination of trade secrets, copyrights, trademarks and a growing family of us and foreign patents, with 13 US and foreign patents having been awarded and 40 US and foreign patent applications pending. The US Patent Office recently awarded us a broad and important patent entitled systems, methods and devices for an enterprise AI development platform. This is the most significant and substantial patents have been awarded to date. This patent essentially secures the fundamental concepts of applying a model driven architecture for enterprise AI applications and it secures as C3.ai intellectual property and we are working now to form a technology licensing office at C3 to license some of this IP to companies that elect to attempt to internally develop their enterprise AI applications.
Big picture, we see a robust and growing interest in enterprise AI software, the solutions that we offer. We continue to make significant progress on all fronts and our objective to establish and maintain a global leadership position in enterprise AI. We continue to aggressively grow the company, diversify our business across industries and geographies, and expand our partner ecosystem. We believe we are very well positioned to address this $200 billion plus addressable market opportunity and we are just getting started. With that. I will turn it over to our CFO, David Barter, for further details on our financial results in the third quarter, David?
Thank you, Tom. We delivered a strong performance in the third quarter, and we're optimistic about the growing demand we’re seeing for our software. Revenue in the third quarter was $49.1 million, up 19% from a year ago, reflecting continued business momentum and increased adoption of enterprise AI. During the third quarter, subscription revenue was $42.7 million, an increase of 23% from a year ago. Professional services revenue was $6.4 million, a decrease of 4% from a year ago. We also saw an increasing diversification of our revenue mix. Our revenue growth in the quarter was highlighted by contributions from over six different industry verticals, including some of our newer verticals, such as life sciences and financial services that delivered approximately 24% of our third quarter revenue. Geographically, EMEA and APAC drove over 30% of our revenue. Finally, it's important to note the improving operating efficiency. This quarter subscription revenue increased 23% year-over-year while non-GAAP sales and marketing expense increased 14% year-over-year.
An important aspect of our model is the usage of market partners. It is a leveraged selling model that significantly extends our sales reach into each industry vertical. Our partners possess deep domain expertise in their respective vertical markets, a commitment to enterprise AI and a significant customer base that will benefit from our enterprise AI applications. This will provide us with sales leverage overtime. It is worth noting that our contract with Baker Hughes, our market partner for oil and gas, includes a very specific contraction commitment. It is a five year agreement and the total contract value is $450 million. Their commitment to C3.ai increases over the five year term of the agreement. This offers predictable growth in the years ahead. For example, the Baker Hughes commitment is $75 million in fiscal year '22 compared to $53 million in fiscal year '21.
Our contractual backlog continues to provide us with healthy revenue visibility. It is important to note that we use adjusted RPO to measure our backlog. This metric includes our GAAP remaining performance obligations. Our contracts with the cancellation clause and the Baker Hughes commitment. At the end of Q3, adjusted RPO was $538 million, up 16% from the prior year. Within adjusted RPO, the GAAP RPO was $247.5 million. Our contracts with a cancellation cost was $48.4 million, and the Baker Hughes commitment was $241.8 million. With the healthy revenue visibility and coverage that comes from an unusually large contractual backlog, we continue to focus on expanding our footprint within existing customers, adding new market partners and adding new customers in existing and new industry verticals.
In discussing our expenses and profitability, I will be referring to non-GAAP measures unless otherwise indicated. A GAAP to non-GAAP reconciliation is provided with our earnings press release. This can be found in the IR section of our Web site and it is on file with the SEC. The difference between our GAAP and non-GAAP financial measures in the quarter was stock based compensation expense. Gross margin in the third quarter was 75.9% compared to 73.8% a year ago. This expansion was driven by subscription gross margin. It increased to 84% compared to 74.7% in the prior year period. Operating expense increased to $49.2 million, up 27% from a year ago. As Tom described, we are making significant investments in R&D in order to bring new products to market that can enhance our growth. C3.ai CRM and C3.ai Ex Machina are two such initiatives. We are also investing in our go to market efforts, including the expansion of our direct sales force. We're also investing in brand awareness, market education and enterprise AI thought leadership. Operating loss for Q3 was $11.9 million or a margin of 24.3% compared to $8.4 million or a margin of 20.3% a year ago.
Turning to our balance sheet and cash flows. We ended the third quarter with $1.12 billion in cash and cash equivalents. This includes net proceeds of $844.6 million from our Initial Public Offering in December. Our operating cash outflow for the period was $24.7 million due primarily to increased investments in sales, marketing and headcount. CapEx in Q3 was $0.2 million. This led to free cash outflow of $24.9 million. The timing of our billings and collections can vary. In order to capture a complete picture of our cash flow margin is recommended and calculated on a rolling four quarter basis. For the last four quarters, our free cash flow margin was a negative 24%. Deferred revenue was $63.3 million at the end of the quarter. It's important to note that deferred revenue is not a perfect measure for our business due to the quarter to quarter variability and the timing of invoices to our customers, and our historical large transaction sizes. In addition, our billing terms vary. For some contractual arrangements, we are not billing at inception. But instead we will build a customer periodically over the duration of the arrangement. So these customers’ revenue commitments do not appear in deferred revenue. As an example, at the end of the third quarter, we booked several deals that will generate $4.1 million of revenue, but this amount is not in our deferred revenue.
Looking to the fourth quarter. We expect total revenue to be in a range of $50 million to $51 million, representing approximately 21% year-over-year growth at the midpoint of the range. Non-GAAP loss from operations is expected to be in the range of $28 million to $27 million. For full fiscal year 2021, we expect total revenue to be in the range of $180.9 million to $181.9 million, representing approximately 16% year-over-year growth at the midpoint of the range. Non-GAAP loss from operations is expected to be in the range of $50.1 million to $49.1 million. Historically, the difference between our GAAP and our non-GAAP financial measures has been limited to stock based compensation expense. Beginning with this guidance, for the fourth quarter and full year fiscal 2021 and in future reporting periods, the difference between GAAP and non-GAAP measures will include stock compensation expense and the employer portion of payroll tax expense related to stock transactions. In summary, we are pleased with our strong performance in our first quarter as a public company. Our results and outlook reflect growing market demand from enterprises across an increasing range. With continued investments in multiple growth drivers, we're increasingly well positioned to capitalize on a large multibillion dollar market opportunity, generating value for all of our stakeholders.
Thank you for joining today's call. Now I'll turn the call over to the operator for questions. Operator?
[Operator Instructions] Your first question comes from the line of Brad Sills with BofA Securities.
I wanted to ask about the vertical partner focus here. Obviously, you talked about some leverage that you'll see here from some of these partnerships could you help us understand for perhaps some of the newer ones like Raytheon FIS, these are relatively new verticals for the company. What kind of resources are committed from these partners? How are you going to market together? How are you expected to get that leverage through these partnerships? Thank you so much.
I’ll fill this one because it’s kind of sales to marketing related. Well, as you know, we're going to market across three plains. We have horizontal market partners like say Microsoft would be the largest, but also IBM NVIDIA. We're building a geographic marketing organization in North America, Asia Pacific, okay, and in Europe. And then across all sectors, we have vertical market sales organizations in oil and gas, in utilities, in financial services, in precision health, et cetera. And you can get expect that each of the -- the goal is that each of these vertical markets we will align with a leveraged market partner. So the classic case is Baker Hughes. So we've aligned in oil and gas with Baker Hughes, this is a 24 roughly, I think, billion dollar oil services company that gives us access to 12,000 people now selling with us around the world. And there’s virtually not one of the largest, say 20 or 30 oil companies, that we’re not in active sales motion with, whether it's Aramco, ADNOC, Rosneft Gazprom, Shell, and 12,000 salespeople is a lot of sales capacity. If we look at financial services as you know we have a couple of -- we have very large and successful relationship in banking with Bank of America. And with a number of applications there and with Standard Chartered Bank and we’ve built a large product line to meet the needs of financial banks, whether it's anti money laundering, cash management, securities lending, Volcker Rule compliance, inter day liquidity, what have you. And so through our partnership with FIS, which I think does roughly $12 billion or $13 billion from the revenue in the banking, these people are now we have access to their I think 20,000 banking customers around the world with their entire sales organization. So this is the entire FIS sales organization.
As it relates to Infor, I don't know how many sales people Infor has, I'll have to look it up. But it's a ton of sales people. These guys sell billions of dollars in software. And now they're selling our software solutions as core to what they're doing. So we're dealing with, by the time you roll in companies like what we're doing with Microsoft and Adobe, in okay, in CRM, what we're doing with oil gas and Baker Hughes, what we're doing in banking with FIS, okay, what we're doing in manufacturing and particularly, with Infor, we're dealing with literally tens of thousands of people that are selling for us around the world. And if we're able to successfully execute this strategy in the next two, three years, honestly, I think we can put the lights out on the market before anybody else gets here. So this is core to the strategy. I think nobody has ever [encountered] this before. We believe it's a core competence for us. And expect and specifically people like me and [Indiscernible] who you know to be spending a significant amount of time on this. And then we brought in a very senior executive by the name of Gene Reznik, who was the Chief Strategy Officer at Accenture, here to head up this initiative of coordinating the vertical market partners. So we expect a significant investment here and expect in the coming quarters, there will be additional announcements with these vertical market partners that we think will give us a -- increase our competitive advantage in the market.
And then one more, if I may, please, the concept of embedding C3’s platform into some of these horizontal categories. You mentioned an early win with the Microsoft partnership here for C3.ai enabled CRM. Infor as a new partner for ERP. Can you remind us a little bit on kind of where you are with those partnerships? I know they're early. How quickly does it take one of these partners to ramp up from an integration standpoint and then go to market? When do you think you'll see the leverage out of those two partnerships in particular? Thanks.
I think we did at Baker Hughes now, if I'm not mistaken five quarters. Is this about right, guys. Okay. And Dave, do you…
Seven quarters. So guys, I stand correctly. Okay, thank you. And I think Baker Hughes is a large global company, and they're fully spooled up. I mean, we're building probably four products together we’re bringing to market. We have two products that we already have in the marketplace. We're fully spooled up. Granted Infor is brand new. Granted FIS is relatively new. And so you can expect us -- I mean it'll take us one, two, three, four quarters to get these things really spooled up and get the momentum going. And it's not going to turn out overnight but we're building a revenue capacity flywheel that I think, you know, once you get in, when you start looking at calendar year '22, '23, I think this is going to be a significant force to be reckoned with.
Your next question comes from the line of Jack Andrews with Needham.
I want to see if you could expand a little bit on your go to market strategy around the Ex Machina product. I believe this is somewhat of a different offering from the solutions that you've historically sold. So how should we just be thinking about your strategy? And is this something that you expect the uptake to be mainly from your existing base of customers or do you expect Ex Machina to effectively draw in some brand new customers to the C3.ai family?
So Ex Machina is a product that we've had in the market for some years, and we've been working and where we are now is we are releasing this and we've done all the work with the documentation, the stack overflow, the online training, the community, basically offer this as a mass market product. You can think of this as serving the same market needs that’s currently served by Alteryx, which is basically associated with the democratization of data science, this is allowing people who would normally be doing analytics with pivot tables, okay, on an Excel spreadsheet to be doing a no code, low code, WYSIWYG, what you see is what you get, point and click drag and drop data science. And what we've released here is kind of a 21st century version of that solution. And the democratization of data science is something that has to happen. I mean, for when we see the expansion of AI, this is not all going to happen through people hiring, 50 PhD level data scientists from MIT and a quarter million a piece, it's just not going to happen.
So we're going to have to provide the tools where mere mortals can apply data science to their business to achieve the benefits, and that's what Ex Machina is all about. And so think about it, I think, Alteryx is a fine company. I'm sure it'll proceed, it'll continue to be successful. But it has some -- but it serves that same market need. Now Alteryx has some constraints. It'll run on any computer you want as long as its on premise on a Windows machine. And Ex Machina will run on any computer that supports a browser. So it's entirely cloud native. Alteryx will use any amount of data as you want as long as it doesn't exceed two gigabytes. Well, we're doing data science at organizations like Aneel, Shell, UnitedHealthcare, Department of Defense, we're dealing with hundreds of petabytes, not a couple of gigabytes. And ALteryx will allow you to perform data science on basically as many rows as you want as long as it doesn't exceed 10,000.
And so think about a 21st century interface on a cloud native WYSIWYG application where the nuclear reactor that’s under the surface that nobody knows is the entire C3.ai platform. And so that is a -- it's a mass market product, it's available on the market right now, you can sign up on the web today and get a free trial, I encourage you to do so. And after you do your 30 day free trial, I encourage you to buy it. I think it cost you $300 a month or something. And so we'll be looking at developing an area where we'll be able to get thousands of customers. It's now a tried and tested proven production product and everybody on the call, I encourage you to go find, C3.ai, click on Ex Machina, sign up for a free trial and send me an email and some of your thoughts of the product before or after you buy it.
Your next question comes from the line of Michael Turits with Keybanc.
So I was really pleased to see the Fortune 500 deal. In the Microsoft partnership, I was wondering if you can fill down a little bit on that and tell us what the value creation was there and why that seems promising going forward? And I've got a follow up.
I mean, we have a number of significant partners and we go to market with AWS, we go to market with Google, we go to market with a lot of these guys. But I would say the organization that is kind of that, as the DNA most aligned with ours is Microsoft, or these guys are really strong and enterprise selling. And so we've been able to align with them. Some of the largest organizations in the world, and the United States government with Shell and with others, I mean we're on speed dial with these guys, they are a great partner. And it looks to me like as it relates to cloud computing, I mean, these guys are going to be a force to be reckoned with, because they really do understand the enterprise and they know how to sell there. And I think they come to believe that we're not entirely unfamiliar with that process also. And so we get along with those guys really well. And we're selling with them in oil and gas, we’re selling them in utilities, we're selling with them in banking. I guess some place around here there's a whole pipeline of transactions that we're doing with Microsoft enterprise transactions, but I'm confident it’s hundreds of transactions that we're working together [Multiple Speakers] they don't hesitate to spool off the big guns, so if they need to close a deal or make a commit to a customer, they'll roll out Satya or Judson or JP or whoever they need to make the commitment and make sure the commitment gets met, they're really strong.
And Tom, if I could just get you to be macro prognosticated here for a minute. You did 20% sequential, almost 9% sequential growth in the quarter. But that's the most you've done in seven quarters. And is this an indication that from -- I mean, obviously, you're doing a lot of things right. But from a demand or macro perspective that people are opening up to these kinds of larger projects that you do?
Well, I think I think our subscription revenue was like 24% growth, wasn't it? And so it’s really pretty healthy. And that's what we're focused on and we're really not focused on the services line. And you want to keep our services line under control and we give you guys the courtesy of disaggregating those two, so there's no question as to whether we're a software company or a services company. Okay. Now, let's talk about -- I mean, coming in to January of 2020 and we were blowing and glowing. I think, our growth rate last fiscal year was 68%, I think top line, okay, and -- 71%, okay, thank you. 71%. I mean, it was we're blowing and going. When COVID came in February, I mean our world just stopped. And understand back then our average transaction size as you know was like, what did I say $21 million. Okay. And so London closed, Paris closed, New York closed, Chicago closed. God knows San Francisco closed, okay, and even the beach in Lower County were closed, I mean, there was nobody else for business.
So we hit a speed bump, okay, in the first two quarters of, I would say, in the February, March, April, May timeframe, we hit a speed bump. Okay, then the all of a sudden, you start getting, all the stimulus starts to take place, people start focusing on digital transformation, they start focusing on AI. And we began to see a significant acceleration in our business once you get in June, July, August, September, which mean it’s feel comfortable talking to you guys about the idea of making this a public company. I mean, there's no question that we're seeing an acceleration. And we believe that as we get into fiscal year '22 and '23, we will see accelerate -- we will see increased growth rates. And so this is a healthy market. And I believe we know how to grow a business rapidly and we're pretty optimistic. But I think the overall growth rate in this year, this fiscal year, David, which will be what?
In April, I think, yes, we would be in the mid point of the guidance [16%].
Mid point of the guidance is like [16%], okay, fine. But we’re dragging Q1 and Q2 like a boat anchor of this fiscal year, it may hurt us. But COVID has turned from a massive decelerant to an accelerant salary. And we're -- I’d like to say COVID was behind us, hopefully, it will be behind this soon, so we all get together for a cocktail in New York. Yes. We're optimistic about the next two years.
Your next question comes from the line of Dan Ives with Wedbush. Your line is open.
So Tom, could you maybe talk about your conversations have changed with customers over the last six to nine months, when I compare it to a year ago. I mean, maybe there it really feel like it almost went from more of a push to a pull. I mean, can you just maybe talk about some of the drivers and some of the anecdotal conversations with the CEOs and CIOs? Thanks.
I did not anticipate the effects of the IPO on the brand. And I guess I should have, because all the papers tell you it’s going to it, and this is going to be important for your business. I just look at it as financing event, to finance growth. And so, I guess I'm kind of slipping, because the effect of the IPO and the associated PR really has been huge on the perceived credibility and gravitas in the business. And so we're seeing a much substantial increase in the growth of the pipeline, conversations are happening faster. The default position of the companies are coming into discussion is that it works and how are we going to make it work. There's virtually -- you know, if we want to hire a senior executive today, there's virtually nobody who won't return the call. And so the effect of the company has been extraordinarily positive. We do -- when you go play on Google Analytics, sometime, and you go do a search on like enterprise AI or C3.ai and look at what has happened in terms of kind of where we are in search frequency in these terms, I mean, we've moved way, way up in the list. So it's -- as I say since the IPO things that -- I think the IPO contributed, I think the global economy is contributing. I think the fact that the government is trading $100 billion a day, this government is in hurry and the focus on digital transformation and AI. So I think we're in the right place at the right time. We have a good solution and I think things have changed significantly.
Your next question comes from the line of Mark Murphy with JP Morgan.
Tom, we have seen quite a bit more excitement about a commodity super cycle and in particular in oil because of the pandemic recovery and response. I'm wondering how different is your energy deal pipeline today versus a year ago. I don't think you've commented on PETRONAS in the script. Wondering if you can comment on that. And just how different is that pipeline, and does a typical oil and gas company feel more pressure to modernize during tough times, or maybe more inclination to spend into a super cycle?
Well, Mark, I don't pretend to be an expert in energy market, as you know, but it's been fascinating to watch. And first of all, the relationship with Baker Hughes is extraordinarily positive. And it's kind of funny that it would be because you couldn't find two companies more culturally different. This is old school oil and gas company, and there’s a new school high tech company but these guys are pretty switched on. And yet today -- I mean, we must talk between the two companies, so 100 times a day, and we finish each other’s sentences. When we first -- you get into February, March, April, oil was negative $37 a barrel. You think that might have a little chilling effect on the energy business, holy moly, I mean these guys were just reeling, right? And they couldn't figure out how to lay off people fast enough. And then after they kind of picked themselves and dusted themselves off, I think they started looking at the reality situation, we had companies like Shell, which is, as I recall, roughly a EUR300 billion business, I think the fifth largest company in the world. And they look at the economics of -- you look at oil futures at the time, nobody was predicting oil over $50 a barrel, I believe it might be over $50 a barrel today, but it wasn't being predicted to be over.
And there's a lot of downward pressure on oil prices for all the reasons that we both know. And I think if you look at a barrel of oil, there might be one company in the world that can make money, okay, at $50 or $45 a barrel, and that would be Saudi Aramco because they have no lifting cost. I think that stuff comes out of the ground at about $6 a barrel. So then you have the rest of these guys, whether it's Rosneft, Gazprom, Exxon or Shell, who are running these large global enterprises and they have the choice to either reinvent themselves or slowly go out of business. And so I mean you get to places like Shell, these guys are not -- I mean, these are highly educated and really bright people. They've made the decision to reinvent themselves. So I think the oil pressure and oil fluctuations are proving an accelerant to the use of AI as these companies -- you really focus on renewable. I mean Shell is turning itself into a renewable energy company. And with the bulk of -- I think their revenues are they going to be coming from electricity in the future. It's a big change. But we see this Rosneft, Gazprom, Aramco, ADNOC, they're -- all the big guys who are thinking about this and we're at the table.
And David, as a quick follow-up, could you clarify how did the GAAP RPO or backlog behave sequentially in the January quarter? I'm just not sure if I heard that number correctly. And I know there's a few different layers on that cake for you. So just how did that GAAP RPO behave sequentially? And then I'm just wondering, do you have the current RPO or the next 12 month’s portion at your fingertips by any chance, or perhaps do we see that later?
Well, you'll see the current RPO in the Q, but the current RPO is $131 million.
And then what was the sequential change just in the total GAAP RPO?
So when you actually look at our total GAAP RPO, you're looking at it from last quarter, it was 267, and it's 247.5. So 267.4 to be precise, 247.5 is the GAAP RPO.
One of the changes that's happened here, Mark, is that back in the old days, like, say, pre-December, we were ruthless about the way that we engaged -- we did software transactions, okay? If the commitment from the customer was not irrevocable, nonrefundable commitment to buy, okay, over multiple years, we did not give business hard stop, okay? Now that we have a -- now we're kind of holding our position where we're focused -- and what we're doing, and we were doing our best to finance the business, and we did a pretty good job, okay, without ringing a lot of doorbells on Sand Hill Road. Another business is pretty well capitalized. We're focused on market share, okay? And we've gone towards more, I would say, market transactions, where we are willing to consider engaging in multiyear transactions with organizations, where some percentage of the transaction is cancelable based upon performance. Well, you could be sure that we're going to bust our assets to make sure that nothing gets canceled. But it does that -- but that decision did have a negative effect on RPO, okay? And I think it's the right decision to make for the business. But we've become -- we're doing -- engaging and we're being more flexible and easier to do business with in order to get contracts signed. And it does have some downward pressure on RPO, but I think it will have upward pressure on revenue prospectively.
And Mark, with that in mind, the cancelable piece went from 37.1 to 48.4, so sequentially up 30%.
So if I'm hearing you correctly, there's a little more inclusion of cancellation clauses in the contracts, so that just the RPO is kind of the mix is shifting a little off the long term and a little off the GAAP piece because it's rolling a little more into the cancelable piece then but it's netting out positively for revenue. Is that a fair estimation?
And it's also positive -- RPO plus cancelable, it's increased year-over-year…
Year-over-year, it is up. Sequentially, it was off a little bit.
Your next question comes from the line of Patrick Colville with Deutsche Bank.
So I guess I've got a two part question. I mean, I think we just touched on it just now, but I just want to just clarify. So the delta between RPO and billings, just help me understand that. Because if I look at billings in the quarter, I think it was off about 5%, whereas the adjusted RPO was up 16%. So just help me understand why that might be?
Well, I think -- and I tried to include it in my prepared remarks on billings was, I think you're looking at perhpas in deferred revenue. Is that correct?
Yes. And so I think what we tried to highlight in our remarks is that in almost in line with Tom's comment on being more flexible, there were some pieces of our deals that are not flowing through deferred. And I called out a portion of what happened at the end of the quarter where $4.1 million isn't flowing through deferred. So I think that that is, to an extent, that's what's impacting your math there.
And are you prepared to call out the Baker Hughes contribution in the quarter?
The Baker Hughes contribution as part of RPO?
As part of revenue.
The Baker Hughes portion, I think we will actually be including in our 10-Q. And so that's where that revenue will be published.
Your next question comes from the line of DJ Hynes with Canaccord.
Tom, when you think about building domain expertise as you enter new vertical, how much of that falls on C3 development efforts and how much of that is influenced by the Lighthouse account with which you're partnering? I'm just trying to think about how you replicate the model as you continue to grow into new markets.
And one of the real gems of the story that really is not realized yet by a lot of the market is that across all of these markets, whether it's AI-based predictive means for the Air Force, whether it's customer churn at Bank of America, whether it's process optimization, AI based predictive maintenance for paper manufacturing at Georgia-Pacific, or whether it's hydrocarbon loss accounting at Shell, 99% of the code we're installing is the same across the entire installed base because that's the beauty of what we've done. So you don't need to know -- we don't need to know first principles of how a turbine works -- a gas turbine works in order to build a digital twin for the turbine, or build predictive maintenance model for that turbine. 99% of the code is exactly the same. All the changes are the data sources, the machine learning models and the user interface. So we agree that the user interface is trivial. We have true realized this idea of aggregating very large data sets into unified federated image.
For example, this is what Palantir calls an ontology. This is what Palantir does, okay? Palantir is a large services organization that takes -- it's a big business, right? Taking, I think, at ENGIE and now we've aggregated 100 trillion rows of data from 50 enterprise data sources, 27 million sensors. And then with regard to the extra price for weather and train social media, that will -- we will update weather, train and social media, 62 billion times a day. And when you aggregate that into the federated image, this is what Palantir says that every sentence of their presentation is ontology, ontology, ontology. Look it up. It's not quite that magic, okay? So we rather do that through services, we do that through software. So we built with our partners at Baker Hughes -- actually before Baker Hughes, we built a predictive maintenance application for an offshore all rig. At Shell, we don't know anything about offshore oil rig/ We've built production optimization for their LNG operations in Australia, that's Queensland Gas. We don't know anything about LNG operations. But we're able to work with their subject matter experts, say our team of six people and their team of six people and build the applications. Today, Shell, God knows how many people they are working on projects, is between 100 and to 1,200s, because they have 100s projects in flight.
And when you did Smartband to create analytics for ML. And ML, I would say, that's the largest AI application in production in nonclassified space on Earth, I mean we don't know anything about how the grid operates, but we know about using AI, to do what they want to do, which is both VAR, predictive maintenance for devices and customer churn and what have you. As we're doing AI based predictive maintenance for the F-35 Joint Strike Fighter, do you think we have any idea how a Joint Strike Fighter operates. I assure you, we do not. Only the people at Lockheed Martin know that. And so where we're able to build the tools, where their subject matters experts can do what they do it, that is the beauty of what we've done. We're able to apply the same code for anti money laundering in a bank, and predictive maintenance for F-35 Joint Strike Fighter. And so I don't think we will be constrained by domain expertise. Now that being said, again, to the extent that we need it, where we’re getting it. We're getting it through partnerships like FIS, like Infor, like Baker Hughes and ENGIE, as it relates to energy efficiency. So I kind of went all around that. But you do not need that they had to clear side as to build a predictive maintenance application for a nuclear reactor and that's the beauty of what we build.
Your next question comes from the line of Peggy Yu with Morgan Stanley.
First, to touch on the pipeline, you've mentioned seeing greater customer interest and the pipeline has continued to grow pretty substantially last quarter. Wanted to see if you could give us a little bit of color on pipeline conversion trends, especially as we go through the year?
I'm sorry, pipeline conversion trends? Peggy, I don't have that answer. And it's a legitimate question. Next time we will talk, I'll have the answer. I don't have it. And so anything I would give you would be misleading? It's a good question. And I'm sorry, it's very, very rarely, I get caught absolutely out of my heels and you caught call me.
Yes, you win the price for the day.
So I guess, could you talk about the trends around margins as we go head into Q4? Looks like the margins -- so Q3, obviously, you saw a pretty good improvement, and Q4 seems to be slightly down. What are some of the puts and takes there?
Q4 margin is down? How was that 100 what…
I think the non-GAAP operating loss that we highlighted in the guidance table.
I think one of the things that is going to contribute to a little bit margin compression in Q4 is, I mean we are focused on being a software company. We're not focused on being a software services company, as lucrative as it could be in this particular market. I know. Okay? I look at Yahoo! Finance every now and then too, okay? That being said, we're going to supplement in order to facilitate our customers and not be in the professional services business, we're going to outsource professional services work with some third party providers that we've enabled. So we have this large ecosystem of third party providers, like IBM’s Global Services, and I think there's about 20 others. And so we will actually give a piece of our contract to them, which will put some downward pressure on, I'm sorry, upward pressures on cost of goods sold and a little bit downward pressure on margins, at the same time, you know keep me out of the services business. And so that's what -- it's not a macro trend, but that's what that's all about.
Your next question comes from the line of Arvind Ramnani with Piper Sandler.
I have a question on your vertical partnerships. As you indicated, it's a very important part of your strategy. And I wanted to understand how you're tracking progress with these partnerships, particularly with the newer ones versus the ones that are more ingrained? Are you looking at deals or pipeline? And are there certain trigger points that will get you to switch from NFIs to Fi service as the partnership isn't quite working out. So just kind of conceptually, how you're managing success at these important partnerships?
We have very, very impressive instrumentation here on kind of what's going on in the business, what's going on with business activity with each partner, with each of the vertical, what's the rate at which we're growing pipeline, what deals are working together, what's the next step of each deal, what's the expected revenue, what are the revenue goals for this year, next year, the year after, by quarter. So we have very tight metrics on that. Fortunately, we haven't gotten to the point where one is not working yet and we need to replace it. We're still focused on making every one of them successful. And we believe that with the partners we have, that we can. That said, are we going to get to the point where one doesn't work? We will. But sometime, Arvind, when you're here, I mean we could kind of show you the metrics that we have for tracking this and sales operations. I believe that we have a finger on the pulse of this business as it relates to the AI enabled CRM system that we have deployed that is absolutely state of the art.
Your next question comes from the line of Pat Walravens with JMP.
And if I can, I'd like to ask two. The first is, for investors, as they look at this, C3 grew 71% in fiscal '20. You hit COVID. And then in fiscal '21, revenue at the midpoint is going to grow 16%. So how should we think about sort of the durable long term growth rate for this business?
We believe the growth rate will expand.
So Tom, can you talk a little bit about the deal with the US Army that you won with Raytheon? I mean, those things are typically super competitive, right? Who is the competition? And then just more broadly, how big or talk about the opportunity that C3 has with the defense industry?
Now this is our -- I know that everyone on this call believes that I close every deal that we do. And I can assure you, that's not true. As a matter of fact, from August 13, 2020, until like December, about 21st. I did nothing but talked to you guys for about 21 hours a day. And I never talk to a customer or prospect. So as much as I would like to talk about that deal with Raytheon on that particular transaction, I don't know anything about it. And we can somehow hold the session to talk about it. How big is the opportunity in defense and Intel, it's freaking huge. And as it relates to applying AI to military and defense, this is going to be a major initiative for the United States government, and you have China spending tens of billions of dollars today, year on AI. And we're -- in the United States, government I think been historically under investing there’s a lot of talk about it and very little action. I think now we're seeing, I think, going forward, as it relates to defense and intel, that market opportunity is going to be virtually limitless.
And I think we're in a position to have -- make an impact there, which is why you see Danny McGinn, the former Assistant Secretary of the Navy, and Rick Ledgett, former Deputy Director of the NSA, joining our advisory boards to help us figure out how to navigate that beyond the phone on Friday with one of the immediate past secretaries of one of the three branches of the US military, to get his advice and how we should structure this going forward. But you can expect us to be making a very, very substantial investment there. I mean, will this ever be anything like 50% of our business? No way, no how, okay? I mean, we're not going to become a federal contractor. But do we expect that they're large and rapidly growing segment of our business? Yes, we do.
There are no further questions at this time…
I think we might be at the last -- we might be about to wrap this up. Is that correct?
Yes. There are no further questions at this time.
Okay. Let me just provide a couple of closing comments, and then we'll put a wrap on this. So first of all, thanks, everybody, for your attention and your thoughtful questions. And to the extent that I think there were two questions that I didn't have the answer to. One was from Pat and one was from Morgan Stanley. And I apologize. So we'll figure out a way to -- I guess I can't get back to you, we need to figure out a way. Can we follow-up? In the world of Reg FD, we can do this [Multiple Speakers] Okay. So we'll do this. I think that we're really pleased with third quarter results. We think they illustrate the power and potential of this highly differentiated model driven architecture that is enabling enterprises across a wide range of industries to rapidly and efficiently develop and operate complex predictive AI applications to scale.
We think we are really well positioned to address this rapidly growing greater than $200 billion addressable market opportunity. And as we enter our fourth fiscal quarter of 2021, I believe that C3.ai has never been more strongly positioned in the market. I believe that we are demonstrating clear technology leadership in enterprise AI. We are building a powerful brand. Our human capital resources are second to none. Our customer and market partners' successes speak for themselves, and the competitive landscape does not appear to be limiting. So thank you so much for your time. We look forward to sharing our progress with you in the months and years ahead. And we wish you all a great day and a great week. So thank you for the courtesy and your time today.
Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.