Transcript
Good day everyone. And welcome to Innodata's Second Quarter 2023 Earnings Call. At this time, all participants are in a listen-only mode. A question-and-answer session will follow the formal presentation. It is now my pleasure to turn the floor over to your host, Amy Agress. Ma’am, the floor is yours.
Thank you, Mathew. Good afternoon, everyone. Thank you for joining us today. Our speakers today are Jack Abuhoff, CEO of Innodata; and Marissa Espineli, Interim CFO. We'll hear from Jack first, who will provide perspective about the business, and then Marissa will follow with a review of our results for the first quarter. We'll then take your questions. First, let me qualify the forward-looking statements that are made during the call. These statements are being made pursuant to the Safe Harbor provisions of Section 21E of the Securities Exchange Act of 1934 as amended, and Section 27A of the Securities Act of 1933 as amended. Forward-looking statements include, without limitation, any statements that may predict, forecast, indicate or imply future results, performance or achievements. These statements are based on management's current expectations, assumptions and estimates and are subject to a number of risks and uncertainties including, without limitation, the expected or potential effects of the novel coronavirus pandemic and the responses of government, the general global population, our customers, and the company; impacts from the rapidly evolving conflict between Russia and Ukraine; investments in large language models; contracts that may be terminated by customers; projected or committed volumes of work that may not materialize; pipeline opportunities and customer discussions that may not materialize into work; acceptance of our new capabilities; continued reliance on project-based work in our Digital Data Solutions segment; and the primarily at-will nature of such contracts, and the ability of these customers to reduce, delay, or cancel projects. The likelihood of continued development of the market, particularly new and emerging markets that our services and solutions support; continued revenue concentration in our Digital Data Solutions segment with a limited number of customers; potential inability to replace completed, canceled, or reduced projects; dependency on content providers in our Agility segment; downturns in market conditions; changes in external market factors; the ability and willingness of our customers and prospective customers to execute business plans that give rise to requirements for our services and solutions; difficulty in integrating and driving synergies from acquisitions, joint ventures, and strategic investments; potential undiscovered liabilities of acquired companies; potential impairment of goodwill and other intangible assets; changes in our business or growth strategy; the emergence of new or growth of existing competitors; our reliance on information technology systems including potential security breaches, cyber attacks, privacy breaches, or data breaches that result in unauthorized disclosure of information; and various other competitive and technological factors and uncertainties indicated in our filings with the Securities and Exchange Commission, including our most recent reports on Forms 10-K, 10-Q, and 8-K and any amendments thereto. We undertake no obligation to update forward-looking information or to announce revisions to any forward-looking statements except as required by federal securities laws, and actual results could differ materially from our current expectations. Thank you. I will now turn the call over to Jack.
Amy, thank you. Good afternoon, and thank you for joining our call. I am tremendously excited to announce that we have landed yet another of the big five global tech companies as a new customer to help it train and scale generative AI and large language models. These, of course, are the technologies behind ChatGPT and stable diffusion, technologies that have caught the world by storm and set off what we believe will be the next revolution in computer technology. We expect our formal agreement with this new customer to be signed tomorrow. With today's deal announcement, we are proud to say that we have won each and every one of the potentially transformative pipeline deals we spoke about last quarter, and as a result of these wins, we are poised to support AI and large language model development for four of the five largest tech companies in the world. These wins, including the deal we are announcing today, have all come in just the last eight weeks, so they are not yet reflected in our financial results. We believe that these wins are potentially transformative for our company. Moreover, we believe we are in the very early stages of exploiting a market opportunity, which itself is in its very early stages and for which, as our last eight weeks of wins demonstrate, we are particularly well suited. In today's call, I will provide some updates and additional perspective around today's announcement and the other deal announcements we made in the last eight weeks. I will then discuss the opportunity landscape we see forming around generative AI technology and how we intend to exploit it. Lastly, I will review our Q2 results and provide some forward guidance. In the last eight weeks, we made four deal announcements, including the deal we announced today. Two of the deals are with an existing customer and two are with two new customers, all of whom are among the top five global technology companies. For the new customer we're announcing today, we expect to provide an array of services required to build and scale large language models. We expect to begin ramping up the engagement early in the fourth quarter. Importantly, the agreement we expect to sign tomorrow with this new customer is a framework agreement that enables the customer's business units to easily add additional programs and allocate additional scope and spend to us. The customer has told us that it may potentially request us to participate in additional programs, which may include supporting the customer’s own customer base with model fine-tuning and integration. Under the agreement, we expect the customer to authorize $3.5 million in spend to get us started, and they have stated that they intend to supplement this authorization as we move forward. The customer has shared with us its vision for the initial program that, if fully realized, we believe could potentially result in approximately $12 million of new quarterly revenues at maturity. That said, at this point, we do not know when or if the initial program will reach this level of spend. While the customer has told us we will be signing tomorrow, it is possible that this could slip as it is outside our control. On July 18, we announced winning another new customer, another of the top five global tech elites. Similarly, we will be helping this new customer develop and train large language models. The customer shared with us that it had been extremely impressed with our pilots and selected us from a field of 17 competitors. The agreement we signed with this second new customer provides for up to $8 million in spend for the remainder of 2023 under an initial program. Although how much of this we recognize is dependent upon multiple factors, including how quickly we ramp up. Ramp up aside, based on discussions with this customer, we believe we could potentially get to an annualized run rate of $15 million with this customer by the end of the year solely on this initial program. The agreement we signed also allows its business units to readily assign new programs and scope to us. We have already begun discussions regarding these.</br> On June 27, we announced that an existing Big Five customer selected us to perform AI data annotation and the LLM fine-tuning as a white-labeled service for its cloud and platform customers. We view this as a particularly strategic win as it provides us the opportunity to scale LLM services for enterprises in a way that would have likely been impossible working solely under our own brand. Driving distribution under the customer's brand is an attractive opportunity. The engagement builds on our 18-month relationship with this company, where we are told we distinguished ourselves based on data quality, responsiveness, and agility. While we stated in our June 27 press release that we expected to kick off the program with three end customers in July, I'm pleased to report that we already have nine end customers signed on for pilots, and we're not even halfway through August. We believe three of these pilots may likely turn into booked business near term, with one having an anticipated booking value of over $1 million. Beyond these nine white-labeled enterprise pilots, we have a direct pipeline of enterprises we are or soon will be engaged in discussions about LLM fine-tuning and integration. Our direct pipeline includes a large legal information company, one of the largest life insurance companies in the world, a leading investment bank, and a leading commercial bank. In the near term, through this program with our customer, we anticipate a cadence of one to two new pilots each week. While it's difficult to forecast the revenue opportunity this program represents because it is so new, we believe that if we can continue to onboard from the potential pool of tens of thousands of our end customers at the pace we're executing currently, this opportunity could dwarf any single initiative we're now pursuing. On June 14, we've stated that an existing Big Five customer engaged with us for its LLM build program. In that announcement, we anticipated potentially exceeding $8 million in revenue this year with this customer. This is up from approximately $3 million last year. We believe there could be considerable opportunities to expand existing programs with this customer and to land new programs. The June 27 announcement I just spoke about regarding the white-labeled service is an example of such a new program. I would like to now discuss our opportunity landscape and put these wins into a strategic framework. First, we believe a technology revolution is before us, that it is in its very earliest of days and will drive profound changes in how people work and engage with technology. Reflecting back, we initially became convinced that we could gain a foothold in the emerging generative AI market by leveraging our experience creating high-quality domain-specific data. Over our seven years of experience in training AI models, we selected the Big Five as our initial target market because we predicted they would be early adopters spending aggressively. They all have products, including search, ad serving, and productivity tools that we believe stand to directly benefit from generative AI, and several are hyperscalers that would seek to provide high-performance infrastructure for training and serving generative AI models. Fortunately, based on what we've seen to date, we believe our assessment was accurate. In their Q2 analyst calls, several of the Big Five tech companies referred to generative AI and LLMs as core to their innovation plans, backed by commitments for billions of dollars per year in capital expenditures. At least one prominent firm stated that every one of its businesses had multiple generative AI initiatives. It is clear that these companies are now in a race to build cutting-edge LLMs that can serve as a foundational layer of generative AI for their products as well as their ecosystem customers. It is early days. One company described it as being just a few steps into a marathon. We expect to perform a range of data engineering work required to build cutting-edge generative AI for the Big Five. This may include collecting data, creating datasets to train the models, teaching the models to follow instructions and chain of thought reasoning, providing reinforcement learning to align models with human values and complex use cases, and providing red teaming and model performance evaluation. Over time, we anticipate working with these companies to address multiple languages, multiple data domains, and multiple data modalities including video. Given that we have now successfully penetrated four of the five largest technology companies, we believe we are poised to drive significant growth in the market in the near, medium, and long term. Beyond the Big Five, there are another 50 to 100 technology companies that we estimate are now building or likely to be building their own LLMs, including some prominent well-funded startups. We intend to build relationships with these companies as well. Another compelling opportunity with the Big Five is establishing white-label programs under which they can provide our services to their customers under their brand. We have a deal in place with one of the largest hyperscalers in the world, and we're exploring similar opportunities with others. We view the white-label arrangement as a particularly strategic opportunity for three reasons. First, given that this customer is one of the largest global cloud providers, we expect many thousands of customers and partners to utilize their infrastructure for generative AI. These customers will likely find a one-stop shop for all things AI to be an attractive value proposition, obtaining compute, storage, foundation models, machine learning, database, and data services, basically everything required to train and serve generative AI under one roof and at an attractive price point. Second, growth in mid-market companies like ours is typically constrained by sales and marketing effort size and scale, but these white-label programs offer us a means of scaling independently of our own sales and marketing, leveraging their brands and customer reach. Lastly, through these programs, we believe we’ll gain early exposure to a wide range of early adopter generative AI use cases. Under our initial white-label program, we've recently piloted use cases ranging from call center summarization, legal and medical question answering, and e-commerce. We believe this exposure will set us up well for what we believe will be our largest and most significant opportunity in LLMs for enterprise. We believe that a decade from now, virtually all successful businesses will be AI companies, and that between now and then, the pace of adoption will dramatically accelerate. There are essentially two reasons for this. First, early adopters of LLMs are likely to be significantly more productive and create more compelling customer experiences. Thus, there will be strong competitive pressure accelerating the adoption curve. A June study by McKinsey estimated that generative AI could add trillions of dollars in value to the global economy, with half of all current work activities automated between 2030 and 2060, with a midpoint of 2045. Secondly, the tech stack will be ready. High-performing commercially usable open-source generative AI models will become increasingly available, and the best performing closed-source models will likely support fine-tuning on proprietary data. We anticipate that this will enable companies of all sizes to customize their own large language models and build generative AI applications in an enterprise-grade fashion. As the enterprise market accelerates, we believe our capabilities will be increasingly in demand. Our plan is to exploit the enterprise opportunity with what we believe are five distinct competitive advantages. First, the skills and referenceability acquired from helping the Big Five build foundational models. Second, the experience gained across a wide range of use cases, working with early adopter enterprises through our white-label arrangements with hyperscalers. Third, the real-world experience we continue to gain in integrating both classical and generative AI into our operations and products. Fourth, our technology platforms for transforming enterprise data into LLM-ready context for model fine-tuning and prompt injection. Fifth, our technology platforms, both existing and in development, that will help enterprises generate reliable, fact-based responses and insights from foundation models. In addition, our strategy is to harness AI and LLM technology within specific workflow applications. For example, in late 2021, we announced a multi-year agreement with one of the world’s largest banks to use AI to re-engineer workflows related to regulatory updates. This project is underway and is being enthusiastically received by the bank, promising to improve regulatory change management, resulting in reduced fines and penalties, while also alleviating the heavy lifting done by personnel. The bank pays us an annual subscription fee, and we’re currently in discussions with three other companies about this product. Let’s now talk about what to expect financially as a result of our new wins. We expect to begin work on the win announced today in early Q4. Concerning the win announced on July 21, we kicked off work about two weeks ago. Based on our experience, engagements start slowly. First, we work with the customer to create detailed specifications and run pilots to ensure that the specifications yield the intended results, often requiring several iterations. Once the specification is locked down, we implement the required infrastructure—which includes custom-configuring technology systems, finalizing process designs, and assembling human resources, data engineers, and subject matter experts. This phase can take two to three months. Following this, we typically scale up slowly to continue testing and refining as necessary with the customer. We expect our programs to be dynamic with customer dependencies and checkpoints throughout, making quarterly forecasting challenging. However, based on our experience, full ramp-up is typically achieved in approximately 12 months. Finally, let’s discuss our Q2 results and guidance. In Q2, our revenue was $19.7 million, reflecting a 4% increase from Q1, and adjusted EBITDA was $1.6 million, a 100% increase from Q1, made possible by the work we did late last year and into Q1 in sharpening focus and finding opportunities to operate more cost-effectively. There was no revenue in the quarter from the wins we discussed today, nor from the large social media company which contributed $2.5 million in Q2 of last year but dramatically reduced spending in the second half of the year due to a significant management change. If we exclude revenue from this large social media company, our revenue growth over Q2 2022 would have been 13%. We expect revenue and adjusted EBITDA growth to accelerate moving forward as the wins discussed today start ramping up. We ended the quarter with $13.7 million in cash and short-term investments, up from $10.8 million at the end of Q1. We continue to have no appreciable debt. To support our growth and working capital requirements in the quarter, we established a secured revolving line of credit with Wells Fargo that provides up to $10 million of borrowing subject to a borrowing base limitation. I’ll now turn the call over to Marissa to go over the numbers a bit more, and then we’ll open the line for questions.
Thank you, Jack. Good afternoon, everyone. Allow me to recap the Q2 2023 financial results. Revenue for the quarter ended June 30, 2023, was $19.7 million compared to revenue of $20 million in the same period last year. The comparative period included $2.5 million in revenue from a large social media company that underwent a significant management change in the second half of last year, resulting in drastically reduced spending. There was no revenue from this company in the quarter ending June 30, 2023. The net loss for the quarter ended June 30, 2023, was $0.8 million or $0.03 per basic and diluted share compared to a net loss of $3.8 million or $0.14 per basic and diluted share in the same period last year. Revenue for the six months ended June 30, 2023, was $38.5 million compared to $41.2 million in the same period last year, again with the comparative period including $6.9 million in revenue from the large social media company mentioned earlier. The net loss for the six months ended June 30, 2023, was $2.9 million or $0.11 per basic and diluted share, improved from a net loss of $6.6 million or $0.24 per basic and diluted share in the same period last year. Adjusted EBITDA for the second quarter of 2023 was $1.6 million compared to an adjusted EBITDA loss of $1.3 million in the same period last year. For the six months ended June 30, 2023, adjusted EBITDA was $2.4 million compared to an adjusted EBITDA loss of $2.3 million in the same period last year. Our cash, cash equivalents, and short-term investments were $13.7 million at June 30, 2023, compared to $10.3 million at December 31, 2022. Again, thanks everyone. I will turn you over to Matthew. We are now ready to take questions.
Your first question is coming from Tim Clarkson from Van Clemens. Your line is live.
Hey, Jack. Good quarter, everything I hope for. So just a basic question. It seems to me looking at the results that your breakeven would be somewhere around, I don’t know, $21 million to $22 million. Is that kind of a ballpark number?
Yes, Tim, I think that’s right. As I mentioned in my remarks, we did a lot of work late last year into Q1 to focus ourselves, achieve a higher level of operational excellence, and that’s benefiting us now. So, I think that’s accurate.
Okay. Now, historically, execution has always been your strength. I mean, when there have been problems, it’s been contracts ending being project-based, but in terms of execution, I can’t remember a screw-up on that end. How hard are these new contracts to execute? Are they within the competencies of the company, or how risky are they?
Sure. These contracts are squarely within our competency, and that’s why we’re so excited. We’re finding ourselves right in the middle of the biggest technology revolution of our lifetimes, armed with competencies that are highly applicable to driving that forward. We’re working with companies that are going to spend billions in CapEx to do exactly that. Now that said, while the contracts are challenging, that’s what makes it good business for us. We have a track record of executing some of the hardest data products and projects over the last couple of decades. We’re several weeks into executing some of these new deals and already demonstrating excellence. The customers are very happy, and we expect things to go very well.
Right. If someone were looking at Innodata from the outside, how would you articulate the competitive advantages you have?
We have several competitive advantages. The work we’ve done in the past, creating high-quality data and building and integrating AI models in real-world applications, has served as our training for today. We’re now prepared for the big game. Competitive advantages include our breadth of domain expertise, competencies in AI and data training, a geographic spread, and the ability to create high-quality training data. Looking forward, we’re building additional competencies, now helping build foundational models, and we’re developing experience through the white-label programs we’ve discussed. We’re also continuing to gain real-world experience integrating classical and generative AI into operations and products. All of these contribute to creating significant competitive advantages for us.
Sure. Let’s imagine these contracts begin to take shape and you’re hitting around $30 million in revenue. At that level, what kind of pre-tax margin would you guys be showing?
I would guide people to think about an adjusted gross margin—meaning, gross margin less severance, G&A, and stock option costs—scaling between 40% and 45%. In terms of SG&A, we have a significant market opportunity ahead, so we’ll likely be adding a bit more in sales and program execution, but not drastically. Our business is not highly capital intensive, so you can predict the cash flow we’ll be able to drive.
Should be at least 10% to 15% pre-tax once you get above $30 million; that’s what I’m thinking.
Yes, I think that’s absolutely right, and there's potential to improve on that.
Everything is looking great. I’m just thrilled, and it’s still quite early. That’s the exciting part. I’ll let other people ask questions. Great quarter. Thanks.
Thank you, Tim.
Thank you. Your next question is coming from Dana Buska from Feltl and Company. Your line is live.
Hi, Jack.
Hey, Dana.
How are you this evening?
We’re doing great. Thank you.
Everything sounds like it’s going pretty good. I have a couple of questions around your white-label program. Could you explain a bit about what exactly you mean by white label?
Sure. We’re going to perform data engineering services while standing behind the brand of one of our large customers. We're currently focused on one customer, which markets compute, storage, foundation models, databases, and will be adding data services. A customer looking to create or build generative AI models will find value in being able to obtain all that in one place.
Are you going to rely on their salesforce to implement that, or does your sales team engage in these types of projects?
Primarily, we will rely on their salesforce, especially in this first iteration. This is attractive for us as it offers an opportunity to scale independent of the limitations our sales and market reach impose upon us. This well-regarded brand has tens of thousands of customers, and I believe programs like this will help us achieve an escape velocity that wouldn't be possible working under our brand alone.
That sounds impressive. Will pricing negotiations be done through your client, or will you have to be involved?
We can provide pricing upfront for the customer, who can then represent that pricing in discussions with their clients.
You didn't discuss Synodex or Agility much. Both seemed to have performed nicely. Could you elaborate on those two businesses?
Absolutely. We had impressive growth year-over-year of 13% in Agility and returned to adjusted EBITDA positive in the quarter. We are now making money in that area with solid bookings, and our gross retention levels exceeded our internal targets, with improvements in average selling price and sales cycle partly due to integrating our new product, PR CoPilot. Regarding Synodex, we experienced issues related to COVID and attrition but are digging out nicely now, with an increased gross margin for the second consecutive quarter.
Looks like everything is going well for you. Regarding DDS, do you anticipate needing to add capacity for these new projects, or do you have enough capacity available?
We have the engineering capabilities and performance management needed. We add human resources as required, but we don’t have issues scaling around that because we can add resources just in time when needed.
Wonderful. That’s all from me. Congratulations on everything going so well. It looks very impressive to me. Thank you.
Thank you, Dana.
Thank you. Your next question is coming from Bruce Galloway from Galloway Capital. Your line is live.
Hey, Jack, congratulations. Great quarter. It looks like you are firing on all cylinders. In the past, you spoke about growth rates and net margins going forward, but I guess you stopped when the world stopped. Will you be providing guidance along those lines? Also, since you are doing so well, will you be getting any Wall Street coverage?
Good to hear from you, Bruce. Thank you for joining the call. We’ve given some guidance regarding expectations for year-over-year and sequential growth in revenue and adjusted EBITDA as we ramp up these programs. However, because of timing uncertainties related to ramp-up and discussions of program expansions, we’re maintaining previous guidance on adjusted EBITDA. Regarding Wall Street coverage, there are a couple of analysts in conversation with us. We’ll see where that goes. Typically, coverage follows investment banking activities. However, we are in discussions with a few people.
Could you elaborate on earnings in IT services that went from 87 million to 82 billion over 11 or 12 years? That’s hypergrowth.
I believe we're a player, and we intend to be a significant one. We're demonstrating that we are. I’ll gladly send you a URL to that Bloomberg report that highlights the magnitude of the opportunity. Yes, the numbers are exciting, and what we've accomplished over just the last eight weeks certainly supports the hypothesis.
It certainly sounds exciting.
It very much is. Thank you, Bruce.
Thank you. Your next question is coming from Tim Clarkson from Van Clemens. Your line is live.
Just one follow-up question, Jack. I am aware you have sold stock before in Innodata, but I notice you haven't sold any stock in the past couple of years. I assume you are vested in the belief that things are going to be really good.
Yes, Tim. We’ve been working at this for a long time. I believe we’ve made some very good decisions and that we are well-positioned for something that is immense and prominent. I think it is extremely exciting. While talking about insider stock sales as predictors of future performance doesn't hold, people have their reasons for raising money or diversifying exposure, and that doesn’t diminish their vested interest in our success.
Right. But you haven’t sold any shares.
I haven't sold any shares. That’s correct.
Thank you. I’m done.
Thank you. That concludes our Q&A session. I'll now hand the conference back to Jack Abuhoff for closing remarks.
Thank you, operator. To quickly recap today, we announced yet another highly anticipated win. We expect to sign tomorrow, completing our eight-week win streak of landing all of the big tech pipeline deals that we presented last quarter. This is potentially transformative for our company and includes two new big tech customers. We believe these companies are positioned to lead generative AI development over the next several years, with billions in capex to support them. Individually and collectively, these new deals present a transformative opportunity for our company. We have a solid track record of land and expand with large tech companies, and coupled with the strong tailwinds of generative AI, we are extremely well-positioned for significant growth in the near, medium, and long term. We are aligned with an exciting, growing, and dynamic market, possessing the right capabilities and a high-quality client base at this point. In a recent Bloomberg report, it mentioned that the generative AI market is poised to explode. It estimated that the overall market for AI-focused IT services will grow from $83 million in 2022 to $21.7 billion by 2027 and to $85.9 billion by 2032, representing a 100% compounded annual growth rate. The report states AI-focused IT services are among the four hottest market opportunities in generative AI, with annual growth rates of 100% or more. The other areas are generative AI-based drug discovery software, cybersecurity spending, and ad spending. Bloomberg also estimates that generative AI will become an essential part of IT spending. I attended a dinner last week where the head of technology innovation for one of the world’s largest investment banks, with a massive $15 billion IT budget, predicted that 20% of his budget will be directed toward large language models within a few years. My team and I are energized by what we've accomplished and the magnitude of opportunities ahead of us. Thank you all for joining us on our call today. We look forward to our next call with you.
Thank you everyone. This concludes today's event. You may disconnect at this time and have a wonderful day. Thank you for your participation.