Palladyne AI Corp. Q3 FY2023 earnings call

• Prioritized end markets to subsea, aviation, solar, and software, narrowing down from 18 product areas to four.
• Suspended hardware commercialization efforts, implemented layoffs, and focused resources on AI platform.
• AI/ML software platform aims to reduce time to program and train robotic systems, expected to be brought to market in first half of 2024 with revenue starting in second half.
• Received a $13.8 million four-year contract from U.S. Air Force to advance AI and ML software.
• Ben Wolff rejoined as Executive Vice Chairman to support AI software platform commercialization and strategic opportunities.

In Q3 2023, revenue was $1.8 million compared to $4.7 million in Q3 2022. The lower revenue was due to completion of certain product development contracts not yet replaced, offset by product revenue from sale of two Guardian Sea Class units. Cost of revenue decreased to $1.2 million in Q3 2023 from $3.6 million in Q3 2022. The company has pivoted to focus on its AI/ML software platform, suspending hardware commercialization efforts for subsea, aviation, and solar robotics.

• Suspended subsea, aviation, and solar robotics hardware commercial efforts, with some small R&D hardware efforts continuing to support software platform.
• Anticipates additional restructuring expenses of $22 million to $24 million in fourth quarter 2023 and first quarter 2024.
• Expect cash to be approximately $39 million at end of fourth quarter 2023, with average monthly cash usage in 2024 at $1.6 million.
• Quarterly R&D expenses expected to decrease by ~80% in first quarter 2024 compared to Q3 2023; G&A expenses expected to trend down, with ~35% decrease in first quarter 2024 vs Q3 2023; sales and marketing expenses expected to decrease by ~60% in first quarter 2024 vs Q3 2023.

• Risks associated with forward-looking statements, including actual results differing from projections due to various uncertainties such as those described in Form 10-Q and earnings press release.

Q: Good afternoon. Hi Laura. Hi Drew. I wanted to touch real quick just around your projected cash burn rate for next year, averaging $1.6 million a month. Do you expect much volatility around that embedded in that estimate?

A: No. The only volatility I’d anticipate around that is if the sales come in faster than we thought. Right now, that’s mostly just the cost of getting the platform up and running. And then as sales starts to materialize, they might bring that use down.

Q: Okay. So the presumption would be once we get past mid-January, say, you should roughly be at that level?

A: That’s right. With all the alignments of the teams, we get down to the 65 people that we are talking about in the press release, that’s right.

Q: Could you just elaborate a little bit on this effort? You’re pivoting to the software model, but your channels to market, how you plan to -- once it’s ready to be commercialized, take to market in the first half of next year. Is this going to be more of a direct model, or do you envision some indirect channels to market?

A: The expectation is you as launch the new service that it will primarily be the direct model. We have a lot of very strong relationships with existing customers on the commercial side and on the government side, and we anticipate being able to enhance those with this new solution. And we’ll use them at least for the early days of selling the solution. And then, as we move forward, then we’ll consider other channels as they present themselves.

Q: Just as a follow-up on that. Should I envision -- as you take this out to existing robotics players, are we mainly talking about traditional robotics? Are these collaborative robotics? Is there a bias one way or another where you think you can find value easier to be realized early on?

A: So, the beauty of the software as it’s been developed is that it allows people to train their arms. As we mentioned, we’ve done it in the lab in a couple of minutes, whereas now it could take potentially weeks or more to train a line and bring it up. And that is just the initial product that would be available. Then the real solutions, Rob, it’s complicated, but the real solutions will allow these systems to train themselves. Sorry, give me one second. I’d like to give you some of the technical words. The platform enables -- the robots learn how to work around unforeseen changes or obstacles by building on their initial programming, and the robots incorporate internal and external inputs that allow them to understand their environment, determine reasonable behavior in unforeseen situations and quickly apply them to a task at hand. And then each newly learned task that the robot has trained itself to do will then be incorporated and used to perform future tasks. And then, it goes to even more sophisticated environment, and then does it in a closed-loop autonomy approach so that the software will help reduce costly workflow stoppages and prevent unnecessary downtimes for customers. So, these are really strong reasons for us feeling confident about taking a path of a SaaS business model using this advanced state of our AI/ML program that began in 2017, with the vision to use these technologies to greatly enhance capabilities in the robotic systems we were developing. And I think as you’re aware, if you look back kind of at our history, we progressed at this stage with our first CYTAR government proposal in 2019. And then we were fortunate enough to have Denis join the company in 2020. He’s been able to really refine the vision for the products and develop them to the state that they’re at right now, and he will continue to head our AI/ML efforts. And then so, with his more than 40 years of experience and everything, we’re in a very good position to capitalize on this opportunity.

Q: Just around -- maybe last question, and I’ll hop back in the queue. But just around the business model itself. How should we get comfortable on the ability to scale this relative to what is a price point? I guess, it will be sold on a per seat basis per arm basis subscription. But just is there any detail you can provide there?

A: Yes. So, there are a couple of things that will enable scaling of the solution. The first -- the very simplest access point for one of our customers will be the ability to do it on a per arm basis. So, they want to be able to train their arm in a few minutes, whereas right now, it could take them weeks or months or longer. And -- so they’ll buy it on a per arm basis. And then there’s all kinds of other modules in the system that they’ll be able to utilize to take on the various tasks that I was just describing a moment ago. And those will be incremental upsells and cross-sells that we’ll be able to provide to them, so that they can take full advantage of the system to the extent that they need it, without requiring it upfront. So they can advance as they go through the various stages and start getting more sophisticated. We’re going to talk about internal and external sources. It could be the camera that’s on -- around the various arm or part of it or it could be external cameras that are not even connected to the arm part of it. So the sophistication of the product and its interpretation of the information is being generated, will then be used by the robot and software to modify the behaviors of the robot for the various needs that customers have designed it for. So, it will be SaaS. There’ll be term licenses. There will be a number of different modules that customers can buy into based upon their various needs and based on the various solutions that they’re trying to put together, and it will be available in a number of different forms so that we can ensure their success.