Cerence (CRNC) is composed of its dominant, extremely profitable, low-to-mid single digit growth Edge business, and its nascent, rapidly expanding, likely loss-making Connected Services segment. Cerence currently trades at around $21.50 per share. I estimate that the Edge business is worth approximately $13-20 per share. I advise that Cerence only offers a compelling risk-reward outlook at around $13 per share, given the long-run uncertainty surrounding Connected Services. Initiating a purchase around $13 per share would represent buying the cash cow Edge unit at a price incorporating its worst-case assumptions while receiving Connected Services for free.
Patient investors should stay abreast of Cerence as there are three headwinds to earnings, unrelated to long-term business value, which may arise over the forward 12 months and permit ownership in Cerence at my recommended price levels.
It should be noted that my valuation, wherein Edge is appraised as a standalone business, rests on the assumption that Cerence’s Edge software platform is compatible with Connected Services provided by other companies, need that be the case if Cerence cedes the Connected Service market to competitors. I implore any readers with domain knowledge of this matter to comment on this assumption.
Cerence is an artificial intelligence voice assistant software company focused exclusively on the automotive industry. Cerence was spun off of Nuance (NUAN), a broader AI speech-recognition enterprise, in October 2019. Cerence is by far the largest vehicular voice assistant software provider in the world. Its platform is installed in approximately 325 million cars, including 48 million shipped during 2019. Cerence’s technology was embedded in 52% of all vehicles sold last year. Cerence sells to all of the major auto-manufacturers and their tier 1 suppliers globally. Its revenue is split between the Americas, Europe and Asia in percentages of 44%, 33% and 23%, respectively.
Cerence offers a “white-label” platform that powers the voice assistant in the vehicle’s infotainment system. The platform is fully customizable by original equipment manufacturers (OEMs) which allows the car companies to personalize the assistant as fit to their brand and vehicular models. The voice assistant interacts with the driver across a variety of commands, including navigation (“which exit do I take for Harrisburg?”), safety (“turn on my windshield wipers”), communication (“text Jake that dinner will be ready at 6”), convenience (“where is the closest Chinese restaurant?”) and entertainment (“play The Investors Podcast”).
Cerence operates through three segments. The first is “Edge” software, which is installed on the vehicle’s head unit and accounted for 56% of 2019 revenue. Edge software is the old-fashioned type – without internet connection. Cerence estimates that Edge voice assistant technology was embedded in 59% of vehicles in 2018 and will reach 85% penetration by 2023. Cerence has 80-85% market share in Edge automotive voice assistants and runs at 99% gross margins.
The second segment is “Connected Services”, which was 26% of 2019 revenue. Connected Services provides an internet-connection to the Edge-powered assistant. Cerence refers to this offering as its “Hybrid Solution” because the voice assistant is equipped with both traditional Edge plus Cloud capabilities. Cerence believes that just 12% of vehicles were furnished with internet-connected voice assistants in 2018, but that this technology will eclipse 50% penetration by 2023. Cerence currently has a 40-50% market share in Connected Services as estimated by the number of vehicles in which its Connected Services operates, management estimates of aggregate Connected Services penetration and total vehicles sales, all linked elsewhere in this write-up. Connected Services generates approximately 52% gross margins.
The final business unit is “Professional Services”, which provides implementation and troubleshooting assistance to Cerence’s auto-manufacturing customers and represented the remaining 18% of 2019 revenue. Pro Services is not genuinely its own business – it reinforces the other two segments. Pro Services is likely profit neutral. I will not offer further analysis on this unit.
Edge and Connected technologies possess different benefits as in-car voice assistants. Edge has superior “communication” capacity because it is tailored to each specific vehicle’s acoustic environment which optimizes the accuracy of dialogue. In addition, Edge is more reliable without the need for internet connection. Connected offerings are plugged in to the internet, thereby providing the most updated and accurate information, but sacrificing language processing capabilities in the process. As a result, a Hybrid Solution combining both technologies is the likely path of evolution for the in-car assistant.
Cerence distributes its technologies by entering into multi-year contracts with OEMs. Cerence charges the OEM a one-time fee for each vehicle that the OEM sells containing its technology. Under Edge, Cerence recognizes revenue each time a vehicle is sold to the end customer. With Connected Services, Cerence collects the one-time fee per vehicle upfront, but recognizes revenue over the term of the contract as the company fulfills its performance obligation by providing internet connection to the voice assistant.
The contractual economics result in a substantial revenue backlog representing Cerence’s estimate of its future revenues under all current contracts. The value of the backlog is the average price per vehicle that Cerence charges for its technology, multiplied by Cerence’s estimate of the total number of vehicles containing its technology that will be sold through the life of the contracts. Because OEMs do not commit to producing a specific number of vehicles in a contract, revenue and the value of the backlog are directly dependent on global vehicle sales. However, over the intermediate-term, Cerence can offset any auto market weakness via increased penetration. As of December 2019, Cerence’s backlog amounted to $1.4 billion.
The Edge and Connected businesses are operating in contrasted competitive landscapes. The following section will examine these aspects of Edge and Connected Services, respectively.
Cerence owns 80-85% of the Edge automotive voice assistant market and runs at 99% gross margins. Cerence is the only speech-recognition company in the world focused exclusively on the vehicle market and has just two competitors in the Edge market: SoundHound, a Silicon Valley based company, and iFlyTek, a very large, partially state-owned Chinese entity. Both companies apply their speech-recognition software across a variety of verticals and use cases, one of which is automobiles. As the only pure-auto company, Cerence should sustain its advantage through superior resource allocation, execution and focus. Cerence is also considerably bigger than its Western counterpart, SoundHound, which has 300-350 employees, compared to Cerence’s 1400, including 700+ in R&D alone.
Cerence’s structural advantage in the Edge market is a powerful form of switching costs created by the contractual agreements with OEMS. These arrangements make changing voice assistance software very challenging for the auto-manufacturers. While the OEM does not commit to producing a certain number of vehicles throughout a contract, they do spend considerable time with the software provider specifying attributes of the voice assistant and making it bespoke to the individual brand and vehicle models. The software provider then has intimate knowledge of the OEM’s requirements and preferences and is best suit to fulfill them.
The evidence is strong that Cerence has a wide moat in the Edge market. Its market share is dominant, and its 99% gross margins are, it goes without saying, incredibly strong. This commanding lead provides Cerence with industry and customer relationships and knowledge that are difficult to replicate. Furthermore, Cerence may stand to benefit from the superior growth in Connected Services as competitors divert resources from the slower-growing Edge technology to cloud-connected offerings.
Connected Services Competition
The Connected Services market is highly uncertain due to its very low penetration, estimated by management at just 12% in 2018. While Cerence is the market share leader with somewhere between 40-50% ownership, it is competing with Google (GOOG), Amazon (AMZN), Apple (AAPL) and Microsoft (MSFT), which are all interested in integrating their voice assistants to the vehicle, as well as SoundHound and iFlyTek, whom provide Connected Services in a similar fashion as Cerence. Cerence’s ability to coexist rather than compete with the Tech Giants rests on a strategy called “Cognitive Arbitration”, which is much simpler than it sounds. Cognitive Arbitration is the integration of all third-party voice assistants into its Connected Services. Through Cognitive Arbitration, vehicles with Cerence’s Connected technology would have access to Siri, Alexa, Google and Microsoft voice assistants, as well as those from companies abroad such as Alibaba (BABA) and Tencent (OTCPK:TCEHY).
There are four players in the voice assistant ecosystem: Cerence, the tech companies, the OEMs and the drivers. Cognitive Arbitration has a strong case as the best proposition for each of them. For Cerence, it provides a route by which it avoids displacement by the behemoths. For the tech companies, Cerence’s Cognitive Arbitration would let it bypass building its own Connected software platform while still entering the vehicle. For the OEMs, Cognitive Arbitration administers maximum flexibility. If an OEM were to partner exclusively with a single tech company to provide a cloud-connected voice assistant, it would impede its ability to attract potential car-buyers who already own an at-home voice assistant from another provider. This is illustrated by the following example. If GM (GM) were to partner exclusively with Google, prospective car-buyers who already own an Alexa at home may be reticent to purchase a GM vehicle if they wish to have Alexa in their car. This conundrum, however, is mitigated when the consumer can select any voice assistant to operate in the vehicle through Cognitive Arbitration. Furthermore, this is very valuable because OEMs are global companies. If GM partners with Google, the voice assistant has no utility to drivers in China where Alibaba’s voice assistant is dominant. And for the consumer, if all cars are enabled by all assistants, the full panoply of vehicles are available so that purchases can be based on traditional preferences.
While Cognitive Arbitration seems to be a wise strategy in cooperating with the Tech Giants, it does not protect Cerence from SoundHound or iFlyTek, whom can and have implemented a similar neutral, third-party strategy. SoundHound received recent investments from Daimler, Hyundai and Tencent, demonstrating that it is a formidable threat in automotive Connected Services. iFlyTek, as the local, state-backed incumbent in China, may have advantages in this market where more than 25% of vehicle sales occur. As the Connected Services market is in its early stages, it remains uncertain how Cerence will fare against these competitors.
Why Connected Services will not Displace Edge
At this point, it would be useful to reiterate why Connected and Edge software are likely to work in tandem with one another in future automotive voice assistants as opposed to cloud-based technologies displacing on-location. The first reason is that on-location technology has distinct advantages in its accuracy and reliability of driver-assistant communication. Second, a voice-assistant powered solely by a third-party is not “OEM-friendly”, because the driver communicates only with the tech company, not the OEM’s brand, rendering the vehicle a commoditized piece of hardware. Furthermore, in this situation, the OEM forfeits all of the monetizable data collected through assistant-driver interaction to the third-party. As such, OEMs are unlikely to favor this option. The final, crucial factor is that on-location software is a very low-cost component of the vehicle, about $3-4 per car, and so OEMs are likely willing to pay this small amount to retain brand visibility and data rights.
Critically, it is my understanding of these technologies that Cerence’s Edge software would be compatible with cloud-connection if it was provided by another company. This infers that even if Cerence cedes the Connected market to competitors, it can retain its dominance in Edge. This is a foundational assumption in my valuation that follows, and whose corroboration I have been unable to fully accomplish through secondary research. As stated in the synopsis, I implore readers with relevant domain knowledge to comment their thoughts.
As I have illustrated, Edge and Connected Services are in quite different locations competitively. The segments, as indicated by their respective penetration rates, also have varying growth profiles. And as I will now demonstrate, I believe that Edge is handsomely free cash flow generative while Connected Services is EBITDA-negative. Therefore, if it is indeed true that Edge technology is compatible with other Connected Service providers, need that be the case if Cerence loses that market, then valuing the businesses separately is the most logical approach.
If we accept management estimates that Edge was in 59% of cars in 2018 and will reach 85% penetration in 2023, we can make some useful, albeit rough assumptions about Cerence’s prospective market share and growth. Cerence reported that it was in 48% of all vehicles in 2018 and 52% in 2019. If 59% of cars had on-location software in 2018, and Cerence was in 48% of total vehicles, this is 81.6% market share. If we guess that the penetration rate increases linearly between 2018-2023, then in 2019, Cerence’s 52% of total vehicle penetration out of the ~64.2% of all cars that had on-location software equates to 80.9% share. Management’s penetration forecasts, assuming flat total vehicle sales, imply aggregate Edge software units should grow 7-8% through 2023. If Cerence cedes 1% of the market annually, estimably the pace between 2018-2019, its Edge business will grow 6% per annum.
But Edge revenue was up just 0.6% between 2018-2019 despite boosting its share of vehicle penetration 8% on a relative basis. This is owed to a lower price per unit and a decline in global car sales. In 2019, Cerence generated $3.58 of Edge revenue per car, down from $3.78 in 2018, a decline of 5.3%. At the same time, global car sales fell ~2%. This suggests that Cerence may have reduced prices to maintain market share. However, between 2017-2018, Edge sales grew 15% year-over-year, indicating that 2019’s pricing pressures might be transitory or exaggerated.
With all of this under consideration and assuming flat car sales, my forecast for Cerence’s Edge growth rate through 2023 is 1-7% annually, composed of 7-8% industry-wide unit growth, 1% annual market share loss, and 0-5% unit price declines.
To value Edge, we need to estimate its contribution to expenses below Cost of Goods Sold in the income statement. While the split between Edge and Connected revenue is 56% to 26%, I do not believe that Edge and Connected are running at equivalent margin profiles. The result is that Edge is responsible for less than 56% of consolidated expenses, while Connected is responsible for more than 26%. This belief is founded on the observations that Edge is extremely dominant in its market and relatively well-established, thereby requiring less R&D and marketing spend than the infantile, highly competitive Connected business.
In addition, I will also assume that Pro Services makes no contribution to R&D, Marketing and CapEx, given that it is not a standalone business.
If Edge and Connected had equivalent margins and Pro Services made no impact on R&D and Marketing, then Edge would be responsible for 68% of such expenses while Connected would account for 32%. Because the reporting history for Cerence traces back only to 2017, and disclosure was minimal under Nuance prior to spin-off, estimating current expense contributions by each segment requires a lot of guesswork.
However, we do know that between 2017-2019, Edge revenue increased 15.5%, Connected revenue grew 74.3% and R&D spend ballooned 62%. Furthermore, in 2019 alone, Edge was roughly flat while Connected Services and R&D increased 33% and 15%, respectively. Additionally, in 2019, Marketing expanded 20%. Edge has grown slowly or stagnated while Connected Services and expenses have grown markedly. This would indicate that a greater proportion of expenses are attributable to Connected Services than is suggested by its proportion of sales.
In what is a modestly informed guess, I will value Edge using the following assumptions. In 2019, Edge was responsible for 50% of R&D and Marketing, 56% of General & Administrative and 68% of CapEx. This implies that Edge was responsible for half of R&D and Marketing, that G&A was split evenly across all three segments, and that CapEx contribution was exactly proportional to that of revenue between Edge and Connected. I will then assume that in future years, R&D and Marketing are 100% variable with Edge sales, such that those expense items contribute a smaller percentage of total expenses as Connected Services becomes a greater percentage of sales, while G&A remains evenly distributed across all segments and CapEx remains proportional to sales. When I apply these assumptions to a DCF, using my range of top-line Edge growth assumptions between 1-7%, a 10% cost of equity and a 2% terminal growth rate, Edge is worth between $13-20 per share.
It is worth noting that I assume that the Edge segment absorbs 100% of debt financing expenses, given that Connected Services is likely loss-making, as will be demonstrated shortly. This has a material impact on valuation, as annual debt financing amounts to roughly half of my estimated EBITDA figures. With that said, forward net-debt to Edge-EBITDA is a mild 2x, and I estimate the unit will generate no less than $200 million in cumulative free cash flow between 2020-2024 with $100 million in 2025 EBITDA. Adding in $85 million cash currently on the balance sheet, Cerence should easily cover its scheduled 2025 $197 million loan maturity.
Connected Services Valuation
Management estimates of Connected Services increasing its vehicle penetration between 2018-2023 from 12% to 50% entails 33% industry-wide annualized unit growth, almost precisely the rate at which Cerence’s Connected Service segment has grown its top line over the past two years. While Connected Services is growing sales quickly, if my assertion is correct that a disproportionate quantity of expenses is attributable to Connected Services, then the business unit is losing money and will not be profitable for the foreseeable future. If we assume that the forward revenue growth persists at the two-year trailing rate (~33%) in 2020 and expense estimates are consistent with those used for Edge, Connected will generate EBITDA of -$22.5 million this year. Moreover, even if Connected Services grows 33% annually through 2023 with significant R&D and Marketing operating leverage, the unit will still be -$5 million EBITDA-generative in 2023.
At current prices around $21.50 per share, and assuming that the market values the Edge business somewhere between $13-20, as I believe it is worth, an investor is paying between $1.50-8.50 per share for the Connected business, or 0.7-3.9x sales. Put another way, the market is ascribing between 7-39% of Cerence’s value to Connected Services.
Because I estimate that the Edge business is worth only $13-20 per share, and Connected Service’s future profitability is so unclear in a nascent, highly competitive market, I would not recommend purchasing Cerence at current prices. However, if prices were to retreat around $13, an opportunity would be present to buy the cash cow Edge business at a price incorporating dreary assumptions while getting Connected Services for free. Following a $25.90 spin-off price, Cerence shares fell to as low as $13.85. These price levels may resurface again as there are three superficial headwinds over the next year, unrelated to long-run business value, that may force Cerence shares lower.
Possible Dislocations in 2020
The first is a feature of Cerence’s contractual economics. Because Cerence bills OEMs in advance of its Connected Service performance obligation, it winds up with large deferred revenue balances that roll off into the income statement but do not generate cash. During 2020, Cerence is set to have a particularly stark gap between revenue and cash inflows, as described in the following image. If the market is not anticipating this decrease in operating cash flows, the share price may decline while the underlying fundamentals will be unchanged.
Source: Cerence Investor Presentation, August 2019, Page 32
Second, Cerence is guiding for $20 million in 2020 CapEx related to “stand-up” investments to get itself situated following spin-off. As such, 2020 CapEx will be around $27 million, compared to $4 million in 2019, and $7 million as the normalized rate in 2021 and beyond. This is another FCF headwind that may hurt the share price.
Last, Cerence has reported a very erratic tax rate, ranging from -796.5% in 2019 to 84% in 2018, with a more regular rate of 25% in 2017. This variance is attributable to “intangible property transfers” and “R&D credits”. I used a 25% rate in my models, but other market participants may be surprised by the normalized tax rate and hence, lower net income, which may cause a share price decline.
The primary assumption underpinning my valuation is that the Edge business can sustain its dominance even if Cerence loses out on the Connected Services market, thereby permitting a standalone Edge valuation. If this is true, then I believe Edge is worth between $13-20 per share. At current prices north of $20, value is being ascribed to the unprofitable Connected Services unit. However, if shares fall to near $13 in the future without a meaningful change in the Edge business, Cerence would provide a favorable risk-reward proposition to buy the cash cow Edge business at a bargain price and the fast-growing Connected Services unit for free.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.