Innoviz: A lidar leader for an emerging market (Part 1)
A 10x company making its breakout
Today I'm diving into one of my 'contrarian, big bets', which fits right into my investment style: Innoviz is a company that I believe can return way above market returns (with larger amounts of risk) that the market isn't 'getting'.
This has been a company I've owned for over a year as a speculative investment. It's been a rough two years for the company and entire Lidar sector, and risk management has been the main challenge (which I'll cover in my 2024 year in review post). Earlier this week the investment case has changed dramatically for the better, which is why I've decided to share the investment case now.
The automotive market has always aimed to improve convenience and safety. This has played out time and again: mapping and navigation, infotainment systems, seat belts, cruise control, and driver assistance systems (such as Mobileye) are just a few examples. Each rollout has played out the same way: an S curve adoption model, that slowly, yet inevitably, leads to mass adoption and new successful companies. Mobileye, now a $15 billion market cap company, is the most recent example. I believe Innoviz is the best candidate for the next wave of such technologies.
Innoviz provides a core technology, LiDAR, which enables the next wave of convenience and safety: semi/fully automated driving from highways to fully functioning autonomous vehicles.
Lidar is a nascent industry, with new hardware and software, which follows the same multi year adoption cycle that other technologies in the automotive industry followed. While the companies in this industry face meaningful challenges today (reflected in their market caps and volatile trading), if they are successful, they will achieve multi billion dollar revenue run rates.
This deep dive will be broken into a few parts include the following:
Company overview
Market need and dynamics
Company analysis:
Traction
Technology
Go to market
Management
Competition
Risks
Investment checklist
Position sizing and buy + sell plan
In today’s post, Part 1, I’ll explore the market need and dynamics (To read part 2 click here and for part 3 here).
Company overview
Innoviz is an automotive Tier-1 (i.e they supply directly to the OEMs) developing LiDAR hardware and software to automotive original equipment manufacturers (OEM). They were founded in 2016 by a very talented management team, were quickly anointed as one of the darling LiDAR companies of the late 2010's (of which there were many many competitors) and took Innoviz public in 2021 via a SPAC. The entire sector which boomed, has been busting in the public markets with the company currently trading at ~$1.30 (down from $10 at IPO).
Innoviz is a well run company in an emerging industry with massive potential throughout the end of this decade and into the next.
Market need and dynamics
This section delves into the automotive market: the need for lidar, technology adoption cycles, competitive market dynamics, regulation and geopolitics all in reference to how this could affect the lidar market.
The need for Lidar
The automotive market has always aimed to improve convenience and safety. This has played out time and again: mapping and navigation, infotainment systems, seat belts, cruise control, and driver assistance systems (such as Mobileye) are just a few examples. Each rollout has played out the same way: an S curve adoption model, that slowly, yet inevitably, leads to mass adoption and new successful companies. Mobileye, now a $25 billion market cap company is the most recent example.
Lidar, or 'light detection and ranging' is the technology that leads to the next levels of enhanced safety and comfort that the automotive industry is looking for. Mainly autonomous highway driving, in city autonomous driving and full on autonomous vehicles. I have no doubt that the adoption that was shown for Mobileye products (lane assist, collision avoidance) will occur for these enhanced versions as well. It's only a matter of time.
What drives the market need for LiDAR?
Safety
Comfort
Autonomous driving
Safety and comfort are two use cases that have driven automotive innovations for years. Autonomous driving is the culmination of these two trends.
A great comparison I like to think of in this market is Mobileye, which developed safety features with added comfort. These include collision avoidance and lane assist and is now migrating upmarket into autonomous driving/advanced driving assistance systems (ADAS). It's taken Mobileye a good 20 years to grow into the industry leader they are today, but it's been worth the wait. Mobileye was founded in 1999, IPOd in 2014 at a $5B valuation, was acquired by Intel in 2017 for $15B, was IPOd again in 2022 and after a rough year is currently worth $15B.
It's taken Mobileye 20 years because of the dynamics of the automotive market. The automotive market is very slow moving. New cars take 3-5 years to be designed and have a long lifecycle. Thus, new products can take years (literally) to hit the market and decades to ramp up.
This time frame for product innovation is something we've seen before with technology products such as mapping and navigation, which took a decade to reach saturation, and Mobileye. While initially slow, once products begin to ramp up, it occurs quickly as entire car platforms join almost all at once.
The Lidar market has been much slower in it's rollout than industry analysts had first believed, by a wide margin, but it's only a matter of time. I'll dive into my thinking regarding the rollout of the timeline below.
First, it's important to understand the industry standards and levels of autonomous driving:
Level 0: The driver does everything, even with warnings or temporary help like collision alerts.
Level 1: The car has some help features like emergency braking and lane assistance.
Level 2: The car gives more help with steering and speed control, but the driver must still be fully engaged.
Level 2+: Similar to Level 2, but can sometimes handle hands-free driving; eyes on the road are still required.
Level 3: The car can drive itself in certain conditions, but the driver must take over when asked.
Level 4: The car can drive itself without any driver input in most situations, but a driver can still take control.
Level 5: The car is fully autonomous and doesn't need a driver for any driving tasks.
Lidar enables all functionalities above L2 for several vehicle segments: commercial vehicles, trucking (whether L2+ or more fully automated vehicles related applications such as platooning) and robotics. The company is focusing on the commercial vehicle segment.
As Innoviz's management puts it:
"The ADAS+/ Level 2+ segment of the consumer automotive market is expected to grow significantly over the short to medium term. ADAS+/Level 2+ systems are still technically in the driver-assist realm but incorporate a whole new layer of functionality on top of the traditional ADAS features. For instance, in Level 2+ vehicles, map data may enable lane-centering functions to remain effective even in areas where sensing-only lane centering systems may face difficulties, such as in areas without visible lane marks or low quality lane markings, ramps with sharp turns, junctions, roundabouts, newly paved roads or urban settings. Level 2+ also supports automatic lane changes by providing information such as lane-marking types and adjusting the drive speed according to road speed/curvature." (2023 Annual filing)
Automotive adoption
The automotive industry's embrace of new technology is a collaborative effort between Tier 1 suppliers and OEMs. Sometimes this process will be 'bottoms up', led by Tier 1s (such as Bosch, Continental, and Denso) who will develop and pitch innovations to OEMs. Other times, this can be 'top down' with OEMs deciding they want to implement specific new features and directing their respective Tier 1 suppliers to find the best solutions.
OEMs usually select their Tier 1s based on longstanding relationships and geography. Toyota for example, has a long standing relationship, and even partial ownership stake, in Denso.
This means that while the automotive market is large and has several dominant players, it's rarely a winner takes all market. This is due to the competitive nature between different Tier 1s and between OEMs. However, sometimes, it can be a 'winner takes most' market. Here and TomTom, mapping and navigation providers were dominant (until Google Maps displaced them) and Mobileye displayed this dynamic in the ADAS market with ~60% market share (or more by some estimates). I'll discuss why Lidar can be one of these adoptions as well.
Market share for innovation in the automotive market usually starts in the luxury segments and 'thought leader' European OEMs (BMW, Mercedes, VW, Volvo) and then trickles down into lower tier cars. For example, BMW was one of the first OEMs to integrate Mobileye's ADAS functionality.
Geography is another important factor. Each major geographic region has their dominant OEMs and Tier 1 suppliers. A few examples are:
APAC: Toyota, Honda, Nissan, KIA, Hyundai, Mahindra, Tata, Mahindra
Europe: Mercedes-Benz, BMW, VW, Renault, Fiat
US: GM, Ford, Stellantis
Winning in one region doesn't necessitate winning in another as suppliers, regulation and geopolitics all play a role. Specifically for Lidar, electric vehicle (EV) and new platform adoption has been a significant factor in lidar adoption (or non adoption as EV platforms have been delayed among most OEMs). China has led the way here, and has the most established EV and lidar market. However Chinese lidar companies are at a disadvantage in the west: they've have made their way into the restricted company list among western regulators and their technological capabilities as a whole have not compared to western lidar companies. The same is true in reverse - it's very hard for western lidar suppliers to win in the Chinese market where price and a local supply chain is an important factor and where safety and quality are less important. Mobileye has recently lost share in China around just these dynamics. Therefore I'm discounting the entire Chinese market for this analysis.
Regulation also plays its part in automotive technology adoption, especially in everything safety and environmentally related. Both regulatory tailwinds for lidar adoption. Since electric vehicles are such a significant change from internal combustion engine (ICE) vehicles, OEMs have been using this platform shift to modernize many different aspects of EV car platforms, making these platforms more software enabled, including ADAS and autonomy functionalities.
On the safety side, NHTSA, the USA's National Highway Traffic Safety Administration released proposed guidelines on May 31st 2024 regarding the implementation of automatic emergency braking (AEB) in all new car models from 2029 onward. Importantly, the mandatory standards include AEB at night, conditions that are challenging for radar and camera only AEB systems. If passed, this would be a significant tailwind for lidar companies. The European regulatory system, NCAP, also mandates safer vehicles by the end of the decade.
Lidar in automotive
Many Lidar company CEOs have voiced their belief that Lidar will play out in a winner takes most market, similar to ADAS and mapping, especially Omer Keilaf, Innoviz's CEO.
The two main reasons for this are:
The technical complexity of a full solution for a new technology
The challenges of integration into a vehicle platform and OEM for a mission critical technology.
Lidar and the perception and driving policy software stack that accompanies it, are core to vehicle safety and performance. These aren't easy to integrate, and aren't easy to replace. We've seen the integration challenges in both Luminar and Volvo's partnership and with GM and Cepton. Both have had multiple delays around software integration: Volvo finally launched their X90 with Luminar’s hardware but no advanced features based on the Lidar. GM and Cepton cancelled their partnership. The hardware isn't simple yet either. Since the industry has yet to standardize on form. This makes Lidar a very sticky product in the market.
I'll let Sumit Sharma, the CEO of Microvision, one of Innoviz's competitors explain the challenges in becoming a tier 1 supplier for Lidar, from Microvision's Q1 2024 earnings call:
"In each RFQ, OEMs require significant customization of hardware, firmware and perception software. Their timelines for customization and qualification are long and would require several hundred engineers for several years...Some OEMs want to see our manufacturing strategy proposals to commit to factories in Asia and North America for volumes that would not justify two factory locations. Some OEMs explicitly want a factory in the U.S. To be clear, they will not accept a NAFTA country, but only a U.S. contract manufacturing factory while expecting cost structures that are only possible from Asia. Others will only review proposals from Asia, while a small group wants to see a diversified operation strategy with multiple continents. Again, the expectation is that we will fund this with our investors.
All OEMs want varying levels of perception features, some running within the LiDAR, some running in their ECU, some claiming they need no perception, but want our source code. Some OEMs want the LiDAR in roofline, others want them integrated in headlamp and some others are only looking at behind windshield integration. They want our core LiDAR to be flexible enough to fit into all their locations. They are aware of the trade-offs in each location, but will require updates to the core hardware.
These challenges in finding the right partner are part of what make it such a challenge to replace an existing partner, and what makes a lidar supplier sticky. Hard to integrate means hard to replace. Multiple lidar management teams have expressed this sentiment over the past several quarters, believing that now is the land grab for the next 5-10 years of the market. This understanding of the market has led most lidar CEOs to opt for a 'scorched earth' policy where any deal they get a foot in the door into an OEM is a good deal, even if the unit economics might not be entirely favorable.
Most of the deals are built on Non Recurring Engineering costs (NRE). The OEM usually provides the tier 1 payments upon certain milestones to fund development until product sales can occur. NREs are often what fund a company until start of general production and are a good leading indicator for real partnerships. NREs can be accounted for in various ways, which are a challenge for investors, but the key is that this revenue take the form of cash that is relevant to funding the company and importantly, isn't adjusted later down the line as a revenue headwind (this is part of what's occurred for Luminar in 2024).
Summarizing the market
Enhanced driving, from L2 - L4, is the future of automotive innovation and will arrive by the end of this decade in mass produced vehicles. Lidar is a key enabler to this. Similar to previous innovations like lane assist, AEB and mapping and navigation, autonomy will be adopted in an S curve that will take multiple years. It's slow going in the beginning, but kicks up quickly in the mid years. Penetration for these products is often done via the premium brands and then trickles down to mass adoption. Relationships with key Tier 1's and OEMs is critical and varies by geography.
These factors give the Lidar market the attributes of a 'winner takes most' market at best and a highly sticky product with strong competitive moats at worst, justifying higher assumptions of future cash flows from winners. As for timing, it's getting clearer that 2026-2027 is the inflection point for early adoption and 2028-2029 the years where the S curve starts to pick up. First L2+ and L3 models are hitting the road in 2024 (BMW's i7) with more models in late 2025 and a larger uptick in 2026. Some notable examples are Ford's Bluecruise, GMs Supercruise and Mercedes.
Picking the right winner and getting the timing right to cross the chasm of adoption are the most important factors in picking a Lidar winner, a winner that will be large indeed in automotive supplier terms. The key questions for picking this winner are:
How many design wins will they have? What market share penetration will they achieve?
Will they be able to survive financially until mass adoption?
Will they be able to achieve high gross margins and not be commoditized?
Now that we've understood the market dynamics, in part two I’ll explore why Innoviz is the right choice for this market. I'll run through:
Traction
Technology
Go to market
Management
Competition
Risks