The LiDAR at the End of the Tunnel
At ECOC 2019 in Dublin, booked-ended by many pints of Guinness, there were several Market Focus presentations dedicated to what has become somewhat of a "beacon of hope" for the broader optical components industry -- the potentially large growth opportunities presented by 3D sensing and LiDAR applications.
The success of Lumentum in generating hundreds of millions of dollars of incremental revenue at attractive margins from 3D sensing has been an important indicator that this market has significant potential.
More than one vendor appeared to endorse a "rosy" outlook for these burgeoning market adjacencies, but there was at least one presenter who was detecting a significantly less bright future for these applications. This had me wondering if some market research vendors have been "kissing the Blarney stone." According to legend, kissing the stone endows the kisser with "great eloquence or skill at flattery" and perhaps, I assume unknowingly, they're telling the industry what it wants to hear.
A company that gave a perhaps unsurprisingly bullish presentation about these markets is II-VI; it had been less than 24 hours since it closed its multi-billion-dollar acquisition of 3D sensing competitor Finisar, which boasted later in the day that it had built the largest vertically integrated VCSEL production facility in the world in Sherman, Texas.
II-VI's Chief Marketing Officer Sanjai Parthasarathi shared a number of bullish market research prognostications:
These forecasts were presented with relative confidence. During the audience Q&A, I asked Parthasarathi about what kinds of assumptions might underpin that forecast for $3.4 billion in consumer-driven revenue for VCSELs, particularly given that "world-facing" applications for 3D sensing haven't materialized in the way investors might have imagined 18 months ago. His answer: "In our opinion, we are just scratching the surface on 3D sensing… $3 billion in three, four years -- frankly to me, I think that's an understatement, I think the market is going to bigger than that."
In a 3D sensing panel led by Robert Blum, Intel's director of strategic marketing and business development for its Silicon Photonics Product Division, other companies shared similarly bullish forecasts. For example, Martin Zimgibl, the former CTO of Finisar, showed a slide projecting the 3D sensing market to grow at a 44% CAGR and to increase 9x from $2.1 billion in 2017 to $18.5 billion in 2023. Of that $18.5 billion, $14.9 billion is expected to be generated by consumer applications and $2.5 billion by automotive. Intel's Blum appeared to present the same data. The CEO of Analog Photonics, Michael Watts, also shared a slide suggesting a LiDAR market of over $7 billion by 2030.
However, there was one more pessimistic presentation sandwiched in between these 3D sensing/LiDAR industry cheerleaders. Vladimir Kozlov, founder and CEO of LightCounting Market Research (my employer), gave a much more conservative outlook in his talk, "Markets for Lasers Used in 3D Sensing in Smartphones and Automotive LiDARS." He cited a number of forecasting principles and insights that he has gathered over the years, including 1) forecasts tend to overestimate markets for new technologies in the short term, but underestimate in the long term; 2) adoption of new technologies takes longer than expected, and legacy technologies continue to improve; 3) price declines are often are surprisingly large. These principles contribute to LightCounting's forecasts for 3D sensing and LiDAR markets that were materially below those shared by other presenters.
For VCSELs for 3D sensing for smartphones only, LightCounting sees the market at around $2 billion in 2024. Kozlov noted that there is a significant portion of the smartphone market that consists of low-end devices sold in developing countries, and LightCounting's forecast assumed only 30% of smartphones (about 500 million units) will use 3D sensors. Additionally, he noted that the upgrade cycle for smartphones has been lengthening from around two years to two to five years. Kozlov also commented that other applications beyond smartphones might contribute another $500 million in 2024, still suggesting a much smaller market ($2.5 billion) than other forecasts are suggesting.
Turning to LiDAR, LightCounting is arguably even more bearish relative to peers -- it sees LiDAR modules for ADAS (advanced driver-assistance systems) revenue to first pass $1 billion in 2026 (with about $350 million for laser sales) and to remain a niche market for the next five years. He pointed out that autonomous driving is somewhat unique from other technologies because potential early adopters need the approval of the more conservative majority of other drivers, as we are all driving on the same roads.
For this reason, Kozlov expects by 2026 only a small percentage (1%) of light vehicles will have a high level of driving automation (L4 or L5 level). Another key assumption concerns pricing; Kozlov noted that L3 modules currently sell for around $600, but LightCounting expects that to drop to $60 over the forecast time period. Kozlov acknowledged vendor skepticism that such price declines will materialize, but noted that 10G transponders in the late 90s cost $20,000; DWDM 10G today is $200 and gray optics $20 -- a factor of a 100x to 1000x lower. Another point was that radar technology could improve and impinge somewhat on the LiDAR opportunity. Kozlov ended his talk by noting the sizeable investments and acquisitions that are being made associated with LiDAR, urged caution, and provided this somewhat "tongue-in-cheek" advice: "Do not panic if you find yourself in a market bubble; enjoy the excitement and all the funding, but in case of a water landing... put on your life vest, look for an exit sign, and do NOT follow anyone’s instructions!"
During his introduction presentation for his panel, Robert Blum expressed skepticism that prices for LiDAR modules would decline as aggressively as LightCounting expects given the sensitivity of the application (i.e., to prevent deaths) and the inherent limitations of competing technologies. He also asks asked his panel participants from Analog Photonics, former Finisar and Lumentum if they were willing to "bet against" Kozlov's bearish outlook for a $1 billion market in 2026.
Analog Photonics CEO Mike Watts gave the most forceful response, suggesting that he was indeed implicitly betting on a more positive outcome, while representatives from Finisar and Lumentum acknowledged that there were significant factors that were out of the direct control of industry participants that create uncertainty around timing, if not the magnitude, of the opportunity.
For private and public investors alike, timing is everything and we'll soon see whether that LiDAR market at the "end of the tunnel" is a beacon of hope or an oncoming train.
— James Kisner, Principal Analyst and Managing Director, Financial Services Business Development, LightCounting