The ecosystem disruptions wrought by ADAS and AV development have made the automotive-semiconductor market more hospitable to startups. This four-part series explores the contextual drivers of the recent increase in startup activity, the challenges these new players will face as they build their businesses, strategies for success, and the potential impact of the Covid-19 pandemic on the startup trend going forward. The following is the Part 2 of a four-part series on ‘On-Ramp Widens for Automotive Semiconductor Startup.’
Proof of concept: The first step on the road to traction
While automotive OEMs and Tier 1 vendors are increasingly open to the idea of working with startups, the high development costs, long product lifecycles, and stringent safety requirements of their industry understandably make them risk-averse. The most common way for an OEM or Tier 1 to engage with a startup semiconductor company without actually making a significant commitment is to engage in a proof-of-concept (POC) project.
In the traditional, hierarchical supply chain structure, “winning” a POC with a major Tier 1 or OEM was considered a major vote of confidence by the customer in the promise of the new solution. While there were no guarantees that production volume would follow a successful POC, the commitment of resources on the part of the customer was generally understood to mean that the customer was relatively serious about the solution.
However, as the automotive industry acknowledged the need to be nimbler and faster in its ability to innovate, it began to experiment with a variety of riskier technologies. Innovation centers were established that became skilled at running multiple projects simultaneously, with the ability to cut projects on short notice before committing anything to a production program. When these POCs are managed by the innovation centers and not out of the actual platform R&D centers of the car OEMs, in many cases they never find their way into a production program.
This is not to say that such POCs are not immensely valuable to the startups. Potential investors will often take POCs into account when assessing a startup’s traction in the market. Feedback from the POCs can be used to fine-tune the product/market fit. Nonetheless, it is critical that the startups have realistic expectations about the likely outcomes from the project and understand both the risks and the benefits, even for POCs that never result in actual recurring revenue.
The long, winding, risky road from POCs to production platforms
Some startup companies have been able generate a reasonable amount of revenue from engaging in paid POC projects with the automotive industry. This has been especially true of very advanced, very expensive technology such as LiDAR, which needs to be tested out across a substantial number of vehicles. According to research firm Yole Développement, the LiDAR total available market (TAM) approached $1 billion in 2019, with less than 10% of that being in actual platform production vehicles.1 However, sustainable revenue growth and lofty enterprise valuations on the basis of POCs are the exceptions, not the rule. In the vast majority of applications, the real growth will only occur after a solution has been “designed in” to a production vehicle platform.
Semiconductor startups and their investors would be wise not to start banking on revenue from an automotive program until after their POCs have started achieving pre-aligned, customer-approved milestones and customer engagements have broadened to include procurement, supplier quality, and platform engineering. Furthermore, since semiconductor companies are typically Tier 2 suppliers, the transactional customer for doing business will normally be a Tier 1 supplier to the OEM. As such, the startup’s customer will need to ensure that the solution is approved, designed-in, and qualified in a production-vehicle program before actual revenue can begin.
It is not uncommon for the full qualification cycle at the Tier 1 and OEM to take up to three years. Once production does start, the volumes during the first few years may only be a fraction of the anticipated volumes because OEMs tend to roll out new features one or two car platforms at a time.
If the startup is sufficiently well-funded, it may be able to hire the resources it needs to ensure it meets the quality requirements of the automotive industry. This often will require going beyond qualification testing to AEC-Q100 requirements. IATF 16949:2016 specifies the quality management system (QMS) requirements necessary for automotive suppliers, with a focus on continuous improvement, defect prevention, variability reduction, and elimination of waste in the supply chain.
The QMS should include advanced product quality planning (APQP) and a production part approval process (PPAP). The PPAP alone is a remarkably detailed process consisting of up to 18 separate elements, which include establishing an automotive-capable verification and validation (V&V) process, complaint-handling flows, design and process failure-mode–effect analyses, and other requirements.2 Furthermore, depending on the application, the quality requirements may include meeting functional safety standards (ISO 26262).
To make things more challenging, the requirements are still evolving. For example, a consortium of automotive companies, including Aptiv, Audi, BMW, Continental, Daimler, FCA and VW, together with semiconductor companies such as Infineon and Intel, recently published a set of “Safety First for Automated Driving” V&V guidelines and established a new standard: ISO 21448, Safety of the Intended Functionality (SOTIF). Far more comprehensive than ISO 26262, SOTIF includes cybersecurity and software verification as well as the more traditional automotive requirements.3
Startups that lack the resources to address these evolving requirements directly may need to leverage the quality assurance capabilities of their larger ecosystem partners, including foundry suppliers and test subcontractors. This does not mean simply leaving the testing up to the partners and trusting that everything will be accepted by the customer; rather, it means the startup may need to have the ecosystem partner provide certain contractual assurances to the customer. In rare cases, if the startup is a Tier 2 vendor or lower in the value chain but has a particularly strong relationship with a customer higher up in the value chain, that customer may be willing to accept the risk with respect to the OEMs.
While perhaps the most daunting, quality assurance is just one element of the complete set of operational-excellence and contractual requirements that startups must meet. A startup may also need to take steps with its predominantly outsourced supply chain to ensure continuity of supply, second-source requirements, and compliance with hazardous-substance and ethical-sourcing regulations, to name just a few. Startups will also require a strategy to deal with contractual liability clauses that may be thrust upon them by their customers or the OEMs.
Selling into the automotive industry involves risks to the startups besides simply overcoming the barriers to entry. Large companies have been known to engage with startups while simultaneously developing their own solutions, sometimes learning from the startups in the process. Even if the relationship develops well, sometimes the best outcome a startup can realistically hope to achieve is to be “acquihired” by the larger company.4 While this is not necessarily a bad outcome, it may prevent the startup from ever achieving its full potential or the full market potential of the technology it developed. Of course, the more partners the startup has, the less likely it is that the startup will be forced into accepting an unattractive offer from a partner.
OEMs and Tier 1 suppliers face their own risks when they choose to work with startups. Occasionally, the larger companies will be developing their own products in a target market while simultaneously sourcing from a startup. That creates a risk of IP leaks within the larger organization. The established company will be very good at protecting its own assets from the startup but may have insufficient protections in place for the startup’s IP, which could damage the startup’s business and create potential liabilities for the established company down the road.
Perhaps a bigger risk for established players in working with startups is the possibility that the emerging company may go out of business or be acquired by a competitor. However, the most important risk is also the most fundamental: program and schedule risk. If the startup is unable to deliver as promised, Tier 1 or OEM customers could be putting entire programs at risk. Delays in R&D could force a platform design overhaul or require a platform to launch without the features that motivated the relationship with the startup in the first place. Delays during the production phase could result in line stoppages at the OEM at a cost of multiple millions of dollars per day.
When the risks are worth the effort
Given how challenging it can be for startups to meet the production requirements of the automotive industry and how risky it can be for car OEMs to incorporate unproven technology into their production vehicles, an obvious question for both sides of the relationship is why they should even make the effort.
Certainly, for the startups, there may be easier markets to go after with faster “time to money.” However, because the automotive industry has such high barriers to entry, the companies that are able to break through will have fewer competitors with which to share the market, resulting in higher profit margins. Furthermore, once production revenue starts, the product lifecycle in automotive is much longer than in most other markets, ensuring a stable source of revenue and profits. Finally, over the long run, across multiple economic cycles, the automotive industry has demonstrated an ability to maintain a fairly healthy and stable growth rate.
For the OEMs and established Tier 1 suppliers, working with select startups can provide rapid access to new technologies, often at a much lower cost than if they had developed the technology themselves or commissioned it from their traditional supply chain partners. Rather than build and resource their own projects from the ground up and compete internally for limited budget inside their bureaucracies, they can quickly adopt new technology from dedicated, focused, and agile external teams. If the larger company aspires to enter the market that is being evangelized by the startup, establishing a solid supplier relationship early on can make it easier to consider that startup as an acquisition target down the road.
So how can both sides increase their chances of making such a seemingly precarious engagement succeed? Part 3 will present some strategies.
1“The Automotive LiDAR Market.” Yole Développement, April 2018.
2IATF 16949:2016. International Automotive Task Force (IATF), October 2016.
3Junko Yoshida, “AV Safety Ventures Beyond ISO 26262.” EE Times, March 5, 2019.
4Tanja Kufner, “Automotive startups grow with increasing OEM uncertainty – here is how.” AutomotiveIQ, Oct. 9, 2018.
The author is Drue Freeman, Board Member, Sand Hill Angels; Adviser Board Member, Silicon Catalyst
The post Risky Relationship between Startups and Auto Industry appeared first on EETimes.