SoftBank’s Arm is seeking the largest Nasdaq IPO in almost two years, targeting a valuation of $60-70bn despite flat revenue and declining net profit.
Third Bridge has been monitoring the company’s IPO prospects since February 2022, providing investors with early-stage, predictive intelligence on Arm’s growth prospects. Gain access to these critical insights through our comprehensive coverage.
Simply submit your details via the form to access the content. Learn how our integrated solutions can help you make faster and smarter investment decisions, when, where, and how you need them.
Specialist
Former Director at Arm
Agenda
- Evolution of Arm’s key use cases for automotive – ADAS (advanced driver-assistance systems), AV (autonomous vehicle), cockpit, powertrain, steering and suspension
- Level 4 and 5 AV development, early adopters and future opportunities
- Automotive royalty rate and licensing fee evolution and key customer dynamics
- Competitive landscape vs Intel (NASDAQ: INTC)
Questions
1.
Arm Ltd reported about USD 2.7bn revenue in 2021, which was above expectations of USD 2.5bn. How much of that comes from the company’s automotive solutions?
2.
Do you expect automotive solutions to increase as a percentage of Arm’s revenue over the next 2-5 years?
3.
When do you think 10% of Arm’s revenue could come from automotive?
4.
How do Arm’s automotive solutions split out across ADAS [advanced driver-assistance systems], AVs [autonomous vehicles], cockpit, power train, steering and suspension?
5.
Intel estimates the automotive chip TAM could be around USD 115bn by 2030, and that chip will make up about 20% of a car’s bill of materials by then. What percentage of semiconductor content in cars could chips made with Arm architecture comprise by 2030?
6.
In terms of the fastest-growing chip segments for Arm, you said the ADAS area will increase the semiconductor content in chips from USD 100-150 to USD 200 per car. When do you think level 4 and 5 AV will be adopted more widely?
7.
What key features will drive up semiconductor content by 2030? What key growth drivers will give Arm the opportunity to have 30-40% of semiconductor content be from its architecture?
8.
To what extent could a changing macroeconomic environment or a potential recession in 2023 slow down demand in the automotive sector and perhaps push out the estimates you mentioned?
9.
How will Arm establish market leadership in AVs, which is one of the company’s aims in SoftBank’s annual report? Would this come from increasing R&D spend and having a better future road map?
10.
Thinking about Arm’s solutions capabilities and looking at its CPUs [central processing units] and GPUs [graphics processing units], have there been any difficulties associated with its key processor offerings and when could those difficulties be overcome?
11.
How much bigger are the royalties from GPUs than CPUs? You mentioned the average royalty was less than 1%.
12.
How does Arm overcome its problems with GPUs and people in-housing them rather than sourcing them from the company?
13.
How difficult would it be for Arm to steal its customers from Imagination to get the GPU business itself?
14.
Arm has obviously done a very good job in the mobile CPU market – I think it mentions having 95% market share in mobile devices, mostly in smartphones. Could we see a similar market dominance in automotive solutions such as AV, ADAS, cockpit and so on?
15.
What’s Arm’s market share in SoC [system on a chip] IP?
16.
Is the main competitor Intel on the CPU side?
17.
Is the main competitor on the GPU side companies such as Imagination and Qualcomm?
18.
You mentioned what Arm could do to improve its market share in GPUs. If the company did exactly what you described, what share could it gain on the GPU side of SoC IP over the next 2-5 years?
19.
Why don’t Renesas, NXP and Qualcomm use Arm’s GPUs if they use its CPUs? Arm claims that based on its architecture, its CPUs and GPUs use less power than competitors’, and having energy-efficient components in a car is obviously very important. If the company’s claims are true, why isn’t everyone using its GPUs or why aren’t as many people using its GPUs as are using its CPUs?
20.
When do you think we could see Renesas and NXP switching to Arm’s GPU IP?
21.
It’s quite difficult for companies to switch GPU providers, even if the benefits of moving to Arm’s architecture from using less power are pretty significant. How much more energy efficient are the company’s GPUs and CPUs than its competitors’?
22.
Customers will look at a broad set of purchasing criteria and energy efficiency isn’t necessarily number one. What would you say are the most important criteria when customers choose who to go with for their IP?
23.
When might Arm see significant royalties from solutions it launched in 2020? I know the company launched the Arm Cortex-A78AE, the Mali-G78AE GPU and the Mali-C71AE image signal processor, and I understand royalties start to come 2-3 years after the licence is assigned. Could Arm see significant royalties in H2 2022 or 2023?
24.
If the C71 will be a component of the camera, would its main application be in AV?
25.
How does the C71 compare to OnSemi, Sony and so on?
26.
When do you think Arm will start to see significant royalties from the C71 and the G78?
27.
When might the Cortex-A78 start to see royalties?
28.
Looking at Arm’s other automotive solutions – perhaps launched since 2020 – what are the other main CPUs, GPUs and ISPs [image signal processors] licensed for automotive use cases?
29.
Why would a company such as Tesla go with Arm’s Cortex-A72 architecture over its newer architecture? Is it because it’s cheaper?
30.
Could Tesla license some of the newer IP solutions that have come out?
31.
What other licensees does Arm need to win to increase its revenue from 5-7% to 10% by 2025-27?
32.
Arm signed 168 licensing deals in 2020. What rough percentage of those would have been for automotive? Would that have been in line with the 5-7% we discussed?
33.
How does Arm increase penetration within its customer base, which sounds as if it’s reached its limit?
34.
The AVCC [Autonomous Vehicle Computing Consortium] is comprised of Arm, NXP, Bosch General Motors, Toyota, Denso and Continental, and I think there are now actually more companies involved. What does that mean and how does it compare to something such as Intel’s Mobileye?
35.
Why did Intel Mobileye deny the AVCC and how do these two compete?
36.
Which robotaxi company do you think would suit AVCC the most?
37.
Will robotaxis start to see high adoption before or after 2030? The challenges that need overcoming seem the same for personal cars vs robotaxis and level 4 and 5 AV. Is that correct?
38.
If personal vehicles won’t come until after 2030, when might we start to see robotaxis driving around with Arm architecture?
39.
Why have robotaxi companies gone with X86 over Arm?
40.
You said the idea of AVCC was to establish a common computing architecture for automotive computing. How is that progressing?
41.
Was the AVCC impacted by the Arm-Nvidia takeover not going through, given I understand Nvidia is part of the AVCC? Has that hindered progress towards establishing the common architecture?
42.
What percentage of R&D attention or spend does Arm give to automotive?
43.
Would you like to highlight anything else around Arm’s automotive chip IP business or automotive chips in general?
