Arm FQ3 24 – AI Pixie Dust

Arm silences the critics of last quarter

• Arm topped off a good set of earnings that silenced its critics and added a sprinkle of AI pixie dust that gave the market confidence that the AI story is real driving the shares up 19% in after-hours trading.
• FQ3 revenues / Adj-EPS were $824m / $0.29 ahead of estimates of $762m / $0.25 and there was a promise of more to come in FQ4 24.
• Here, Arm increased its FY2024 guidance by 5% from $2.96bn – $3.08bn ($3.02bn) to $3.16bn – $3.21bn ($3.18bn) and reduced its expenses forecast by 4% driving a 16% increase in the adjusted-EPS forecast from $1.05 (midpoint) to $1.22 (midpoint).
• This quarter is important because in FQ2 24 Arm told the market that its slight revenue miss was due to the timing of deals being signed which the market chose not to believe and sent the shares down anyway.
• This quarter has shown that the company was telling the truth meaning that next time there is a timing issue, the market will be more inclined to take the company at its word.
• The key to the forecast increases is licensing revenue which came in at $384m for FQ3 24 and was stronger than both the company and the market expected.
• More companies are choosing to license Arm for use cases outside of smartphones and tablets and crucially, companies are choosing Arm to run AI on edge devices.
• RFM research has long concluded that AI, and generative AI in particular, is going to be most effective when inference is executed on the user’s device rather than in the cloud.
• This makes economic sense for the provider of the AI service and also means a faster response time and greater data security as the user’s data never leaves the device.
• 2023 was the year of training while 2024 is shaping up to be the year of inference and Arm’s license revenue performance is early evidence of this trend coming to fruition.
• Visibility in the short term is also good as Arm has roughly 12 months of revenue rolled up in signed agreements that has yet to be recognised as revenue.
• These remaining performance obligations (RPO) grew by 38% YoY in Q4 to $2.43bn which is a very good indication of the kind of revenue growth that Arm could be about to experience going into FY2025.
• Arm does not expect AI to be a big royalty driver for the next year or so and it looks like the main story is going to be licence revenues but there is potential elsewhere that no one seems to be paying much attention to.
• This is the potential for the Windows ecosystem to migrate to Arm on laptops where Qualcomm looks like it has a compelling product and Arm has a processor in the works called Blackhawk that should provide other chipmakers the ability to produce processors with similar performance.
• The key to this is a seamless migration of Windows from x86 to Arm which has already been tried twice with dismal results.
• However, Arm says that it is confident that Microsoft will get this right 3^rd time around, especially in light of how flawlessly Apple managed to do the same thing.
• Hence, I think there is scope for Arm to benefit from taking share from x86 in a large market where the ASP of the chips are high and the processing requirement large.
• I don’t think that there is very much of this at all in the expectations for Arm and in many ways, this could be a more interesting short to medium-term revenue proposition than AI.
• AI will allow Arm to earn more revenues from higher prices and an increased core count which is good but not as good as laptops where it currently has very little.
• If the port to Arm goes well, then Arm’s revenues here could substantially increase which I think would have a disproportionate and larger impact on its financial performance than AI in the same time frame.
• However, the story of these results is the first tangible impact of the promise made at IPO that Arm would become an AI processor business and right now the market loves anything that has exposure to the current frenzy.
• This means that as long as the AI bubble does not pop, the market will be willing to look at Arm more like Nvidia and less like Intel which will provide good support to the valuation.
• Furthermore, Arm is more diversified than Nvidia so if the AI bubble does pop in 2024, the impact on its fundamental financial performance will be much less than it will be for Nvidia.
• The net result is that Arm has silenced the critics of FQ2 2024 and is beginning to produce the goods that it needs to justify the valuation at which it trades in the market.