Jul 29, 2026 earnings call — Arm AGI CPU demand and record quarter
Record quarter: Revenue $1.29 billion, and AGI CPU demand exceeds $2 billion.
Revenue $1.3B (est $1.3B) · EPS $0.45 (est $0.40)
Key quotes
"Demand now exceeds $2 billion as we continue to add new customers, including multiple customers in the U.S. and China, while the overall value of our pipeline has continued to strengthen."
"Arm Neoverse shipments have now surpassed 1.5 billion cores with most recent 500 million shipping in just the last 9 months, where the first 1 billion took 6 years."
"Total revenue grew 22% year-on-year to $1.29 billion."
Analyst Q&A 9
Supply confidence for AGI chips
Q (Joseph Quatrochi · Wells Fargo): Analyst asked why the company is more confident about upside above $1 billion in AGI revenue and what is driving that. He also asked how gross margins for the extra revenue compare to the original $1 billion plan.
A: Management said demand has grown above $2 billion and their confidence in securing supply for more than $1 billion has improved across wafers, substrates, test capacity, and memory. They said gross margin expectations have not changed: first-generation margins are expected in the high 30% to low 40% range, with a path to 50% over a couple of years. They will give a more detailed revenue and margin update before Q4 at the end of Q3.
"If I look back to the call we had 90 days ago, we had committed to supply of $1 billion." — Rene Haas (CEO)
Smartphone royalties and demand mix
Q (Sebastien Cyrus Naji · William Blair): Analyst asked how higher handset memory costs and OEM price changes affect smartphone unit demand, product mix, and the company's royalty revenue outlook for the fiscal year.
A: Management said they have been somewhat insulated because many customers moved to V9 and CSS, which has supported royalty growth, so they expect double-digit growth in smartphone royalties versus a declining handset market. They also said they see some incremental slowdown across the market and now expect royalty growth for the year closer to the high teens, while cloud AI overperformance helps offset smartphone weakness.
"Right now, if I had to guess, that's probably somewhere closer to the high teens right now." — Jason Child (CFO)
Longer-term AGI demand and supply
Q (Gary Mobley · StoneX Group): Analyst asked about the demand pipeline beyond the next two years (fiscal '29 to '31) and whether the $15 billion AGI revenue target for '31 assumes supply constraints are handled.
A: Management said supply has been a constraint but that their previous TAM view may have been conservative, noting peer estimates have risen well above $100 billion as inference and agentic workloads drive much more CPU demand. They said supply should expand in later years as wafer and memory capacity grows, and internally they have higher targets but are focused near term on reaching the $2 billion level and will update as supply comes online.
"I think all of that is really being driven by the increase of overall compute capacity, which will be inference-based and all of that inference-based compute is going to largely be running agentic workloads." — Rene Haas (CEO)
Where AGI CPU fits in workloads
Q (Thomas O'Malley · Barclays): Analyst asked whether the Arm AGI CPU will serve traditional servers, head nodes, and agentic applications, and which area it will favor.
A: Management said the Arm AGI CPU will play in all three categories and they have customers across each type, citing examples like Cerebras and OpenAI for head nodes and Meta and Cloudflare for agentic and general-purpose server use. They described the CPU as a very good fit across those workloads.
"Short answer is those 3 categories that you defined as described earlier, the Arm AGI CPU is a great fit for all 3, and we have customers in all 3." — Rene Haas (CEO)
Why extra supply is hard; accelerators
Q (Vivek Arya · BofA): Near term: what prevents Arm from securing supply for another $1 billion next year? Longer term: do proprietary accelerators from others limit Arm's opportunity and will Arm add other components beyond a CPU?
A: Management said the market is extremely tight across memory, test equipment, substrates, and TSMC wafers, and even companies with fabs have struggled to meet demand, though they are more optimistic now than 90 days ago. On accelerators, they said Arm can work with custom accelerators: Google uses Axion to interface to TPUs and NVIDIA's NVLink Fusion shows a mix-and-match approach, so there are paths for Arm to connect to proprietary accelerators even though they have nothing to announce today.
"It is a very, very, very tight market across everything, whether it's memory, whether it's test equipment, whether it's substrates, whether it's TSMC wafers, it's a very, very, very tight world." — Rene Haas (CEO)
Core counts and silicon revenue breakout
Q (Vijay Rakesh · Mizuho): Analyst asked whether current AGI CPUs are mostly 128 cores and when 192/256 core CPUs might appear, plus any ASIC color and when silicon revenue will be broken out separately.
A: Management said the Arm AGI CPU today is a 128-core design and the trend is toward more cores, citing Graviton5 at 192 cores as an example, though they had no roadmap details to share. They added that silicon shipments start at the end of this year and once silicon is at least 10% of revenue they will break it out as a third revenue line, which they expect in FY '28 based on their forecast.
"The Arm AGI CPU today is 128 core-based design." — Rene Haas (CEO)
Licensing vs royalty trends
Q (Yu Shi · Needham & Company): Analyst asked whether licensing revenues could offset any royalty weakness and whether a 20% royalty growth rate is still reasonable for FY '29 (CY '28).
A: Management said they are guiding to 13% royalty growth in Q2 and have increased license growth, so on a combined basis they are ahead of earlier expectations and do not expect an overall slowdown. They also said cloud AI overperformance could provide upside and that the 20%-plus royalty growth outlook for next year and beyond still looks intact.
"We're now guiding to 13% royalty growth in Q2." — Jason Child (CFO)
Impact of low-cost / open-source AI
Q (Kinney Chin · TD Cowen): Analyst asked whether the rise of low-cost or open-source AI models is a net positive or negative for the AGI CPU franchise and how it affects Arm's IP business for data center and edge.
A: Management said if the world moves to more open-source models, the impact is neutral to positive because Arm is agnostic: all models still need CPUs and IP to build custom chips. They noted open-source could enable smaller, more efficient models that run at the edge, which would be beneficial for Arm's footprint in smaller devices.
"If the world moves to something that's more open source-based, AKA open weights, for us, I think it's somewhere between a net neutral to a net positive." — Rene Haas (CEO)
R&D timing and single-part supply risk
Q (Timm Schulze-Melander · Rothschild): Analyst asked how R&D spending timing works (milestone-based, capitalization, or other) and whether a single missing component could derail fulfillment of AGI CPU demand.
A: Management said R&D variance was driven more by tool and emulation cloud spend versus headcount, so the Q1 savings can be banked and they expect OpEx to grow mid-single-digit quarter-on-quarter, with tool use flexing as development needs change. On supply risk, they said they are not concerned about a single point of failure and are confident in their partners' capabilities so one missing component should not tip the plan.
"So there isn't one single point of failure that to your point of the 1% that could cause an issue." — Rene Haas (CEO)