The long-term story for the company remains intact, despite near-term uncertainties introduced by competition.
Software revenue accelerated to 15% and the company guided for a significant sequential increase, which translates to an additional $1 EPS annually.
Management provided unusually detailed customer and GW color, including significant commitments from Anthropic, OpenAI, and Meta, and announced an AI XPV platform with Apollo/Blackstone to deploy over 20GW of compute capacity through 2028.
Infrastructure software showed strong performance, with the FQ3 guide significantly exceeding Street expectations, driven by VCF 9.1, strong server demand, and enterprise AI inferencing workloads.
The decline in FQ3 gross margin guide is due to product mix, not a structural change in semiconductor margins, and management expects gross margins to remain above 74% in the coming quarters with minimal operating expense increases.
The company has become a clean way to invest in the AI infrastructure cycle, with unique exposure to custom XPUs and Ethernet/networking.
Bulls are focused on AI bookings exceeding $30B, a six-customer roadmap, multi-GW commitments, and the extension of the AI opportunity into FY28 and FY29.
The combination of commitments from Anthropic, OpenAI, Meta, Google, two additional customers, and the >20GW XPV platform through 2028 indicates that FY27 is a bridge to a much larger FY28 deployment year.
The company now has three growth vectors: AI semis, software, and non-AI semis, with software reducing concerns about AI impacting VMware and potentially becoming a tailwind as server CPU core counts rise.
The unexpected software beat adds $1 incremental EPS, which at a 20x multiple, contributes $20 to the stock price.
Non-AI semis showed bookings exceeding $6B, suggesting a cyclical recovery.
Management maintained operating margins despite gross margin pressure, supporting the idea that AI mix does not necessarily destroy EPS leverage.
AVGO: Numbers missed very high expectations; Hock Tan with a sub-par call performance only reiterated F27 and spoke of GOOGL competition; but F28 commentary on callbacks positive.
The numbers missed very high expectations, and the CEO's sub-par call performance, coupled with comments about Google competition, introduced uncertainty.
Fast money leaned short due to the Anthropic revenue recognition issue and the shift from racks and chips to just chips, which impacted the Q3 guide.
The F27 outlook of >$100B was merely reiterated, rather than raised, which disappointed buyside expectations.
The FQ3 gross margin guide of ~74% was below Street expectations, and management stated that AI mix is dilutive to corporate gross margin.
Bears worry that as AI semis become a much larger share of revenue, the company's premium margins could be structurally challenged, even if management attributes the decline to mix.
The comment about Google's source diversity fuels the bear case regarding competition from MediaTek and MRVL, despite Google's use of multiple SKUs.
Bears question why the company deserves a higher multiple than NVDA, as it trades at 25x Street FY27 earnings compared to NVDA's mid-teens multiple.