Ananda Baruah

Charles gave qualitative mix commentary and a directional 'consistently grow' answer but no specific FY27 gross margin figure.

David deflected, saying only that an update would come shortly and that nothing more could be shared on the call.

David confirmed no additional company employees involved beyond those named. No restatement expected. 10-Q filing being prepared. Independent investigation ongoing with Munger Tolles & Olson and AlixPartners.

Charles reported no negative customer sentiment from the indictment based on direct communications. Customer base continues to grow with more large and mid-sized customers.

Charles confirmed margin improvement expected quarter-over-quarter from maturing GP300 production, reduced expedite costs, improving tariff impact, and growing DCBBS contribution.

Charles confirmed economies of scale will drive both gross and operating margin improvement. DCBBS software and infrastructure services add positive margin mix.

Charles confirmed the $36B guide is conservative. First time achieving $10B+ quarterly revenue and capacity is now scaled globally to support higher volumes.

Charles expressed confidence in continued growth through the calendar year, citing scaled capacity and market leadership in both technology and market share.

David confirmed Q2 is expected to be the gross margin low point for FY2026, attributing it to the unprecedented revenue doubling and new platform ramp-up costs.

Charles acknowledged ongoing product transition dynamics with multiple NVIDIA product lines but positioned SMCI's breadth as an advantage rather than directly addressing whether cycles have normalized.

Charles did not specify customer types or domains. Gave a generic response about strong CSP demand and improved cash flow readiness.

Charles expressed hope for sequential growth trajectory similar to Hopper cycle but stopped short of firm commitment.

David referenced historical H100 cycle margin improvement to 18.8% as precedent. Still targeting 14-17% but acknowledged competition will influence Blackwell margins.

Charles framed the $40B target as conservative without directly confirming the quarterly math or customer composition.

David noted $4B in working capital raised over 8-9 months. Company does not believe it will have impediments to raising working capital.

David confirmed they are still targeting margin improvement and guiding conservatively. Charles added competition is putting some pressure.

George said 1P and marketplace storage continue to grow above the cloud portfolio; Wissam cited the 80-85% target and mix tailwind.

George outlined two agentic AI opportunities: AIDE (AI Data Engine) addresses data preparation, guardrails, and enforcement requirements; AFX and caching technologies optimize recursive data access patterns used by reasoning agents. More innovation in pipeline.

George cited semiconductor equipment manufacturer unifying millions of documents for AI, content production house using AI for digital experiences, multimodal applications growing, and video analytics in weather forecasting. No dramatic inflection yet but multimodal demand growing with more complex model support.

George confirmed customers want areal density improvements. NetApp qualifies different media sizes and density points following customer demand. Performance, cost, and density are key drivers of new silicon adoption.

George confirmed sovereign cloud providers expanding from enterprise cloud to AI-as-a-service, with NetApp NVIDIA cloud partner certification enabling multi-tenancy and model training for sovereign models. Still believes enterprise inferencing is primary TAM but being opportunistic in other use cases.

George highlighted two opportunities: data organization tools (tagging, vectorization, graphs, RAG readiness) to monetize massive unstructured data install base, and reasoning model optimization capabilities that reduce repeated data source access. More details coming at Insight user conference.

George cited unique hybrid cloud positioning to bring enterprise data to hyperscaler platforms. Shared gaming company AI win using on-prem plus cloud. Noted vectorization and multimodal datasets driving data growth, though early to quantify. Proof of concepts showing clear data growth indication.

Mike reaffirmed Investor Day targets for mid-to-upper single-digit growth and margins through FY27.

George explained midrange brings new logos while high-end ASA expands wallet share with existing customers.

Irving framed Asia and open-source frontier models positively, arguing more models drive compounding storage demand.

Kris confirmed ~10% cost/TB reduction as baseline, with potential to accelerate if roadmaps advance faster. Driven by continued innovation, higher capacity drives, UltraSMR upshift, and manufacturing efficiencies.

Irving deferred to Innovation Day the following week. Did not provide specific capacity ceiling for ePMR/UltraSMR before HAMR crossover.

Irving laid out roadmap: current 26/32TB ePMR → next-gen 28/36TB ePMR (qual H1 2026) → HAMR 38/44TB (qual H2 2026, ramp H1 2027). HAMR crossover at 4TB per platter.

Irving confirmed one large cloud provider and significant Storage-as-a-Service providers using platforms. Neo clouds focused on compute first; storage follows as seen in mainstream cloud AI adoption.

Irving acknowledged HAMR requires rack-level and software changes similar to Ultra SMR adoption. Sophisticated customers are familiar with needed architectural adjustments.

Don confirmed faster-than-expected 11-disc platform ramp with high quality, reliability, and yields was key driver of gross margin beating guidance.

David maintained stable Flash/HDD mix view, noting AI training favors Flash but data generation flows back to HDD. Both businesses benefit from AI demand.

Asked about the areal-density roadmap ahead of HAMR; management detailed the bridge on the current platform before HAMR is introduced, direct.

Dave pointed to unstructured video and sensor data across enterprises and the edge, not only hyperscalers, as the applications driving storage demand today.

Mosley confirmed 20TB HAMR on Mozaic 4 is technically possible but high-end demand is too strong to divert volume. Romano agreed demand too strong in public cloud. Mozaic 5 may enable addressing lower-capacity markets later.

Mosley pushed back on the warm-tier-becoming-NAND narrative, noting HDDs work 24/7 in optimized streaming workloads and the architectural tier for mass data storage will remain constant for the next decade.

Mosley noted customer pull for higher capacity points is much stronger than prior PMR transitions due to compelling TCO benefits. Human creativity driving video demand also fueling the pull.

Mosley identified three video demand types: existing video properties, unstructured data processing, and emerging AI-generated video content creation.

Mosley acknowledged difficulty predicting GenAI data center storage but noted disaggregated storage as an interesting trend that could benefit HDDs.

Mosley confirmed HAMR is currently accretive to gross margins even at today's elevated PMR margin levels, and Seagate will ship as much HAMR as possible while navigating customer qualification timelines.

Asked what adding capacity would look like if demand outstrips HAMR/areal-density gains; Dave explained that pushing areal density frees components to build more drives efficiently.

CEO confirmed Greensboro capacity is incremental to existing targets, representing over $5 billion of potential additional revenue. Fab expected to come online in early 2028, roughly 6 quarters away.

Michael sees OCS as multi-hundred-million-dollar opportunity coming sooner rather than later. No specific timeframe but increasingly clear on the trajectory.

Michael expects cloud modules at roughly 30% GM at asymptote, a drag on corporate margins. Wajid sized it at 15-20% of revenue at $600M. Components and DCI will be larger contributors.

Alan said first new customer reaches full motion by end of CY2025 with second product. Third customer starts fiscal Q4 and hits run rate by Sep-Oct. No conflict between external EML sales and internal use, won't reduce external supply.

Alan emphasized U.S. headquarters, Thailand manufacturing, laser chips from U.S./U.K./Japan. Strategy unchanged regardless of administration. Resonates with customers wanting supply chain resilience.

Alan acknowledged pricing leverage in constrained EML market, noting need to fund capacity additions. Discussed on portfolio basis rather than specific figures.

PS head said customers are naturally choosing higher-configuration products to run AI at the edge, driven by token costs, cybersecurity, privacy and latency needs, and to govern agents securely as they deploy them more broadly.

PS head confirmed the same procurement playbook will apply for FY27, including qualifying new suppliers and accelerating component quality checks across the portfolio.

Karen clarified 35% is of PC BOM specifically. One-third of PS margins come from businesses not reliant on memory including higher-margin attached business, Workforce Solutions, and hybrid systems.

Bruce outlined search committee established ~3 weeks ago. Seeking proven shareholder value track record, ability to navigate complex environments, and global multi-segment business experience. Board taking broad view including candidates outside PC/print industry.

Enrique outlined consumer continuing to decline, transactional office declining, contractual office performing better (opportunity for share gain), industrial growth continuing.