Seagate @JPM 54th Annual TMT: Key Takeaways & Full Transcript
Date: May 2026 (TMC 2026, Boston)
Format: Fireside chat, ~34 minutes
Source: J.P. Morgan TMC 2026 conference audio webcast
Participants
Samik Chatterjee: Hardware and Networking Analyst, J.P. Morgan (moderator)
Dr. Dave Mosley: Chief Executive Officer, Seagate Technology
Mosley has been at Seagate for nearly three decades and has been CEO since 2017.
Key Takeaways
The cycle framing has changed. Mosley refuses to call this a normal HDD cycle. The 2012 unit peak (66 million drives in a single quarter) was a low-capacity client/notebook business with idle drives. Today’s installed base is fewer units but “chock full,” running at 100% utilization in data centers. He is selling a structurally different industry, not a cycle to fade.
Build-to-order has converted the franchise into a visibility business. Recording-head wafer lead times are now >9 months; drive build adds another quarter on top. BTO is locked in for 4-5 quarters but customers are signaling demand 2-3-4 years out. As contracts roll, re-up pricing is moving up, not down.
HAMR / Mosaic is in real volume. Millions of HAMR drives shipped (not “one or two million” - “many more than that”). Mosaic 3 is qualified at all planned CSPs. Mosaic 4 (4 TB/platter) is qualified at two large CSPs and Mosley is confident on the path to 5 TB/platter. He referenced 40 TB and 50 TB drives by end of calendar 2027.
Mosaic 4 qualified at “PMR-equivalent” speed. The painful multi-year Mosaic 3 qualification at the first hyperscaler appears to have been a one-time learning-curve event. Mosaic 4 ran through two large CSPs at traditional qualification timelines. Expectation set: Mosaic 5 and 6 should look similar.
Capacity expansion is happening through technology, not floor space. Seagate is targeting mid-20s percent exabyte growth and explicitly not adding new factories. The lever is platter density - 3 TB to 4 TB to 5 TB - which keeps componentry roughly flat per drive. This is the cost-leverage engine behind incremental margins.
SSDs are not the bear case anymore. Mosley argues the data-tier in hyperscale is “stuck on hard drives for a long way into the future.” NAND price increases have made hyperscalers ask HDDs for more, not less.
Margin structure has stepped up. Gross margins are progressing toward 50% (vs. 20-25% pre-COVID); incremental margins have been ~70% and Mosley is selling this as sustainable, not cyclical-peak. 1,400 bps of operating margin expansion YoY in the March quarter against flat OPEX.
Let’s dig in…
