I know many here are tracking Intels progress on 10nm. Here are the 10nm comments extracted from their latest earnings conference call.
Our ecosystem partners have already received Icelake server samples. We are making good progress on Icelake server and are now planning to start production wafers in the first half of 2020 with the volume ramp in the second half of the year. Both yield and defect density are ahead of schedule for our 10-nanometer data center products.
We also launched our new 10th Gen Core product family codenamed Ice Lake which integrates Wi-Fi 6, Gen11 graphics and AI acceleration.
Our process technology road map continues to improve and we're making excellent progress on 10-nanometer. We began shipping Ice Lake clients in the second quarter supporting systems on the shelf for the holiday selling season and expect to ship Agilex, our first 10-nanometer FPGA later this year.
We now have two factories in full production on 10-nanometer. We are also on track to launch 7-nanometer in 2021. With a roughly 2x improvement in density over 10-nanometer, our 7-nanometer process, which will be comparable to competitors' 5-nanometer nodes, and will put us on pace with historical Moore's Law scaling.
We expect Q4 gross margin to be down 3 points to 3.5 points sequentially, as we continue to ramp 10-nanometer.
The increased savings are being offset by higher spending on 10-nanometer and 7-nanometer processes and product R&D.
From the Q&A session (less prepared remarks)
We did say that we think we'll bottom out at 57% in 2021 (referring to gross margin). We didn't guide specifically for 2020, but implied it would be closer to the 60% range. And I would say, one of the things that's going to help us -- two things that are probably going to help us as we think about 2019 to 2020. One is we're going to see more of the benefit of moving up the yield curve in 10-nanometer, which is pretty painful now and you're seeing that really in the fourth quarter gross margin.
To George's point the -- while we're not big fans of growing inventory we feel great about qualifying the Icelake product in the second quarter. And that qualification is the single biggest reason for the step function and inventory from Q1 to Q2. So we knew if we executed on our plans of the Icelake clients qualifications we'd have a step-function inventory and we feel good about getting that qualification done and behind us. The implications of that is the higher balance in Q2 versus Q1.
They had enough ice lake product built before the product qualified that it shows up in excess inventory now. Ice Lake product cost is very high, and will be a drag on profitability. Gross margins have dropped from the low 60's to the high 50's. It looks like 14nm product cost is better than 10nm through much of next year.
