Record revenues of $672M. . but gross margins still under Orton's magic 30% at 28.2%
http://ir.ati.com/phoenix.zhtml?c=105421&p=irol-newsArticle&ID=836826&highlight=
CC in a few minutes, notes to follow.
I dislike the morning ones, because I end up doing it real-time and miss some stuff because I don't have a recording to go back and forth on.
Having said that, usual disclaimer --What I found interesting, as best I heard it:
Orton
"Exceeded revenue guidance. . ."
10% higher than previous record.
chipsets 25% of total sales.
Imageon love, oh yes.
50,000 units in first 6 weeks of x1900. (I think I heard that right, will check later).
"honestly believe" X1900 "shocked our competition". Huh.
Crowley
Company record comes in "seasonally down" quarter.
Strong ASP growth in desktop discrete.
Notebook discrete ~20%
Desktop/notebook 60/40.
Royalties $10M. NRE less than $10M.
Gross margins down a bit.
Finished clearing out desktop old write-down stuff in quarter.
Chipset growth responsible for gross margins decline.
Purchasing substrates directly now; change has helped supply, but increased inventory.
Southbridge shortages prevented some sales (their own tho, it sounds like), and increased northbridge inventory.
Q3 expected gross margins to get to about 30%
Orton
"Extremely excited" about Imageon acceptance. "At cusp" of major inflection point in handhelds.
Mobile TV "set to explode".
From 10s of thousands to millions in just 4 quarters on chipset growth.
Express 3200 love from the community. Quotes provided. Going to drive the margins recovery on the chipset side.
Ode to X1K and IQ here. Love HQV benchmark.
X1600 "incredible demand". (Hmm. Disconnect there from conventional wisdom!)
"Will continue to enhance stack" in Q3. (RV570/R590 hint, I'd guess)
Q/A.
Margins? A: Chipsets pulling us down. RS6xx needed for complete fix. 34-38% *excluding* chipsets today.
Excess inventory in market? A: Europe some excess notebooks in channel, but short-lived.
Revenue growth by segment? A: Some limit on chipset growth by capacity. Handhelds growing. Desktop growing but Q3 is a seasonal down quarter. DTV "extremely robust".
OEMs given oops last year? A: Notebooks 2/3s and up. Desktop lined up well for Spring. Stay tuned for OEM announcements.
Handsets design wins? Qualcomm? A: Qualcomm 7000 sampling now. customer ramps in late summer to Fall (Finally!). Design wins with LG. Platform announcements coming. Spring/Summer.
Desktop market share? A: We believe in CY Q1 we gained share. Wait for Mercury. Believe we'll gain in Q2 as well. We're only high-volume high-end shipping. Expect to stay in 65-75% notebook. Integrated growing rapidly on notebook
Royalties progression? A: Consoles seasonal. Difficult to predict. Expect Q3 flat with Q2.
Pricing environment for desktop, and die size issues? A: (Sounded amused.) Fall of 2002 same dynamic with "focus" on die size. Not as important at high-end as performance and value (which is basically the same thing Jen-Hsun told Wavey re G70 last year!). Mid and value are similar cost structure to competition. Advantage in time to market on 90nm. "Yields phenomenal" and better because they are further in their cycle than NV. Pricing looks stable.
ULi? A: One large customer uses ULi, Asus, and they have good supply. "We don't see it as an issue at all."
80nm and margins? Writedown done completely? A: "Pretty well done." 80nm production rollout in lowend of the line. Press release from fab partner today. "Easier transition." "Going well".
Vista? Delay impact? A: CY 4Q. Primarily OEM at first anyway. No impact on corporate. On consumer, hard to say. Interested to see if OEM's offer upgrade (for Christmas market, presumably). "Enthusiasts will take the fastest" whenever.
80nm and R6xx or 65nm? A: 80nm is cost reduction. Die shrink rather than performance. Some 80nm mixed with some 65nm in Fall/Winter. R6xx will NOT be 90nm. (Like Wavey said, I too read that as low/mid first to 65nm).
80nm again. A: Don't freak on die size comparisons. Redundancy important. Clear feature performance advantage over competition, including HDR, 1GB support. Cost structure of low/mid very competitive, and that's where it counts.
Gross margins on chipsets to get to 25%? A: Making progress but not quite guiding at that level yet.
Visibility for rest of year? A: Macro feel is very positive rest of the year.
More Vista. A: Higher demand for visual capabilities, dx9 floor. Focused on bringing low power notebook Vista stuff. Integrated is Vista-ready.
Substrates: Geo missed it. Solly.
Southbridge? A: Order n&s simultaneous. Lost 1.5 weeks of supply on SB due to fab issue of some sort. That was the constraint. Always sell in pairs (which doesn't sound right --people are buying SB and using ULi anyway?)
DX10 transition. A: Next major platform transition, first since 2002. Big deal. Focused for several years. Think they have clear advantage over NV architecturally, including Unified (implication think NV will not). Advanced Rendering Model, first time in history "fine grained context switching". Hussah whatsa?
AR growing disproportionally A: No big thang. Just a timing issue.
Southbridge shortage nothing to do with ULi? A: No. 180nm capacity is constrained, so tight on southbridge anyway and Fab problem hit. 1.5 weeks is a million units. NB is different process (huh!).
Handheld visibility? A: Market moving to 3G, not there yet. We have media-coprocessor, it's indifferent to whether 2G or 3G is on the other end.
http://ir.ati.com/phoenix.zhtml?c=105421&p=irol-newsArticle&ID=836826&highlight=
CC in a few minutes, notes to follow.
I dislike the morning ones, because I end up doing it real-time and miss some stuff because I don't have a recording to go back and forth on.
Having said that, usual disclaimer --What I found interesting, as best I heard it:
Orton
"Exceeded revenue guidance. . ."
10% higher than previous record.
chipsets 25% of total sales.
Imageon love, oh yes.
50,000 units in first 6 weeks of x1900. (I think I heard that right, will check later).
"honestly believe" X1900 "shocked our competition". Huh.
Crowley
Company record comes in "seasonally down" quarter.
Strong ASP growth in desktop discrete.
Notebook discrete ~20%
Desktop/notebook 60/40.
Royalties $10M. NRE less than $10M.
Gross margins down a bit.
Finished clearing out desktop old write-down stuff in quarter.
Chipset growth responsible for gross margins decline.
Purchasing substrates directly now; change has helped supply, but increased inventory.
Southbridge shortages prevented some sales (their own tho, it sounds like), and increased northbridge inventory.
Q3 expected gross margins to get to about 30%
Orton
"Extremely excited" about Imageon acceptance. "At cusp" of major inflection point in handhelds.
Mobile TV "set to explode".
From 10s of thousands to millions in just 4 quarters on chipset growth.
Express 3200 love from the community. Quotes provided. Going to drive the margins recovery on the chipset side.
Ode to X1K and IQ here. Love HQV benchmark.
X1600 "incredible demand". (Hmm. Disconnect there from conventional wisdom!)
"Will continue to enhance stack" in Q3. (RV570/R590 hint, I'd guess)
Q/A.
Margins? A: Chipsets pulling us down. RS6xx needed for complete fix. 34-38% *excluding* chipsets today.
Excess inventory in market? A: Europe some excess notebooks in channel, but short-lived.
Revenue growth by segment? A: Some limit on chipset growth by capacity. Handhelds growing. Desktop growing but Q3 is a seasonal down quarter. DTV "extremely robust".
OEMs given oops last year? A: Notebooks 2/3s and up. Desktop lined up well for Spring. Stay tuned for OEM announcements.
Handsets design wins? Qualcomm? A: Qualcomm 7000 sampling now. customer ramps in late summer to Fall (Finally!). Design wins with LG. Platform announcements coming. Spring/Summer.
Desktop market share? A: We believe in CY Q1 we gained share. Wait for Mercury. Believe we'll gain in Q2 as well. We're only high-volume high-end shipping. Expect to stay in 65-75% notebook. Integrated growing rapidly on notebook
Royalties progression? A: Consoles seasonal. Difficult to predict. Expect Q3 flat with Q2.
Pricing environment for desktop, and die size issues? A: (Sounded amused.) Fall of 2002 same dynamic with "focus" on die size. Not as important at high-end as performance and value (which is basically the same thing Jen-Hsun told Wavey re G70 last year!). Mid and value are similar cost structure to competition. Advantage in time to market on 90nm. "Yields phenomenal" and better because they are further in their cycle than NV. Pricing looks stable.
ULi? A: One large customer uses ULi, Asus, and they have good supply. "We don't see it as an issue at all."
80nm and margins? Writedown done completely? A: "Pretty well done." 80nm production rollout in lowend of the line. Press release from fab partner today. "Easier transition." "Going well".
Vista? Delay impact? A: CY 4Q. Primarily OEM at first anyway. No impact on corporate. On consumer, hard to say. Interested to see if OEM's offer upgrade (for Christmas market, presumably). "Enthusiasts will take the fastest" whenever.
80nm and R6xx or 65nm? A: 80nm is cost reduction. Die shrink rather than performance. Some 80nm mixed with some 65nm in Fall/Winter. R6xx will NOT be 90nm. (Like Wavey said, I too read that as low/mid first to 65nm).
80nm again. A: Don't freak on die size comparisons. Redundancy important. Clear feature performance advantage over competition, including HDR, 1GB support. Cost structure of low/mid very competitive, and that's where it counts.
Gross margins on chipsets to get to 25%? A: Making progress but not quite guiding at that level yet.
Visibility for rest of year? A: Macro feel is very positive rest of the year.
More Vista. A: Higher demand for visual capabilities, dx9 floor. Focused on bringing low power notebook Vista stuff. Integrated is Vista-ready.
Substrates: Geo missed it. Solly.
Southbridge? A: Order n&s simultaneous. Lost 1.5 weeks of supply on SB due to fab issue of some sort. That was the constraint. Always sell in pairs (which doesn't sound right --people are buying SB and using ULi anyway?)
DX10 transition. A: Next major platform transition, first since 2002. Big deal. Focused for several years. Think they have clear advantage over NV architecturally, including Unified (implication think NV will not). Advanced Rendering Model, first time in history "fine grained context switching". Hussah whatsa?
AR growing disproportionally A: No big thang. Just a timing issue.
Southbridge shortage nothing to do with ULi? A: No. 180nm capacity is constrained, so tight on southbridge anyway and Fab problem hit. 1.5 weeks is a million units. NB is different process (huh!).
Handheld visibility? A: Market moving to 3G, not there yet. We have media-coprocessor, it's indifferent to whether 2G or 3G is on the other end.
Last edited by a moderator:
