Hey look at that! A set of solid Tech earnings tonight: AMD +9%; ANET +10%; ALAB +3%; DUOL +12%; BKNG -2%…
AMD +9%:
Looks solid. Much better top line, but GM missed. 3Q revenue of $13.38B (+45%) beat Street ($13.11B) and was in line with the low end of the ~$13.35-13.5B bogeys, Data Center beat at $8.55B (+97% y/y, a record +$1.83B q/q) vs. $8.25B Street, and the 4Q guide of ~$17.0B +/- 3% (+27% q/q, +66% y/y) beat Street ($16.29B) and sits at the top of the ~$16.6-17.0B bogeys, at ~55% GM (30bps under Street’s 55.3%, but GP ~$9.35B vs. $9.0B Street and ~$9.1-9.4B bogeys); with opex guided to ~$3.75B (Street $3.99B), that’s ~$2.92 of 4Q EPS by our math vs. $2.65 Street and ~$2.71-2.84 at the bogeys. 3Q GM of 55.94% missed Street, the ~56% guide and the bogeys, and Client missed (-3.7% q/q vs. guided “slight growth”).
Revenue beat Street and was in line with the bogeys at $13.377B, +44.7% y/y (+16.0% q/q), vs. Street $13.11B (+41.8%), the ~$13.0B +/- $300M guide and ~$13.35-13.5B bogeys (+44-46%).
That’s a 2.9% beat on the guide midpoint vs. 3.0% in 2Q, the smallest since 4Q24 (a third straight smaller beat: 7.0%, 4.6%, 3.0%, 2.9%); 12-quarter average 3.3%. It cleared the $13.3B top of the guide by 0.6% (2Q: +0.3%).
99% of the $1.84B q/q revenue add came from Data Center (+$1.83B) vs. 73% in 2Q.
Data Center beat Street at $8.546B, +96.9% y/y (+27.2% q/q), vs. Street $8.25B (+90.0%). Release: “driven by strong demand for AMD EPYC processors and the launch of AMD Instinct MI450 Series GPUs.”
The q/q add of +$1.83B is a record vs. +$0.94B in 2Q and the prior record +$1.10B in 3Q25; DC was 63.9% of revenue vs. 58.2% in 2Q.
The y/y decel from 107% is the comp: 2Q25 DC was the MI308-ban trough ($3.24B) and 3Q25 then grew +34% q/q, so the 4-quarter accel run ends but the 2-yr stack only slips to 119.2 from 121.7.
Beat Street by 3.6% vs. 1.2% in 2Q (12-quarter average 0.8%).
DC op income $2.88B, a 33.7% margin vs. 31.3% in 2Q and 24.7% a year ago.
No server CPU vs. Instinct split in the release. Anthropic, META, MSFT, OpenAI and ORCL (OCI) are “running models and workloads on the first AMD Helios rackscale solutions.”
Client missed at $2.949B, +7.2% y/y (-3.7% q/q), vs. Street $3.08B (+11.9%) and the “slight growth” q/q guide; Gaming beat at $773M (-40.4% y/y, -0.8% q/q) vs. Street $674M and a “strong double-digit decline” q/q guide.
Client slowed for a 6th straight quarter, to 7.2% from 22.5%. That’s not the comp: it eased (3Q25 client grew 46.2% vs. 67.5% for 2Q25) and growth still slowed, so the 2-yr stack fell to 53.4 from 90.0. That’s before the 4Q PC slowdown INTC flagged.
Client and Gaming segment $3.722B (-8.1% y/y, -3.1% q/q) vs. Street $3.75B; op income $573M (15.4% margin vs. 15.2% in 2Q).
Embedded beat at $1.109B (+29.4% y/y, +13.5% q/q) vs. Street ~$1.09B (+27%), a 4th straight acceleration (18.6% in 2Q); op income $453M, a 40.8% margin vs. 39.5%.
GM missed at 55.94% (non-GAAP), ~17bps below Street’s 56.11%, 6bps under the ~56% guide and ~26-46bps short of the ~56.2-56.4% bogeys.
The first miss vs. the guide since 3Q25 and vs. Street since 1Q25 (2Q beat the guide by 24bps). Down 30bps q/q from 56.24% in 2Q, +195bps y/y (53.99%).
GP of $7.48B beat $7.36B Street but missed the ~$7.50-7.61B the revenue and GM bogeys implied. Incremental GM 54% q/q (Street 55%), with MI450 starting to ship (the CFO said in August DC AI runs “slightly below corporate average”).
Non-GAAP opex $3.56B, $90M under the ~$3.65B guide and $131M under $3.69B Street; op income $3.92B / 29.3% OM beat Street $3.67B / 28.0% (~130bps above); 26.8% in 2Q and 24.2% a year ago. Incremental OM y/y 41% (Street ~37%).
EPS of $2.09 (+74% y/y) beat $1.95 Street ($1.93 at the guide by our math). About half the $0.14 beat is GP dollars (+$124M vs. Street, ~$0.065) and half opex ($131M under Street, ~$0.07), by our math. Diluted shares 1,661M, in line with the ~1.66B guide.
GAAP EPS $1.46, incl. a $1.10B ($0.66) non-cash charge from marking the customer warrants to market (driven by AMD’s stock price) and $67M ($0.04) of investment gains.
FCF $554M (4% margin) vs. $1.56B (14%) in 2Q and $1.53B a year ago: inventory rose $3.0B q/q to $11.46B (from $8.47B) and receivables $1.9B, partly offset by payables +$2.3B. Capex $648M. Cash and ST investments $15.8B; total debt $7.08B after $4.7B of new debt; no buybacks and $1.5B of tax withholding on employee equity.
4Q guide at the top of the bogeys: revenue ~$17.0B +/- 3% (+65.5% y/y, +27.1% q/q at the mid) vs. Street $16.29B (+58.6%) and ~$16.6-17.0B bogeys (+62-66%). Guide, verbatim: “approximately $17 billion, plus or minus 3%. The mid-point of the revenue range represents year-over-year growth of approximately 66% and a sequential increase of approximately 27%. Non-GAAP gross margin is expected to be approximately 55%.”
The mid is $710M above Street, $400M above the $16.6B bogey low end and at the $17.0B top. The range is +/- 3% (~$16.49-17.51B) vs. the usual +/- $300M; even the low end is ~$200M above Street.
+27.1% q/q vs. the +3.8% they guided for 4Q25 and +12.7% for 3Q26, roughly double their biggest guide step in the last eight quarters (+13.2%); the bogeys needed +23-27%. Su: “we expect growth to accelerate further in the fourth quarter as we scale Helios and ‘Venice’ production.”
A 3Q-sized 2.9% beat on the $17.0B mid gets 4Q to ~$17.5B vs. Street $16.29B.
GM ~55% vs. 55.3% Street (~30bps below) and ~55-55.5% bogeys (the low end), ~94bps below 3Q’s 55.94%, so GM falls q/q for a second straight quarter on MI450 rack mix. GP ~$9.35B (55% x $17.0B, our math) beat $9.0B Street by ~$340M and sits in the upper half of the ~$9.1-9.4B bogeys.
Profit (deck): opex ~$3.75B ($240M under Street’s $3.99B; +5% q/q vs. revenue +27%), OI&E ~$4M (vs. ~$55M guided for 3Q), 13% tax and ~1.67B shares (1,668M). With the GP guide that’s ~$5.60B of op income, a ~32.9% OM vs. Street $5.06B / 31.1% (~180bps above), and ~$2.92 of 4Q EPS (+91% y/y on $1.53) vs. $2.65 Street and the ~$2.71-2.84 the bogeys implied, by our math. That fits the CFO’s “accelerating revenue and earnings growth” (EPS grew 74% in 3Q).
Implied FY26 revenue ~$52.2B (+50.6%) vs. $51.2B Street
The Dec-Q set hit the top of the bogeys on revenue: revenue at the ~$17.0B top ($710M above Street) and GP ~$9.35B in the upper half of the ~$9.1-9.4B bogeys ($340M above Street), with GM ~55% at the low end of the ~55-55.5% bogeys and 30bps under Street, and the ~$3.75B opex guide takes implied 4Q EPS to ~$2.92 vs. $2.65 Street and ~$2.71-2.84 at the bogeys (our math).
2027: nothing in the release beyond Su’s “confident in our ability to sustain very strong revenue growth and expanding profitability over the coming years.” In August they said 2027 DC would “more than double” (“well over 100%” in Q&A) with server CPU growing more than 70%
ANET +10%: Looks really good. Another big beat and a 4Q guide well above the bogeys — much better despite being a crowded long. Revenue of $3.66B (+58.5%, the fastest growth since the IPO) beat Street ($3.33B) and the ~$3.43-3.48B bogeys, a 4th straight acceleration (from 37.7% in 2Q) and a 10.8% beat on the $3.3B guide (8.4% in 2Q), GM of 64.0% beat Street (63.1%) and the ~63-63.5% bogeys, and the 4Q guide of $3.9-4.0B (+59% at the mid) beat both Street ($3.60B) and the ~$3.68-3.78B bogeys, implying FY26 of ~$13.35B vs. ~$12.85-13.0B bogeys. The one blemish: total deferred revenue fell $26M q/q, the first decline since 2Q23. Will wait for 27 color on the call.