Almost all of this busy week's content has been sent. In case you missed any of it:
I will be sending out more detailed reviews of Microsoft and Starbucks earnings as soon as I can. Brief coverage is provided below.
Next week, I will be providing earnings coverage on:
Mercado Libre
Palantir
Axon
Uber
Shopify
AMD
AppLovin
Datadog
SpaceX
The Trade Desk
1. Starbucks (SBUX) – Brief Earnings Snapshot
The full review is coming soon. For now:
a. Demand
Beat 5.7% comp store sales growth with 7.9% growth. Really good.
Traffic and ticket size were both better than expected.
Beat revenue estimate by 2.2%.
Store count was roughly in line with expectations.
More Y/Y Comp Store Sales (CSS) Growth Data:
Within global CSS growth.
Transactions rose by 4.2% compared to 3.8% last quarter and 1.0% the quarter before.
Ticket size rose by 3.5% compared to 2.3% growth last quarter and 1.0% the quarter before.
Within North American CSS growth:
Transactions rose by 4.5% compared to 4.4% growth last quarter and 3.0% the quarter before.
Ticket size rose by 3.5% compared to 2.7% growth last quarter and 2.0% the quarter before.
USA CSS growth was 7.9% thanks to 4.2% transaction growth and 3.6% ticket size growth.
Within International CSS growth.
Transactions rose by 3.1% compared to 2.1% growth last quarter and 3.0% the quarter before.
Ticket size rose by 2.6% compared to 0.5% growth last quarter and 2.0% the quarter before.
b. Profits
Beat EBIT estimate by 23%.
Beat $0.65 EPS estimate by $0.20.
c. Balance Sheet
$3.6B cash & equivalents.
$2B investments.
$13B debt.
0.4% Y/Y dilution.
d. Guidance & Valuation
Raised 5% global & U.S. comp sales growth guide to 6% & 6%+.
Raised revenue growth guide from flat to slight growth, meeting estimate.
Raised EBIT margin guidance.
Raised $2.35 EPS guide to $2.60, beating estimate by $0.21.
SBUX trades for 36x EPS. EPS is expected to grow by 20% in each of the next 3 years.
2. Microsoft (MSFT) – Brief Earnings Snapshot
a. Demand
Beat revenue estimate by 2.6% & beat guidance by 3.1%.
More Personal Compute (PC) revenue beat estimates by 5.5%;
Intelligent Cloud (IC) revenue beat estimates by 3%. Satya talked a lot about open models during the call. We'll get to all of that in the review.
Productivity & Business Processes (PBP) revenue beat estimates by 1.6%.
Unearned revenue beat estimates by 2%.
Seat growth remains stable at 6%. No issues from AI disruption here.
90%+ of its cloud revenue is outside of frontier model companies.
Microsoft Foundry (AI developer platform) customers "more than doubled" Y/Y to 100,000+.
b. Profits
Beat 64% Cloud GPM estimate & beat identical guidance by 1 point each.
Beat 66.5% GPM estimate by 70 basis points (bps; 1 basis point = 0.01%).
Beat EBIT estimate by 4.1% & beat guide by 5.7%.
Beat $4.25 EPS estimate by $0.56.
Beat $15B FCF estimate by $4.6B. CapEx was 1.5% higher than expected.
c. Balance Sheet
$77B cash & equivalents.
$40B debt.
-0.3% Y/Y dilution.
d. Guidance & Valuation
Q1 2027 revenue guidance was 0.8% ahead with EBIT 3% ahead.
Lowered $190B calendar year 2026 CapEx guide to $15B due to shifting some of the expenses to OpEx. No change in build-out plans.
Reiterated Azure growth acceleration for 2H of calendar year 2026.
Fiscal year (FY) 2027 CapEx is expected to grow Y/Y.
Reiterated 10%+ revenue growth with modest mar leverage for FY 2027.
Committed to positive FCF for FY 2027.
Microsoft trades for 24x forward EPS. EPS is expected to compound at a 16% clip for the next two years.
3. AI Infrastructure Developments
This week, there was a Chinese state-backed initiative announced that combines various pieces of that nation's supply chain to build Deep Ultraviolet (DUV) Lithography machines. These are more basic versions of Extreme Ultraviolet (EUV) Lithography machines. EUV machines are used for manufacturing best-in-class GPUs, custom accelerators and high bandwidth memory (HBM) chips dominating the AI buildout. The goal of the project is to move from 5 functional machines this year to 20 next year.
These developments weighed heavily on ASML, which is the EUV/DUV industry king. They're banned from selling EUV machines to China, so they rely pretty heavily on DUV demand for their overall growth. Specifically, China is estimated to be about 15%-20% of total revenue for 2026, so this is material for them. The company is deeply capacity constrained and will find other demand for these machines if they need to (China will still keep buying these from them). And while that may happen in some spots here and there... I don't think they'll need to. Again, the China joint initiative pooling together various supply chain players is expected to generate 25 total DUV machines over the next 2 years. ASML shipped nearly 300 DUV machines last year alone, making this seem even less concerning for them. And furthermore, ASML's DUV machines are expected to be years ahead of what China first debuts. I get that the infrastructure sector reacted to this piece of news today. People are grappling with how easy ASML's dominant positioning is to rival and whether or not this supply changes things. To me, after contemplating this news, its moat and its demand are very safe.
These machines also could eventually make China's domestic supply chain stronger for more in-country memory and AI accelerator production. That could also limit demand from U.S. vendors, which is why some memory and GPU names sold off. Importantly in memory, Micron's heavy skew towards Dynamic Random-Access Memory (DRAM) and leading high bandwidth memory (HBM) chips is seen as a lot more defensible than SNDK's flash and Not AND (NAND) storage businesses. You need EUV to make best-in-class HBM chips. You do not need EUV to make flash and NAND storage. For one more relevant reminder, Nvidia didn't recognize any H200 revenue in China last quarter and didn't include it in forward guidance. AMD's China business was "not material" in Q1 (was worth $390M in revenue in Q4 2025).
4. Headlines and Macro
The Fed left rates unchanged with 3 members voting to hike and 9 voting to leave rates unchanged. The statement didn't materially change and the Fed seems to be in wait and see mode before making any big rate decisions. Economic expansion was called solid, inflation remains elevated largely due to the war, and employment conditions are healthy and stable.
This week, OpenAI slashed GPT-5 Luna model pricing by 80%. More signs of model pricing wars, which tends to happen when competing products are becoming more similar. Think that favors the public big boys (big 3 cloud providers + Meta) with other businesses and less reliance on directly making money from inference.
Fortinet earnings were a notable standout. Revenue was more than 8% ahead. $0.90 in EPS beat $0.75 expectations. Its new full year revenue guidance is nearly 4% above consensus. Full year EPS guidance well ahead too.
