Alphabet (GOOGL) is one of the hardest companies to read because it is so large and does so many things. Each business has to be judged in context.
Search still the core
Over 50% of Alphabet's (GOOGL) revenue still comes from search. The worry, which I shared, is that AI would pull market share away from Google because AI gives better answers. That has not happened. Search traffic is holding. Sentiment data shows higher search volume and higher price-per-click. On a pure search-volume and consumer-demand measure, GOOGL sits at plus one year-over-year - not amazing, but steady. Likefolio tracks nearly every Google search and knows which brands and companies people look for; total search volume is not falling and is not being taken by OpenAI, Anthropic, or others. Even tech-savvy people still go to Google as much as before. Gemini is not trading blows with Anthropic and OpenAI the way those two keep leveling up against each other, but Gemini is built into Google, so searchers now get good AI answers inside Google. That defense of search traffic means the long-predicted death of Google search - talked about for decades, not years - will either take much longer than expected or never arrive. Search may be safer than many, including me a few months ago, believed.
Streaming and YouTube
Across streaming, search traffic and social-media chatter are declining industry-wide, with one exception: YouTube TV, growing about plus 4% while rivals slip. This is year-over-year data with seasonality and swings, but it is good for a GOOGL investor to see YouTube holding. YouTube is roughly 10% of revenue - a small piece, doing well.
Cloud and the TPU story
Google Cloud is about 15% of revenue, carries big margins, and is where the growth and excitement sit. Cloud revenue rose 82% year-over-year and is expected to grow more; income is up 3x. Margins run around 35%. Much of that margin comes from Alphabet (GOOGL) making its own chips.
Those chips are tensor processing units (TPUs), Google's version of the Nvidia (NVDA) chip - which I believe Google built first. TPUs keep costs down and lift margins, and now Alphabet is renting them out, opening a new revenue stream. Anthropic agreed to use about a million. OpenAI is rumored to be buying a large number.
This repeats what Amazon (AMZN) did with AWS: AWS was born when Amazon built hosting internally to cut costs, found spare capacity, and sold it. Google is doing the same - internal tool turned into a product and a revenue stream. That path is attractive to investors because it clearly works.
Among comparable demand numbers, Amazon (AMZN) is declining while Microsoft (MSFT) is growing fast. These are year-over-year reads against each company's own past. Amazon faces hard comparisons against a strong year earlier and is still growing, though its growth rate may be slowing. Microsoft, once a laggard in the group, sits at plus 14 - take that with caution.
Capex is the test
For several quarters the story has been capex. Alphabet (GOOGL) is talking about spending $200 billion in 2026 and raising that in 2027, which makes many nervous. Investors want capex to turn into income. If Google Cloud keeps growing, margins hold near 35%, and TPU sales and rentals build, that validates the capex - the boogeyman of recent earnings reports. Microsoft (MSFT) has already shown this works. If GOOGL posts strong growth and attractive margins, scrutiny on capex should drop sharply, because the spending will have visibly produced income, making the case to keep investing.
Valuation and the stock
The stock is $90 off its recent high. Large conglomerates tend to get judged by their worst segment - there is always something wrong when you own a big company - and stagnant streaming growth has weighed on it even after it cleared its legal problems and rallied. GOOGL trades at a PE in the low-to-mid 20s with 25% revenue growth. Earnings per share is growing only 5% because of all the spending. It becomes more attractive, especially if it holds search dominance against AI. It is a tortoise-and-hare story: Alphabet is not flashy about Gemini, with no weekly releases, yet the company is doing great things.
Nvidia angle
On NVDA, TPUs are still a very small number next to Nvidia. Nvidia holds 90% of the market with 75% margins. Those 75% margins make it tempting for a competitor to step in and do it cheaper, which Google can. It is a tiny competitor and a long way from real competition, but for Nvidia (NVDA) holders it is uncomfortable to watch another AMD-style rival appear when a monopoly was more comfortable. Owning both companies, the question of who wins stays open.
The verdict
Heading into this report, capex should be justified through higher revenue, Google Cloud's success, and sustained margins, relieving investors and letting them look at the PE with search holding and cloud ramping. Last quarter investors disliked the capex, so the question is whether the tune has changed. The near-term earning score is plus three - basically neutral. The long-term score is 87 out of 100. Alphabet (GOOGL) sits in the "infinite hold" bucket: buy it, lock it away, never touch it, and expect it to be great in 10, 20, or 30 years. Neutral on this call, very bullish long term - a home run over time.
