LikeFolio Weekly Roundup: $725B AI Surge

Four of the most important companies on Earth just made one thing clear: the AI race is accelerating—fast.

In a single round of earnings, Amazon (AMZN), Google (GOOGL), Microsoft (MSFT), and Meta Platforms (META) pushed projected 2026 AI capex from roughly $670 billion to $725 billion.

That’s a full-scale buildout of the next economic backbone.

And this week’s results show exactly where that money is going, and why it matters now.

Infinite Hold Updates

Amazon’s (AMZN) AI Engine Hits Another Gear 

Amazon just put up one of its strongest quarters ever.

Revenue jumped 17% to $181.5 billion. Operating margins hit a record 13.1%.  

But the number that matters most sits inside AWS. 

Cloud revenue surged 28% – with a $2 billion sequential jump, the biggest Q4-to-Q1 increase in company history.  

This is the AI engine we’ve been tracking. 

AWS is now running at a $150 billion annual pace… and already pulling in over $15 billion from AI services alone.  

Even more telling: Amazon’s backlog hit $364 billion – and that doesn’t include a new $100+ billion deal with Anthropic. That’s locked-in demand. 

At the same time, Amazon is becoming a major chip player. Its custom silicon business is already at a $20 billion run rate, growing nearly 40% quarter over quarter.  

Translation: Amazon isn’t just selling AI infrastructure. It’s building the stack from the ground up. 

Yes, costs are rising – from chips to satellites to logistics. That will pressure near-term cash flow.  

But we’ve seen this playbook before. Amazon spends big when demand is clear. Then it scales into it. 

For long-term holders, this report didn’t crack the story. It reinforced it. 

Amazon just showed why we own it – and why we’re not letting go. 

Google’s (GOOGL) AI Stack Hits Escape Velocity 

Google also opened at an all-time high on Thursday as revenue jumped 22% to $109.9 billion. Profits surged even faster, up 81%. But the real signal sits underneath those headlines. 

Cloud revenue exploded 63%… and the backlog quietly doubled to $462 billion. 

That’s committed future demand – and more than half of it converts within the next two years. 

This is what we’ve been tracking from the start. 

Google owns the full stack – chips, models, data, and distribution.  Now that edge is turning into signed deals. 

GenAI cloud revenue grew nearly 800% year over year. Enterprise adoption is accelerating fast, with Gemini usage up 40% in a single quarter.  

At the same time, Search is expanding – not shrinking. AI Overviews and conversational queries are pushing usage to all-time highs.  

There is one constraint: supply. 

Management made it clear they can’t keep up with demand for compute. In plain English – they could be growing even faster if they had more capacity. 

That’s why CapEx is climbing toward $190 billion this year.  

Short term, that pressures margins. 

Long term, it builds the moat. 

We’re not chasing a headline here. We’re watching a platform scale across every layer of AI – with demand already locked in. 

Tesla’s (TSLA) Semi Moves from Concept to Scale 

The first Tesla Semi rolled off the high-volume line this week. And just like that, Tesla enters a new lane. 

Tesla is now pushing into a 500,000-unit global trucking market with a product that changes the math overnight. At $116 oil, fuel is a pain point. The Semi cuts that cost by as much as 70%. For fleet operators, that’s a reason to switch. 

And Tesla is aiming straight at scale – 50,000 units per year. 

Most investors are looking right past this. They’re locked in on robotaxis and robots. Fair enough. Those matter. 

But this is how Tesla builds. One layer at a time. Each one tied to the same core advantage: batteries, software, and real-world data. 

The Semi adds another demand engine. More vehicles. More miles. More data flowing back into the system. 

That feeds everything else. 

And Tesla is scaling that system fast – lifting 2026 capex to $25 billion, building 130,000-GPU training clusters in Texas, and pushing deeper into its own AI hardware. 

These ramps can be slow at first. Then they stack faster than most investors are ready for.