Today’s theme: the big tech companies are ‘backing up the truck’ on AI-infrastructure spend. Google is the latest example this earnings season — ramping AI capex hard even as its free cash flow goes negative for the first time in decades — with Meta, Microsoft and Amazon up next week on their own earnings calls. Meanwhile, Big Tech’s combined on- and off-balance-sheet AI debt and equity now tops $1.65+ trillion. This is the AI Tech Wave at its most capital-intensive. Three events, each with my Take, then my Overall Take. Plus a Gadget AI on Ford running Apple Maps in a new EV, and two questions. Let’s get started.

(1) Google — Ramping AI Capex Even as Free Cash Flow Turns Negative

MP TAKE: Google’s quarter again came in above expectations on all the key businesses and metrics. Even the worry over AI Search being a touch weaker was overdone, given the current success of Google Gemini AI Mode I discussed recently. The company raised its 2026 AI-capex guide to as much as $205 billion, up from $190 billion-plus — even as its free cash flow went negative for the first time in a long while (FT clocks a $6 billion cash burn on the quarter), and it’s bolstering its balance sheet with a recent $80 billion-plus equity raise. This from a company still throwing off $112 billion-plus in profits — no dearth of cash; the spend is simply running ahead of the balance sheet. The demand signal underneath it is real: Bloomberg reports the cloud backlog ballooning to $514 billion. As the WSJ puts it, Google has the muscle to overpower spending worries.

Its approach is different from its frontier-lab AI peers. Google isn’t competing as aggressively on front-of-the-line benchmark models. It’s instead shipping a slate of Gemini models that are fast, efficient and relatively cheap to run at mass-market scale — this week’s new, cheaper Gemini 3.6 Flash, 3.5 Flash-Lite and 3.5 Flash Cyber are the latest. That’s a ‘good enough’ AI strategy the market hasn’t fully absorbed yet. It worked for Microsoft with Internet Explorer browser vs Netscape in the Internet Wave. And likely will work for Google in the AI Tech Wave. In my view, Google remains the best-positioned global consumer-AI firm — along with Apple (Not stock advice).

Sources, in narrative order: FT — Google Burns Through $6 Billion in Cash as AI Spending Climbs Again; Bloomberg — Google Says Cloud Services Backlog Expands to $514 Billion; Bloomberg — Google Boosts 2026 Spending Estimate to as Much as $205 Billion; WSJ — Google Has the Muscle to Overpower Spending Worries; Axios — Google Releases Series of New, Cheaper Gemini AI Models; Google — Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. For longtime readers: ‘Google’s Strong Quarter and Doubling AI Capex’ in AI-RTZ #989; ‘Flashing Green Lights for AI in 2025’ in AI-RTZ #791; and ‘Google under fire again for “breaking the internet”’ in AI-RTZ #1155.

(2) The Rest of Big Tech Is Right Behind — Earnings Calls Next Week

MP TAKE: After Google this week, Amazon, Meta and Microsoft are all scheduled to report their second-quarter earnings next week. I’d expect every one of them to be just as committed to their AI-capex spend as Alphabet was this quarter. No one is ready to ‘blink’ — which is exactly what the market is focused on right now. Each of them is still seeing risk/reward calculations that look favorable for their investors, and markets are leaning in, in a ‘green light’ environment. The New York Times’ DealBook frames it well: Google sets the pace on hefty AI spending — and the rest follow.

Sources, in narrative order: NYTimes DealBook — Google Sets the Pace on Hefty AI Spending. For longtime readers: ‘Big Tech roars “AI” on mid-year earnings calls’ in AI-RTZ #800; and ‘Booms that Rhyme — Internet/Telecom Capex vs AI Tech Wave Capex’ in AI-RTZ.

(3) Global AI Capex at Unprecedented Levels — the $1.65+ Trillion Question

MP TAKE: Cumulative debt at the US tech giants is up roughly 8x in four years, to an estimated $1.65 trillion, as AI capex keeps ramping quarter by quarter. The latest Nikkei Asia study looks at Google, Microsoft, Amazon, Meta and Oracle: the debt on their actual balance sheets comes to about $1.35 trillion — but the off-balance-sheet piece is the eye-opener. Meta’s off-balance-sheet debt alone is pegged at ~$420 billion — about 3x its recorded debt — largely through SPVs (special-purpose vehicles) that PE and VC firms set up so institutional investors from around the world can pour money into AI data centers.

It’s framed as ‘hidden, opaque and scary,’ but these SPVs are well understood institutional investment vehicles — the kind BlackRock and the big PE/VC funds run all the time — not the 2008 CDO shenanigans the comparison implies. If I’d draw one real parallel to ‘08, it isn’t the balance-sheet debt; it’s the NASDAQ index fast-tracking SpaceXAI’s IPO into the index in weeks rather than the usual seasoning — and maybe doing the same for freshly-minted Anthropic/OpenAI IPOs. Investors are currently flashing green lights on this kind of multi-year funding. We’ve seen versions of it in prior tech waves — but never at this scale. Alarmists will, of course, point to the swelling debt. But for now, most of the investors leaning in are calculating the risk-reward as positive.

Sources, in narrative order: Nikkei Asia — Five US Tech Giants’ Hidden Debts Soar to $1.65 Trillion on Opaque AI Funding. For longtime readers: ‘Multi-trillion $$ AI Data Center Compute builds ahead’ in AI-RTZ #529; and ‘Latest report on an AI Bust and the Global Economy’ in AI-RTZ #1134.

MP OVERALL TAKE

Google and the rest of Big Tech are ‘backing up the truck’ on AI-infrastructure capex — into the trillions. The markets are talking about all of it, especially through this week’s and next week’s quarterly results, but they’re still leaning into the spend. My view: the risk/reward analysis favors continuing to ‘step on the gas’ after ‘backing up the truck.’ Even with cumulative industry investment running into the $1.6-trillion-plus range, the enterprise and institutional demand — globally — is there for all the AI-data-center infrastructure being rolled out, right down to the earlier generations of chips (just look at the spot and long-term contract markets).

And here’s the counter-intuitive part that keeps the whole thing relatively intact: the flow of new, more efficient open-source AI models doesn’t shrink demand — via the ‘Jevons Paradox,’ cheaper compute expands and accelerates the global market for AI compute. It sounds backwards, but it’s a fundamental, bottoms-up market trend for now — which is also why barring or slowing the best open-source models, many now from China, would be detrimental to US long-term interests. There will be volatility and ‘air pockets’ — stocks are down today — but the comparison with previous tech ‘bubbles’ — like the internet — is premature. I know a little about that one; I was at the front tables then.

GADGET AI

Ford Puts Apple Maps in the Driver’s Seat — a First

MP TAKE — Gadget AI (take-first, Ep 103+)

Here’s a notable first: Ford is building Apple Maps into a new, ‘skunk works’-driven ~$30,000 EV for next year — at the operating-system level. That’s different from the usual Apple CarPlay or Google Android Auto, which run off your smartphone inside the car; this bakes Apple’s mapping into the vehicle itself. Both Apple and Google have been working for years to get their software deeper into cars, with limited success — GM has moved away from CarPlay entirely, as has Tesla with its full vertical stack. So Ford’s move is particularly notable for Apple, which has spent years as the #2 in maps behind Google Maps and Waze. A sub-$60 billion Ford leaning on a ~$5 trillion Apple underlines the software age we’re in. The car operating-system race worldwide is still in its early innings — with Chinese players like Xiaomi, Geely and others racing ahead too.

Sources, in narrative order: NYTimes — Ford Will Use Apple Maps in New Self-Driving System. For longtime readers: ‘Apple Car, Lessons Learned’ in AI-RTZ.

QUESTIONS

Q1 — What’s Michael’s favorite map system while driving?

Answer: Still partial to Google Waze, particularly for its community-sourced road updates — but I’ll have to give Apple Maps another try on my next road trip soon.

Q2 — What could they do better?

Answer: Truly integrate live AI features with full accuracy on ambient conditions and audio alerts as and when needed — and do on-the-fly detours to the nearest coffee or gas stop on command, without the driver taking hands, eyes or attention off the road. The simple things are still the hard things: we’re getting there, but it’s still easier said than done.

WRAP

Today’s AI-RTZ #1156 — Anthropic and OpenAI Now Against China Open-Source AI — the two US frontier labs lining up against China’s open-source push, ahead of their own mega-AI IPOs, enlisting Washington as the ‘bad cop’ to wall off US businesses from Chinese open-source models. My read: that’s the wrong move — it would block US customers from serving the world with the best software available.

AI Ramblings Daily on AI-RTZ is here to think through AI and reset. Together.

Tomorrow — back with ARD 126 on AI-RTZ 1157.

Thanks for joining us today, AI Curious Folk. Stay tuned.

— MP

Full Source Reading —

For the broader context, see the canonical sources for ARD 125 — in today’s narrative order:

Event 1 — Google Ramps AI Capex at Negative Free Cash Flow

• FT — Google Burns Through $6 Billion in Cash as AI Spending Climbs Again · Bloomberg — Google Says Cloud Services Backlog Expands to $514 Billion · Bloomberg — Google Boosts 2026 Spending Estimate to as Much as $205 Billion · WSJ — Google Has the Muscle to Overpower Spending Worries · Axios — Google Releases Series of New, Cheaper Gemini AI Models · Google — Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber

• AI-RTZ #989 — Google’s Strong Quarter and Doubling AI Capex · AI-RTZ #791 — Flashing Green Lights for AI in 2025 · AI-RTZ #1155 — Google under fire again for ‘breaking the internet’

Event 2 — The Rest of Big Tech Reports Next Week

• NYTimes DealBook — Google Sets the Pace on Hefty AI Spending

• AI-RTZ #800 — Big Tech roars ‘AI’ on mid-year earnings calls · AI-RTZ — Booms that Rhyme — Internet/Telecom Capex vs AI Tech Wave Capex

Event 3 — Global AI Capex at Unprecedented Levels ($1.65T)

• Nikkei Asia — Five US Tech Giants’ Hidden Debts Soar to $1.65 Trillion on Opaque AI Funding

• AI-RTZ #529 — Multi-trillion $$ AI Data Center Compute builds ahead · AI-RTZ #1134 — Latest report on an AI Bust and the Global Economy

Gadget AI — Ford Runs Apple Maps in a New EV

• NYTimes — Ford Will Use Apple Maps in New Self-Driving System · AI-RTZ — Apple Car, Lessons Learned

Clips from today

Clip 1 — Improving Voice-Activated Navigation

Watch on YouTube Shorts

Even with advanced Siri and Google Gemini, in-car voice navigation is still a bit clunky.

MP Take: The simple things are still the hard things. When you want to say ‘just drive me to the next gas stop’ or ‘the next place for coffee,’ the systems — even integrated into the nav — still make you take your eyes and attention off the road. I’d like that a lot smoother and more seamless. We’re getting there, but it’s not quite there yet.

Clip 2 — AI Compute Demand Surges

Watch on YouTube Shorts

Why the huge AI-infrastructure spend isn’t as scary as the headlines suggest.

MP Take: The demand on a bottom-up basis is very much there. Enterprises are scaling up for all manner of AI applications layered on top of 50-100 years of existing software — that takes years to deploy. It’s why Elon can rent SpaceX’s spare capacity to Anthropic and OpenAI at 2-3x market rates. And there’s the Jevons Paradox: new, cheaper ways to do things don’t shrink demand, they accelerate it — especially with strong open-source models from China.

Clip 3 — AI Investments: No Signs of Retreat

Watch on YouTube Shorts

Everyone’s waiting for one big tech company to ‘blink’ on AI spending. Don’t count on it.

MP Take: After Google, Amazon, Meta and Microsoft all report next week, and I expect every one just as committed to their AI capex. No one is ready to take a step back — and that’s exactly the signal the bears are waiting for. The moment one company says ‘we’re pulling back’ and its stock goes up, that’s the tell that would bring the whole AI-infrastructure complex — Nvidia, memory, the infra names — down with it. I’m not expecting it. (No stock advice.)

Clip 4 — Ford’s Pragmatic Approach to Tech Integration

Watch on YouTube Shorts

Ford is building Apple Maps into a new EV at the operating-system level — a first.

MP Take: This isn’t CarPlay or Android Auto running off your phone; it’s Apple Maps built into the OS of a new ~$30,000 Ford skunk-works EV. GM has barred CarPlay, Tesla runs a full vertical stack, so this is notable for Apple — long the #2 in maps behind Google. A sub-$60-billion Ford leaning on a ~$5-trillion Apple shows the software age we’re living in. And Chinese players like Xiaomi and Geely are racing here too.

About AI Ramblings Daily (ARD), and AI-RTZ

Both are daily. Both are free. Both are about AI. But they’re different mediums carrying different messages.

AI-RTZ is the morning text — a deeper written take on one idea, published by at least 5 AM EST. Today: post #1156.

AI Ramblings Daily is the afternoon video + podcast — my ad hoc takes and perspective on the day’s AI issues & news flow, around 20 minutes, with short 1-2 minute clips for quick topic views. Today: episode #125.

Subscribe to either or both on michaelparekh.substack.com. They run as separate Sections you can opt into or out of.

Links used in today’s show (already embedded inline above; listed here for reference)

Take 1 — Google Ramps AI Capex at Negative Free Cash Flow:

• FT — Google Burns Through $6 Billion in Cash as AI Spending Climbs Again

• Bloomberg — Google Says Cloud Services Backlog Expands to $514 Billion

• Bloomberg — Google Boosts 2026 Spending Estimate to as Much as $205 Billion

• WSJ — Google Has the Muscle to Overpower Spending Worries

• Axios — Google Releases Series of New, Cheaper Gemini AI Models

• Google — Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber

• AI-RTZ #989 — Google’s Strong Quarter and Doubling AI Capex

• AI-RTZ #791 — Flashing Green Lights for AI in 2025

• AI-RTZ #1155 — Google under fire again for ‘breaking the internet’

Take 2 — The Rest of Big Tech Reports Next Week:

• NYTimes DealBook — Google Sets the Pace on Hefty AI Spending

• AI-RTZ #800 — Big Tech roars ‘AI’ on mid-year earnings calls

• AI-RTZ — Booms that Rhyme — Internet/Telecom Capex vs AI Tech Wave Capex

Take 3 — Global AI Capex at Unprecedented Levels ($1.65T):

• Nikkei Asia — Five US Tech Giants’ Hidden Debts Soar to $1.65 Trillion on Opaque AI Funding

• AI-RTZ #529 — Multi-trillion $$ AI Data Center Compute builds ahead

• AI-RTZ #1134 — Latest report on an AI Bust and the Global Economy

Gadget AI — Ford Runs Apple Maps in a New EV:

• NYTimes — Ford Will Use Apple Maps in New Self-Driving System

• AI-RTZ — Apple Car, Lessons Learned

Companion text:

• AI-RTZ #1156 — Anthropic and OpenAI Now Against China Open-Source AI

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here.)

Subscribe to AI: Reset to Zero for daily AI Ramblings + Sunday Bigger Picture posts.