The Cloud Is the Engine Now

For months, Wall Street has been asking the same question: Is Google spending too much on artificial intelligence? The company’s latest earnings report offers a clear answer — not yet.

Alphabet’s cloud business just delivered its best quarter ever. Google Cloud revenue hit $24.8 billion, an 82% jump from the same period last year. That handily beat the $22.46 billion analysts had predicted. Even more telling: last quarter’s growth was already a robust 63% to $20 billion. The acceleration is real.

What’s driving it? Enterprise AI adoption. Companies are buying into Google’s AI infrastructure and enterprise AI solutions at a furious pace. The cloud contracting backlog — work signed but not yet billed — now stands at a staggering $514 billion.

Profit Explodes, Revenue Hits New Highs

The numbers go beyond cloud. Alphabet’s quarterly profit soared to $112.1 billion, up from just $28.1 billion a year ago. Total revenue climbed 24% year-over-year to $119.8 billion. Even the core Google Services segment grew 15% to $94.5 billion.

This marks the company’s 12th consecutive quarter of double-digit revenue growth. But even by that standard, this quarter stands out. Alphabet earnings reports rarely show this kind of acceleration across every major business line.

Gemini Hits 950 Million Users

Google’s AI chatbot, Gemini, is also gaining traction. The app now has 950 million monthly active users, up from 750 million in Q4 2025. That’s a 27% increase in just a few months.

CEO Sundar Pichai summed it up during Wednesday’s earnings call: “Our AI investments are redefining what’s possible across every part of our business. We have exciting momentum across the board.”

The $180 Billion Question

Alphabet isn’t slowing down its spending. Capital expenditures — the money poured into data centers, chips, and infrastructure — are projected between $180 billion and $190 billion for the year. Analysts pressed Pichai hard on when those investments will actually pay off.

His answer was specific: “I think our compute capacity investments in ’27.” He pointed to “strong demand indicators, including long-term deals” and added that “the dynamics look healthier than where we were about a year ago.”

In other words, the payoff isn’t imaginary. It’s just not fully here yet.

What This Means for Investors

For anyone worried that AI spending is a black hole, the cloud numbers offer real evidence otherwise. Enterprise customers are voting with their wallets, and Google is collecting the revenue. The $514 billion backlog means future quarters are already largely locked in.

That doesn’t mean the risk is gone. Spending at this level is unprecedented. But the cloud business is now big enough — and growing fast enough — to justify the bet. If you’re tracking enterprise AI adoption trends, Google just became the strongest case study yet.

The next milestone? 2027. That’s when Pichai expects the infrastructure buildout to start generating its full return. Until then, the cloud numbers will keep telling the story.

Google Home just made automations a whole lot simpler

Google is rolling out a fresh update to the Google Home app — version 4.20 — and the big news is aimed squarely at people who’ve never bothered with automations. Instead of staring at a blank screen full of triggers and conditions, you’ll now find a handful of pre-built routines ready to go. Tap one, and it’s live.

For years, setting up an automation meant piecing together your own triggers, conditions, and actions. That’s a lot of mental overhead for anyone who just wants their lights to turn off at midnight or the thermostat to drop when they leave the house. The result? Most users never touched the feature at all.

That changes now. The new Google Home automations tab comes stocked with ready-made routines for common scenarios — think home security, morning wake-up sequences, and energy-saving schedules. Google says you can also tweak any suggested automation to fit your exact needs, or swipe it away if it doesn’t apply.

No more starting from zero

Until this update, the barrier to entry was real. You had to know your way around the app’s logic builder — pick a trigger (like motion detected or time of day), set conditions (only when you’re home), and choose an action (turn off the lights). That’s three steps of decision-making for one routine. Multiply that by a few automations, and it’s no wonder adoption was low.

Now, you open the Automations tab and see a list of suggestions. One tap enables the routine. It’s that simple. And if you want to customize — say, change the time a morning routine runs or add a device — you can edit it just as easily.

Google’s support page confirms users can also dismiss suggestions they don’t find useful, so your tab won’t get cluttered with irrelevant options.

Gemini for Home gets smarter too

Alongside the automation overhaul, Google is upgrading Gemini for Home. Weather forecasts and general knowledge answers on smart displays now appear with refreshed visual cards — cleaner layouts and better polish on screen. Sports fans get more accurate scores, schedules, and team standings, which is a welcome fix for anyone who’s seen a wrong score flash up mid-game.

Gemini’s Continued Conversation feature is also more reliable now. Google says back-to-back commands and follow-up questions should no longer get interrupted by a voice verification prompt in the middle of a chat. That’s a subtle but meaningful improvement for anyone who talks to their smart display regularly.

Smaller fixes and camera improvements

The update also brings a handful of bug fixes and minor features. A dedicated light toggle is now built into the camera view for the onn Outdoor Camera Plug-in and onn Floodlight Camera Hardwired — a small convenience that saves you jumping into settings just to flip the light on.

Device setup for older Nest cameras has also been made more reliable, which should reduce the frustration of a camera that refuses to pair on the first try.

Altogether, version 4.20 of the Google Home app is a meaningful step toward making smart home automations accessible to everyone — not just the tinkerers who enjoy building routines from scratch. If you’ve been putting off setting up automations because it felt too complicated, now might be the time to give them a try.

The comparison that didn’t embarrass anyone

I braced for disaster. Digital Trends put Meta Muse, Gemini Nano Banana 2, and OpenAI ChatGPT Images 2.0 through a battery of prompts designed to expose their worst habits. Plastic faces. Mangled hands. Fake products with text that reads like a curse in an unknown alphabet. Instead, the models performed competently. That competence felt more unsettling than outright failure.

These aren’t the same engine wearing different skins. Meta pitches Muse as a social-image model baked into Meta AI and its family of apps. Google frames Nano Banana 2 around speed, editing, and Gemini’s broader world knowledge. OpenAI sells ChatGPT Images 2.0 on text rendering, visual control, and better prompt handling. Different ambitions. Same polished little showroom.

The uncanny valley got quieter

The first prompt asked for a tired office worker eating instant noodles at midnight. The detailed version added a messy kitchen, a laptop, dishes in the sink, and harsh refrigerator light. Meta and Gemini generated quickly. ChatGPT gave me the most distinct simple image, then mostly repeated that mood when I added more detail.

That wasn’t a failure. It obeyed. The problem was that all three seemed more interested in quality than realism. The rooms were composed. The lighting was handsome. The exhaustion looked art-directed by someone who’s never eaten noodles over the sink at midnight.

Readable text became the easy win

The poster test should have been a mess. I asked for a fake coffee shop poster for ‘Bad Wi-Fi Café,’ then gave the detailed version exact text to include. All three made decent posters. The text was readable enough to count as progress. AI image tools used to treat letters like cursed decoration. I’ll give them that.

The finish still had the usual stain. Everything carried that warm AI yellow tint, as if every café, kitchen, and bedroom had been lit by a sponsored sunset. Gemini made what looked like a photo of a poster rather than the poster itself. ChatGPT had the roughest showing, taking more than three minutes, failing three times, and only working after I started a new chat.

Product imagery and the generic premium mush

The cat, suitcase, and umbrella prompt was cleaner. The models followed placement instructions well. Product imagery was shakier. A sleek earbud photo was easy. But ‘open-ear wireless earbuds’ pushed all three into generic premium earbud mush. Every model produced something that could sit in an ad, but nothing that looked like a real product you’d find in a drawer.

That’s the new failure mode. These tools don’t break in the old funny ways. The scenes are coherent. The text is readable. The objects usually go where you ask them. The failures have moved from competence to taste.

Manila looked familiar through a filter

The Manila street food prompt could’ve collapsed into tourist-board soup. Meta and Gemini did better in the runs I completed. They included plastic stools, wet pavement, motorcycles, steam, tarps, and the cramped casualness of eating outside while the city keeps moving.

That made the weakness easier to see. The ingredients were right, but the finish still felt too smooth. Like the street had been cleaned for a brochure shoot five minutes before the rain started. Even when the details matched, the soul didn’t.

If you dropped these images into a set of real photographs, would you be able to tell? Probably not on a quick scroll. That’s the point.

The fake premium era is harder to ignore

None of these tools collapsed in the old funny ways. The scenes were coherent. The text was readable. The objects usually went where I asked them to go. The failures have moved from competence to taste.

That leaves a weirder question than which model ‘wins.’ In the post-truth era, does it matter which tool is the most realistic, or which one follows instructions best, when all three can already make a fake thing look plausible enough to survive a quick scroll?

These models have learned how to make images look expensive before they’ve learned how to make them feel lived in. The fake premium era is less funny than nightmare fingers, but it’s probably more useful. That’s exactly what makes it harder to ignore. For a deeper look at how AI image generators handle text, check our breakdown of ChatGPT image text rendering capabilities.