NVIDIA Wants Every Watt Working Harder

AI is moving beyond better models. NVIDIA is squeezing more compute from the same power. Bill Gates is drawing boundaries around AI’s role in relationships. And Google is testing purchases through Gemini with Flipkart. Together, these stories show AI entering infrastructure, commerce, and everyday life.

In today’s post:

• NVIDIA found 49% more AI throughput without adding power

• Google wants Gemini to become your checkout

• Bill Gates won’t use AI for this

SPONSORED BY

A free newsletter with the marketing ideas you need

The best marketing ideas come from marketers who live it.

That’s what this newsletter delivers.

The Marketing Millennials is a look inside what’s working right now for other marketers. No theory. No fluff. Just real insights and ideas you can actually use—from marketers who’ve been there, done that, and are sharing the playbook.

Every newsletter is written by Daniel Murray, a marketer obsessed with what goes into great marketing. Expect fresh takes, hot topics, and the kind of stuff you’ll want to steal for your next campaign.

Because marketing shouldn’t feel like guesswork. And you shouldn’t have to dig for the good stuff.

Sign Up Free

What’s Trending Today

AI INFRASTRUCTURE

The next AI breakthrough might come from using electricity smarter.

NVIDIA just published an interesting AI infrastructure experiment. The problem wasn’t a shortage of GPUs. It was how much power those GPUs could safely consume. Data centers reserve electricity for worst-case demand. But GPUs rarely hit peak power simultaneously. That leaves valuable capacity sitting unused.

• NVIDIA tested DSX MaxLPS under a fixed 264.4 kW budget.

• The baseline supported 140 GPUs using traditional static provisioning.

• Dynamic power allocation increased that number to 192 GPUs.

• That meant 37% more GPUs without increasing provisioned power.

• Aggregate throughput jumped 49.2%, reaching 1.62 million tokens per second.

• Throughput per provisioned watt climbed from 4.10 to 6.12 tokens.

• There was a trade-off: P99 time-to-first-token increased by 17%.

AI infrastructure is becoming an optimization problem. Buying faster chips is only one part. Power, cooling, networking, and utilization increasingly determine useful compute. The interesting question may soon change. Not “How many GPUs do you have?” But “How much work can those GPUs actually produce?”

TESTING

AI shopping is moving from recommendations to transactions.

Google is testing something important in India. Some Gemini users can now find a product. Then they can buy it through the same interface. The experiment starts with Flipkart. But the bigger story is what happens to ecommerce.

• Google is testing “Buy” buttons on select Flipkart products.

• The test covers some electronics, smartphones, and accessories.

• Shoppers enter a Flipkart checkout flow from Google’s AI.

• Google reportedly plans broader availability during October.

• Google already built UCP for AI-powered commerce interactions.

• Flipkart is also strategically important because Google invested in it.

• Amazon listings appeared without the same direct purchase option.

Search used to send people somewhere else. AI could change that relationship completely. Discovery, comparison, recommendation, and checkout may merge together. That creates a new question for every ecommerce company. What happens when customers stop visiting your digital storefront? The next storefront might simply be the conversation.

RESEARCH

Bill Gates uses AI extensively. But one area remains off-limits.

Bill Gates has spent decades around transformative technology. He also uses AI regularly today. But there’s one place he won’t bring it. His personal relationships. That boundary reveals something bigger about AI’s growing role.

• Gates says he doesn’t use AI for personal relationships.

• He does use it extensively in his philanthropic work.

• His foundation applies AI to malaria and malnutrition research.

• Gates believes immediate misuse deserves more attention than distant scenarios.

• He argues self-regulation alone won’t provide enough protection.

• He wants monitoring and safeguards built into AI systems.

• He says global AI agreements could prove unusually difficult.

The interesting question isn’t whether we should use AI. That question has mostly been answered. The harder question is where we stop using it. Technology becomes powerful when it removes friction. But some friction may be worth keeping. Especially when being human is the entire point.

Free Guides

My Free Guides to Download:

🚀 Founders & AI Builders, Listen up!

If you’ve built an AI tool, here’s an opportunity to gain serious visibility.

Nextool AI is a leading tools aggregator that offers:

• 500k+ page views and a rapidly growing audience.

• Exposure to developers, entrepreneurs, and tech enthusiasts actively searching for innovative tools.

• A spot in a curated list of cutting-edge AI tools, trusted by the community.

• Increased traffic, users, and brand recognition for your tool.

Take the next step to grow your tool’s reach and impact.

Reach 150,000+ READERS:

Expand your reach and boost your brand’s visibility!

Partner with Nextool AI to showcase your product or service to 140,000+ engaged subscribers, including entrepreneurs, tech enthusiasts, developers, and industry leaders.

Ready to make an impact? Visit our sponsorship website to explore sponsorship opportunities and learn more!