We may earn a commission from links on this page. Deal pricing and availability subject to change after time of publication.

Amazon Big Deal Days is October 7-8, and Lifehacker is sharing the best sales based on product reviews, comparisons, and price-tracking tools before it’s over.

• Follow our live blog to stay up to date on the best sales we find.

• Subscribe to our shopping newsletter, Add to Cart, for the best sales sent to your inbox.

• New to Prime Day? We have a primer on everything you need to know. Sales are accurate at the time of publication, but prices and inventory are always subject to change.

We’re right in the heart of October Prime Day again, and together with the rest of the Lifehacker team I’m hunting through Amazon to find the best details for you—and the $315 reduction on the Netgear Orbi 970 Wi-Fi 7 mesh series has definitely grabbed my attention. This is a system I’ve tested in the past for review purposes, alongside many others, and it’s the most powerful Wi-Fi 7 system I’ve used yet.

The price drop here is 16 percent, so you’re going to be paying $1,684.99 instead of $1,999.99. I know that’s a lot of money for a mesh system, but this three-pack is wifi networking for power users—and if you weigh just how important home wifi is, and how much use you get out of it, then it’s a worthwhile investment if you’ve got the budget. According to price trackers, it’s also the lowest this system has ever been.

Netgear Orbi 970

$1,684.99

$1,999.99
Save $315.00

$1,684.99

$1,999.99
Save $315.00

This beast of a Wi-Fi 7 system will completely eliminate dead spots across 10,000 square feet of space, and you can connect up to 200 devices to it without issue—enough to handle even the busiest of households. You’ve got blazing fast wired Gigabit ports here too, and quad-band support to ensure a rock solid connection every time.

Though the theoretical maximum wifi speed is an astonishing 27 Gbps, this won’t actually upgrade the broadband speed coming into your home—but it will make sure that wifi speeds are maximized in every corner of your property. In the time I spent testing out the Orbi 970 I was seriously impressed with the upload and download rates, and the connection stability across dozens of phones, tablets, laptops, and other devices.

Looking for something else? Retailers like Walmart and Best Buy have Prime Day competition sales that are especially useful if you don’t have Amazon Prime.

• Walmart’s Prime Day competition sale runs from Oct. 6 at 7 p.m. ET through Oct. 12 and includes deals up to 50% off. It’s an especially good option if you have Walmart+.

• Best Buy’s Prime Day competition sale runs from Sept. 27 through Oct. 12, and has some of the best tech sales online. It’s an especially good option if you’re a My Best Buy “Plus” or “Total” member.

• Target’s Prime Day competition sale runs from Oct. 5 through Oct. 11, and it has deals going up to 50% off. You can become a Circle member for free.

AmiMoJo shares a report from the BBC: Renewable energy overtook coal as the world’s leading source of electricity in the first half of this year — a historic first, according to new data from the global energy think tank Ember. Electricity demand is growing around the world but the growth in solar and wind was so strong it met 100% of the extra electricity demand, even helping drive a slight decline in coal and gas use. However, Ember says the headlines mask a mixed global picture.

Developing countries, especially China, led the clean energy charge but richer nations including the US and EU relied more than before on planet-warming fossil fuels for electricity generation. This divide is likely to get more pronounced, according to a separate report from the International Energy Agency (IEA). It predicts renewables will grow much less strongly than forecast in the US as a result of the policies of President Donald Trump’s administration. Coal, a major contributor to global warming, was still the world’s largest individual source of energy generation in 2024, a position it has held for more than 50 years, according to the IEA.

Read more of this story at Slashdot.

In a significant shakeup of the AI hardware landscape, AMD has recently secured a major contract with OpenAI, signaling a shift in the high-stakes arms race for dominance in artificial intelligence infrastructure. The deal, rumored to include large-scale deployment of AMD’s MI300X and upcoming MI400 accelerators, positions AMD as a powerful alternative to NVIDIA’s near-monopoly in AI compute.

While full financial details remain undisclosed, insiders suggest the agreement involves a multi-billion-dollar commitment over several years, with OpenAI leveraging AMD’s GPUs across both inference and training workloads. The move underscores OpenAI’s growing appetite for diversification and hints at dissatisfaction with NVIDIA’s supply constraints, pricing strategies and closed ecosystem.

How This Puts AMD Ahead of NVIDIA’s Pending Deal

NVIDIA, long considered OpenAI’s closest hardware partner, has been in discussions for an extended renewal and expansion of their supply agreement. However, sources close to the talks indicate friction around costs and NVIDIA’s bundling practices, which often force cloud providers and AI labs to buy complete DGX systems rather than standalone GPUs.

By contrast, AMD has reportedly agreed to far more flexible licensing terms. AMD’s ROCm software stack, now mature and open-sourced, provides OpenAI with greater control, enabling them to fine-tune hardware-software alignment. This could tilt the scale in AMD’s favor, especially as OpenAI seeks to deepen vertical integration and potentially develop its own chips in-house.

AMD’s History as the Quiet Challenger

AMD has always played the underdog role—competing against Intel in CPUs and NVIDIA in GPUs for decades. Under the leadership of Dr. Lisa Su, AMD has methodically clawed market share across both fronts. First with the Zen architecture against Intel, then with CDNA and ROCm in AI compute, AMD has gone from being dismissed to dominating benchmarks and securing high-profile design wins.

This new OpenAI deal adds another major player to AMD’s growing list of customers—joining Microsoft Azure, Meta and Oracle—all of whom are already deploying MI300X or preparing for its successor.

How AMD Caught Up So Fast

NVIDIA had a ten-year head start in AI, but AMD has used three key strategies to close the gap quickly:

1. Leverage of CPU+GPU integration – With EPYC and Instinct on the same board, AMD delivers dense, high-bandwidth compute nodes ideal for large AI models.
2. Smart acquisitions – Xilinx and Pensando added vital IP around FPGAs and DPUs.
3. Open collaboration – AMD actively works with the PyTorch and Hugging Face communities, optimizing performance across standard ML stacks.

Thanks to these efforts, AMD’s MI300X now matches or exceeds the H100 in many real-world inference workloads, making it a compelling alternative.

The Open-Source Advantage

Where NVIDIA locks customers into CUDA and closed software stacks, AMD offers transparency and openness. Their ROCm platform has become increasingly robust, supporting industry-standard frameworks and delivering near parity with CUDA on key benchmarks.

OpenAI, a long-time advocate of open research (before its partial shift to closed-source models), still relies heavily on open software workflows. AMD’s open ecosystem fits far more comfortably into that vision. Moreover, the ability to modify and adapt ROCm to its own needs gives OpenAI technical flexibility it could never get from NVIDIA.

NVIDIA’s Arrogance Is Catching Up

NVIDIA’s historic success has bred complacency and hubris. Jensen Huang’s company has repeatedly been accused of vendor lock-in, excessive pricing and squeezing smaller partners. As demand for H100s exploded, NVIDIA prioritized top-tier clients, leaving many cloud providers and labs scrambling.

Such behavior has alienated customers. AMD, in contrast, is taking the opposite route by working closely with partners, offering transparent roadmaps and delivering value-based pricing. OpenAI’s shift may be just the beginning of a broader customer exodus.

AI Market Bubble and the Looming Economic Downturn

It’s no secret that the AI market has reached fever-pitch levels of hype. With over $150 billion poured into AI startups since 2023 and GPU scarcity pushing data center budgets to extremes, many analysts believe the industry is nearing a saturation point. A market correction—or even a recession—could reset priorities toward cost-efficiency and ROI.

In that context, AMD’s competitive pricing and more flexible supply terms become not just attractive, but essential. The OpenAI deal may be the first of many such strategic moves as companies hedge against an over-concentrated, NVIDIA-dominated supply chain.

What’s Next for AMD in the AI Race?

Expect AMD to push aggressively on four fronts:

• MI400 Series Launch: Early 2026 will likely bring the MI400X with even more memory bandwidth and AI-dedicated tensor engines.
• AI Chiplet Customization: AMD may co-design chips with major customers, a la OpenAI, using modular chiplet architectures.
• ROCm Ecosystem Investment: Continued funding for open-source libraries, driver optimization, and developer tools.
• AI Software-as-a-Service: There are rumors AMD may partner with or acquire a small cloud AI platform to showcase ROCm performance at scale.

With momentum on their side and a long history of overcoming giants, AMD is no longer just the second choice. It’s a front-runner.

Wrapping Up

The AMD–OpenAI contract is more than a business deal—it’s a signal of shifting tides in the AI hardware wars. AMD has proven that a blend of open standards, customer-centric strategy, and technical excellence can challenge even the most entrenched monopolies. While NVIDIA will remain a powerhouse, it’s dominance is no longer assured.

For OpenAI, diversifying away from a single hardware vendor is a smart hedge. For AMD, this deal opens the door to deeper influence over the future of AI.

In a cooling market, only the most flexible and collaborative companies will survive. Right now, AMD looks like it’s playing the long game—and winning.

By now, a handful of technologies are leading contenders for producing a useful quantum computer. Companies have used them to build machines with dozens to hundreds of qubits, the error rates are coming down, and they’ve largely shifted from worrying about basic scientific problems to dealing with engineering challenges.

Yet even at this apparently late date in the field’s development, there are companies that are still developing entirely new qubit technologies, betting the company that they have identified something that will let them scale in ways that enable a come-from-behind story. Recently, one of those companies published a paper that describes the physics of their qubit system, which involves lone electrons floating on top of liquid helium.

Trapping single electrons

So how do you get an electron to float on top of helium? To find out, Ars spoke with Johannes Pollanen, the chief scientific officer of EeroQ, the company that’s done the new work. He said that it’s actually old physics, with the first demonstrations of it having been done half a century ago.

Read full article

Comments