HARDWARE

Nvidia's Vera Rubin finally has a worthy rival

Much of the value of the current AI boom has accrued disproportionately to one company: Nvidia. But the biggest challenge to its flagship product has now emerged. 

AMD Helios has arrived to take on Nvidia's Vera Rubin, the most powerful system in the world for both training the most advanced AI models and running them at scale. While nearly all the major AI and tech companies are trying to build their own AI chips now, Vera Rubin — and now Helios — represent the top of the food chain. 

These are the rack-scale systems that combine CPU, GPU, high-bandwidth memory, networking, and software into one platform. And they are the kinds of chips that will power the AI factories, those massive data centers custom-designed to efficiently run the most advanced cutting-edge systems. 

"Helios is truly a game-changer," AMD CEO Lisa Su said last year when the company first announced the product. "For the first time we architected every part of the rack as a unified system. That's combining our CPUs, our GPUs, our Pensando NICs, and our ROCm software all together in one platform. And it's really purpose-built for the most demanding AI workloads from training the largest frontier models to scaling inference across thousands of nodes."

It's not an exaggeration to say that AMD bet the company on this product. AMD's data center GM Forrest Norrod told CNBC that it has a town hall every quarter for everyone working on Helios, and attendance has swelled to 8,000 employees.

On Monday, AMD Helios got a major vote of confidence from the world's second-largest hyperscaler. Microsoft officially signed on as a partner to deploy the full-stack version of Helios on Azure to allow its customers to train frontier models, deploy AI inference at scale, and offer the most advanced AI agent capabilities. 

AMD claims that its first-generation AI rack system will rival Nvidia's second-gen system in virtually every aspect, and will exceed it in some. Agents are expected to be one of its strengths because agents rely on CPU as much as GPU, unlike chatbots, and AMD has a much longer history in building CPUs than Nvidia. AMD has also signed up Meta, OpenAI, Oracle, and Tata as partners planning to deploy Helios.

AMD's rival to Vera Rubin doesn't mean Nvidia is in trouble. This is not a zero-sum market, but one that is growing rapidly and doesn't have enough chips or servers to grow as fast as it would like. Nvidia has reportedly already sold its next 12 months of products before they are even manufactured. Demand is so strong for the infrastructure to power the growth of AI that there's plenty of room for AMD to ease some of that burden, if its products turn out to be in the same ballpark as Vera Rubin. Helios will cost more, around $5 to $5.5 million compared to $3.5 to $4 million for Vera Rubin, according to the Futurum Group. But it's also promising greater efficiencies and potentially cost savings, so it could be less expensive to operate over time. I'll be at AMD's Advancing AI 2026 event this week in San Francisco and will have more to report on the latest details of Helios and the future of AI infrastructure. 

Disclaimer: Jason Hiner's travel to AMD's Advancing AI 2026 event was paid for by AMD. The Deep View's coverage is editorially independent from the companies we cover.

TOGETHER WITH ATTIO

The next era of revenue runs on Attio.

Get agents working on every account, surfacing new opportunities, and handle the work that used to take your team days. With Attio’s new workflows engine, you can build agents for any GTM play you need. Just describe the goal, and Attio assembles the workflow for you to approve.

Then Ask Attio any question, from the weekly forecast to performance by rep, and get the answer in seconds. So you keep every deal moving.

Trusted by some of the world’s most successful startups, like Granola, Modal, and Railway, Attio is the agentic CRM that transforms how revenue work gets done.

Try Attio for free

CULTURE

AI porn is outpacing safeguards for teens

Generative AI has given anyone the ability to create practically any content with a few simple prompts. That can be dangerous when it's used to distort reality. 

In a Common Sense Media survey of more than 1,300 teens ages 13 to 17, 44% reported that they believed they have seen AI-generated sexual content, including explicit imagery and videos, with this content largely showing up on social media platforms like TikTok, Instagram and X. Of those who reported seeing this content, 79% said their exposure to it was accidental. 

What's more: A quarter of those who have seen AI-generated sexual content say they have seen it depicting themselves or people they know personally. 

• Around two-thirds of teens surveyed worry about being the target of deepfake sexual content without their consent. 

• As a result, 38% reported that they are more careful about what they post, 30% have made their social media accounts private, 23% post pictures and videos of themselves less often, and 9% have deleted social media entirely. 

AI has lowered the barrier for entry to creating and accessing this content, Common Sense Media noted. "Technologies that can generate sexual content are too easily accessible to teens, and the social media platforms that allow the distribution of this content are not doing a good enough job of detecting, blocking, and taking it down," the report wrote. 

Many reported understanding the impacts that AI-generated sexual content could have. Roughly two-thirds of the survey pool reported that they believe this kind of content distorts their peers' ideas of body image, shifts trust in media and has changed expectations of romantic partners. And 60% of those who have seen this kind of content agree that it has warped their perception of what real bodies look like. 

While 82% surveyed agreed that creating this content without the subject's consent should be illegal, nearly a quarter view the content as less harmful than real explicit imagery, according to the report, because "no one gets hurt." 

The report highlights a problem that has manifested in recent years as increasingly powerful AI models continue hitting the market. In particular, it's an issue that xAI's Grok has faced. In January, Grok flooded X with AI-generated deepfake nudity, forcing it to limit the model's image generation capabilities. Still, problems persist: Last week, xAI sued a user who allegedly used Grok to generate child sexual abuse material. 

And the issue extends beyond just one chatbot. On Friday, San Francisco City Attorney David Chiu sent a cease-and-desist letter to Apple and Google, demanding that the tech giants take down 13 "nudify" apps that allow users to digitally undress people. 

“AI companies have made this content easy to create and hard to avoid,” Supreet Mann, director of research at Common Sense Media, told The Deep View. “App stores should stop hosting 'nudifying' apps, incorporate age assurance into platforms that are not intended for minors, and invest in sexual deepfake detection rather than relying on kids to report it after the fact.”

AI-generated explicit content and deepfakes are just one example of a broader issue: When powerful AI can be accessed by anyone with a wifi connection, who is culpable for the damage when things go awry? Broadly, it's true that some AI model providers have tried to account for existing and potential damage, such as OpenAI's parental controls and age prediction system to flag harmful behaviors of young users. Still, for now, given how nascent the tech remains, responsibility for harm is largely on a case-by-case basis and handled through the legal system, such as in the several allegations against AI companies for their models' role in suicides. However, not only is the cat out of the bag, but that cat is growing more powerful by the day. These tools are already in the hands of young users who do not know how to wield them safely, as well as adults who seek to wield them maliciously. Without binding legislation around these issues and as the scale of the damage continues to grow, are the AI labs responsible for cleaning up the aftermath?

TOGETHER WITH CRUSOE

Only pay for fine-tuning that works

GPU-hour billing charges for setup, idle time, queueing, and failures — infrastructure efficiency you don't control.

Crusoe Serverless Fine-Tuning, now GA, bills per token processed during training, so costs are predictable from your dataset size and epoch count.

Early stopping ends the job, and the billing, the moment your model stops improving — you're never charged for epochs that add no accuracy. Iterate faster. No surprise bills.

Start fine-tuning today

WORKFORCE

AI layoffs expose US tech’s efficiency gamble

American tech companies are bearing the brunt of this year's tech layoffs.

About 82% of the 156,975 tech jobs eliminated globally in the first half of 2026 came from US companies, according to a new report from TradingPlatforms, a UK-based research firm that analyzed workforce reductions using data from TrueUp.io, Layoffs.fyi, TechCrunch, and state WARN filings. Almost half of the cuts were tied to US companies restructuring around AI and automation, led by bellwethers such as Oracle, Meta, Block, Cisco, and PayPal, according to the report.

Cloud computing and software-as-a-service companies recorded the highest number of layoffs globally, TradingPlatforms found. Many of those reductions came from Oracle, Microsoft, and Salesforce, among other firms. E-commerce and marketplace companies, including Amazon and eBay, were the second hardest hit. California-based companies accounted for the largest share of US layoffs among the states, followed by Texas, Washington, and New Jersey, according to the report.

Compared to other sectors, tech companies are often the first to feel major economic and technological shifts because they're quicker to adopt new tools, according to Stanislava Savisheva, an analyst at TradingPlatforms.

"Nearly every major technology company is currently looking for ways to operate with leaner teams, greater automation of workflows, and more efficient cost structures, with AI accelerating a shift that was already underway," Savisheva wrote in a statement.

The report comes as US tech companies reorganize their workforces as they bet billions on AI. On July 6, Microsoft announced it would immediately cut about 4,800 jobs, 1,600 of which were across its Xbox gaming division, as it continued ramping up AI investments. In June, Oracle disclosed that it had reduced its workforce by 21,000 employees over the past year to invest in AI. A month prior, Meta cut about 8,000 employees, shifting thousands more workers into AI-focused roles. That same month, Cloudflare eliminated about 1,100 jobs as CEO Matthew Prince said the company was preparing for the "agentic AI era."

"The next phase of this transformation will not simply be about replacing jobs," Savisheva wrote, but about "fundamentally changing how businesses build teams, distribute work, and measure productivity in an AI-driven economy."

AI-driven restructuring is becoming the norm. But companies may want to think twice before cutting staff in anticipation of greater efficiency. Ford, for instance, recently hired more than 350 veteran engineers, some of whom were former employees, after finding its AI-powered quality systems couldn't perform on their own. At the same time, many companies are still struggling to show meaningful returns on the billions they're pouring into AI. As investors and boards push executives to justify those investments, layoffs may prove easier to announce than to live with. According to data from PwC, the companies that benefit most from AI are often the ones that use it to augment employees rather than replace them. Meanwhile, a study from Ramp found that the companies investing the most heavily in AI grew headcount by 10% over the past two years.

LINKS

• China's z.AI completes 1GW data center using only Chinese chips

• Anthropic offers $50,000 in Claude credits for rare disease researchers

• Google is reportedly developing specialized server chip, targets 2028

• OpenAI pauses internal use of unreleased model due to misalignment

• Chris Fall, Head of Trump's Center for AI Standards and Innovation, resigns

• Archer, Anduril plan autonomous aircraft platform for military use

• Qwen3.8 Max: Alibaba released previews of the latest iteration of its flagship model. 

• Claude Fable 5: Starting Monday, Anthropic's flagship models will be included in Max and Team Premium plans, at 50% of limits. 

• Google Gemma: Developers can now use the Gemma 4 31B model as the basis for voice AI. 

• Ramp Router: The payments platform launched a tool to automatically direct AI requests to the right model based on cost, intelligence and latency.

• Anthropic: Standards Editor

• LinkedIn: Staff Security Engineer - AI 

• Nvidia: Senior Scientist, Synthetic Data and Privacy

• Google: Policy Development Lead, YouTube Trust and Safety

GAMES

A QUICK POLL BEFORE YOU GO