Nvidia has long been the go-to name in AI chips—it’s what famously made the chip giant the world’s most valuable company amid the generative AI boom.

The dominance of its graphics processing units (GPUs) isn’t likely to change soon, but certain tech giants have increasingly been eyeing more customized alternative processing units from the likes of Broadcom and Marvell. These specialized chips might make operations cheaper and more efficient for big cloud providers and AI companies’ specific AI tasks. But they also come with drawbacks.

Custom AI chips snagged the spotlight last month when OpenAI announced a blockbuster deal with Broadcom around building its own data center infrastructure. Anthropic also recently announced it would tap up to 1 million of Google’s custom TPUs, which it co-developed with Broadcom.

Jefferies analysts wrote in a research note on Broadcom last week that custom application-specific integrated circuits (ASICs) have hit “an inflection point.”

The X factor: Broadcom’s bread and butter has historically been networking infrastructure—the chips and hardware that tie together data centers, cloud platforms, and internet service providers. But the sprawling Palo Alto-headquartered company has carved a fast-growing new niche for itself in XPUs—the “X” being a variable for any given application of AI.

The two business lines complement each other, Peter Del Vecchio, product line manager for Broadcom’s Tomahawk family of data center switches, told Tech Brew, and Broadcom develops the XPUs in tandem with the networking to make them as compatible as possible.

“The two kind of go hand in hand, where a lot of customers say, ‘I’ve got to get an XPU that’s going to talk to a switch,’” Del Vecchio said. “So the switch team here works very closely with the ASICs products division. And a lot of the products come out hand in hand.”

Keep reading here.—PK

EV battery suppliers are under pressure—pressures, actually, according to a new report from AlixPartners.

Against the backdrop of significant policy changes on vehicle electrification in the US and Europe, EV battery makers face growing uncertainty and challenges related to plant utilization and finances, according to the report.

US EV sales are falling after tax credits of up to $7,500 expired at the end of September under President Donald Trump’s tax and budget bill. Automakers and suppliers are canceling EV projects, mulling cuts to existing electric models, and pivoting into other lines of business like energy storage.

To sum up the situation: “Billions of dollars have been invested in R&D, product development, and associated costs to rapidly scale batteries, lightweight materials, and other vehicle technologies designed to make EVs more attractive and affordable,” per AlixPartners.

But the demand that industry players projected just a few years ago hasn’t materialized, and now EV sales are falling. In 2023, AlixPartners’ own forecast called for EVs hitting 36% US market share by 2030. The consultancy has since slashed this projection in half.

“What we see in our study for the EV battery landscape is that there is a real change or a reset going on,” Rohit Gujarathi, SVP at AlixPartners, told Tech Brew. “Because the past few years, the industry built a capacity for EV batteries for demand that has really not materialized at the scale where people expected it to be. We are entering a phase where the supply is outstripping the demand, and it’s creating challenges both at the operational and financial level for the auto suppliers.”

Keep reading here.—JG

What do companies want most when hiring for AI-related roles? Someone with ethics.

According to a recent IEEE study, 44% of technology leaders ranked AI ethical practices as a top skill they’d like to see from candidates applying for AI roles.

The study surveyed 400 technology leaders from Brazil, China, Japan, India, the UK, and the US. Other top skills they wanted from AI hires included data analysis (38%), machine learning (34%), and data modeling (32%).

What exactly are AI ethics skills? IEEE fellow Karen Panetta told IT Brew she defines AI ethical practices as a professional’s ability to keep humans in the loop, as well as their ability to evaluate data funneled into AI for bias.

“[AI] should never replace a human in the final decision process, especially if it affects somebody’s career, their health, or some other performance decision,” Panetta said. “The other piece of it is understanding where or what’s in the model, and where it came from and its limitations.”

Keep reading here.—BM

Listen up: Ear protection is something almost everyone needs at some point—whether mowing lawns, attending rock shows, or just sleeping in a noisy environment. But jamming a bit of foam in your ears comes with drawbacks. Retail Brew has notes on how 3D-printed Loop Earplugs made doing something good for your health downright trendy.

Black box: A new video game console just dropped: The Verge gets into the nitty gritty of the Steam Machine, a new PC-based console from Valve that aims to take on the likes of Microsoft and Sony.

More focus, less fluff. CollabWORK filters out the noise and delivers jobs that actually match what Tech Brew readers are looking for. Click here to see the full board of curated roles.

• Principal Data Modeler and Database Engineer at Raytheon (in person)
• Senior Software Engineer, LiveDesign at Schrödinger (in person)
• Senior Software Engineer, AI at Lyft (in person)
• Principal Electrical Engineer at Raytheon (in person)
• Sr. Principal Software Engineer at Alcon (in person)