cd /news/ai-infrastructure/ai-inference-servers-shrink-to-fit-t… · home › topics › ai-infrastructure › article
[ARTICLE · art-172797] src=siliconangle.com ↗ pub= topic=ai-infrastructure verified=true sentiment=↑ positive

AI inference servers shrink to fit the air-cooled racks enterprises already run

Axelera AI B.V. unveiled Titania, a third-generation inference architecture that scales to a 400-watt chip and a six-kilowatt server designed to fit existing 20- to 30-kilowatt air-cooled enterprise racks, vice president and head of product Manuel Botija said at SC26. Axelera's Europa chip has shipped since September, and the company positions the power-efficient design, built on digital in-memory computing, against GPU racks that cost at least $1 million and typically require liquid cooling. "Enterprises cannot own their own infrastructure. They're locked out because of the cost and the power," Botija said, arguing the hardware lets companies amortize their own AI infrastructure instead of renting cloud compute.

key takeaway

Axelera's Titania architecture puts a 400-watt inference chip into a six-kilowatt server that fits 20- to 30-kilowatt air-cooled racks, avoiding the liquid cooling and $1 million-plus cost of GPU racks.

by read3 min views1 publishedNov 19, 2026
AI inference servers shrink to fit the air-cooled racks enterprises already run
Image: Siliconangle (auto-discovered)

AI inference servers shrink to fit the air-cooled racks enterprises already run

Power-efficient inference chips are becoming the entry price for enterprises that want to run AI on hardware they own.

As AI shifts from building models to running them every day, costly graphics processing unit racks and the liquid cooling they demand keep many companies renting compute instead. That gap is opening room for challengers to the incumbent chipmakers. Axelera AI B.V. built its business at the edge with low-power chips and edge AI development tools, and it is now taking that technology into data center infrastructure, according to Manuel Botija (pictured), vice president and head of product at Axelera.

“Enterprises cannot own their own infrastructure. They’re locked out because of the cost and the power,” Botija said. “We are enabling them now to own it. It’s not that we are helping the world save power or save money — which we are — but it’s that we are unlocking use cases that were simply not possible before.”

Botija spoke with theCUBE Research’s John Furrier and Rebecca Knight at SC26, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how power-efficient inference chips are heading for the data center and how enterprises locked out by the cost and power of GPU racks can own their own AI infrastructure instead of renting compute. ( Disclosure below.)*

Power-efficient inference chips head for the data center

Axelera’s Europa chip has been shipping since September. Ahead of SC26, the company unveiled Titania, a third-generation architecture that scales the design up to a 400-watt chip. The resulting six-kilowatt server is built to perform like leading data center GPUs while fitting into the existing 20- to 30-kilowatt air-cooled racks of an ordinary enterprise data center, Botija explained.

“Having a six-kilowatt server that can perform like a data centre DPU is a pretty big deal,” he said. “It basically means that they have some storage equipment, some networking equipment, some traditional CPU equipment, and next to all of that, they can also insert some AI inference hardware acceleration without needing to go and rebuild another server room or another data center.”

For enterprises that would rather own their AI hardware than pay a recurring cloud bill, the barriers have been steep. A GPU rack costs at least $1 million, often far more, and usually requires liquid cooling, Botija noted. Axelera’s answer rests on digital in-memory computing, a technique that grew out of its founders’ research and gives the company’s power-efficient inference chips a performance-per-watt advantage. “Now you can own your own infrastructure,” Botija said. “You can amortize those investments and … reduce your OPEX, and therefore be more profitable as a company.”

Stay tuned for the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of SC26.

( Disclosure: NYSE Wired sponsored this segment of theCUBE. Neither NYSE Wired nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)*

Photo: SiliconANGLE

Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.

  • 15M+ viewers of theCUBE videos , powering conversations across AI, cloud, cybersecurity and more
  • 11.4k+ theCUBE alumni — Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network

Are you an AWS customer?  Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/

About SiliconANGLE Media

SiliconANGLE,

theCUBE Network,

theCUBE Research,

CUBE365,

theCUBE AIand theCUBE SuperStudios — with flagship locations in Silicon Valley and the New York Stock Exchange — SiliconANGLE Media operates at the intersection of media, technology and AI.

Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.

── more in #ai-infrastructure 4 stories · sorted by recency
── more on @axelera ai b.v. 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/ai-inference-servers…] indexed:0 read:3min 2026-11-19 · —