IBM CEO 表示开源权重模型在本地部署中表现优秀且成本低廉

Jump to main content

Search

TOPICS

• Security

  • All Security
  • Cyber-crime
  • Patches
  • Research
  • CSO

• Off-Prem

  • All Off-Prem
  • Edge and IoT
  • Channel
  • PaaS and IaaS
  • SaaS

• On-Prem

  • All On-Prem
  • Systems
  • Storage
  • Networks
  • HPC
  • Personal Tech
  • Cx0
  • Public Sector

• Software

  • All Software
  • AI and ML
  • Applications
  • Databases
  • DevOps
  • OS Platforms
  • Virtualization

• Offbeat

  • All Offbeat
  • Columnists
  • PHP2Shell
  • BOFH
  • Legal
  • Site News
  • About Us

• Special Features

  • All Special Features
  • Nvidia GTC Berlin 2026
  • Cybersecurity Month 2026
  • VMware Explore 2026
  • Cloud Infrastructure Month 2026
  • HPE: AI Explainers
  • Agentic AI
  • The Future of the Datacenter
  • AWS:Reinvent
  • Nvidia GTC
  • Supercomputing Month
  • Computex 2026
  • AI Infrastructure Month 2026
  • The State of Storage 2026
  • RSA Conference

• Vendor Voice

  • All Vendor Voice
  • Modernizing Financial Services with FIS and AWS
  • Barco
  • Infinidat
  • Everpure
  • Rubrik
  • Make it real with Capgemini and AWS
  • Money Movement Hub
  • ZTE
  • Nutanix: Scale Kubernetes. Not Chaos.
  • AWS New Horizon
  • Digicert
  • Netscout

• Resources

  • Intelligence
  • Webinars & Events
  • Newsletters

Search

• Sign in

• AI

• Security

• Software

• Microsoft

• Developer

• Open Source

• BOFH

• Who, Me?

• On Call

• PHP2Shell

REG AD

systems

IBM CEO says open weight models are now irresistibly good and cheap on-prem

90 percent of the performance at ten percent of the price, in just two or three racks, is hard to beat

Simon SharwoodSimon SharwoodAPAC Editor

33

Published Thu 22 Oct 2026 // 04:26 UTC

Make us preferred on Google

READ MORE

• After threatening AI export bans, China allows DeepSeek to keep selling now

• Autonomous attack agent in 'Skynet mode' hijacks other models to carry out its evil deeds 1 day ago

• OpenAI disables internet access during model training while Anthropic issues another mea culpa 2 days ago

• Red-team report found OpenAI utterly failed to monitor its agents during Hugging Face attack 2 days ago

• AI models describe humans as a bunch of amoral sheep 2 days ago

IBM CEO Arvind Krishna has told investors that open weight models now make brilliant sense for on-prem enterprise use.

Speaking on the company’s Q3 earnings call, Krishna said that in 2024 closed frontier models were plainly superior to open alternatives.

“If you ran some assessment, some benchmark, and it was broad-based, not narrow, the top model might score a 70, and the open model might score a 20 or a 30.”

REG AD

“In late 2025, we began to see that gap shrink, where maybe the gap was not 70 to 20 or 30, it was 70 to 50,” he continued.

REG AD

The CEO said open weight models currently deliver “anywhere from 25 percent to 10 percent of the token cost of the frontier models. That is very attractive given that they’re now 90 percent as effective.”

“It’s a simple question for any enterprise. Would you like something that’s 10 percent of the cost if it’s 90 percent as good as the best, or do you want to pay that 10 times higher cost?”

Krishna thinks businesses will most often choose not to incur the higher costs.

“Then the question is: how many tasks require that? I would tell you some do, but the vast majority, summarizing an email, examining a contract, answering a customer call, composing an email, all of those can be done as well with the much cheaper answer.”

The CEO said those numbers mean IBM is seeing booming interest in on-prem AI infrastructure. While he thinks most businesses can build useful AI infrastructure in just two or three racks, he feels many will struggle to get it running.

“The question becomes: how do you do that?” he said. “The question that the enterprise struggles with is: ‘Where are my kernel engineers? Where are the people who worry about the underlying network? Who can make sure that the model is clustered on GPUs that are close to each other in network proximity? Who can make sure that you’re doing the right caching of results so you get an incredible response?’”

He didn’t have answers, other than to say IBM will produce software that eases AI deployments. Krishna also predicted demand for on-prem AI will spur sales of Big Blue’s Spyre mainframe accelerator cards.

The CEO said current-gen z17 mainframes have delivered $1bn more revenue than any previous model, and said Big Blue’s new rackable models delivered “the strongest midrange launch in IBM Z in nearly two decades.” Mainframe revenue, however, fell 20 percent – as IBM expects given it’s now 18 months since it launched the z17.

REG AD

IBM’s infrastructure revenue grew five percent year-over-year to $3.7 billion, much slower improvement than recorded by rivals Dell, HPE, and Lenovo.

Software did rather better, scoring nine percent growth to reach $7.8 billion for the quarter. Red Hat was the star, growing eleven percent and increasing annual recurring revenue by 30 percent. Some of that future revenue is for Red Hat’s virtualization products, which have now won $800 million of deals, with a pipeline of a further $750 million. That’s a solid result for a product line launched in 2024, and IBM believes there’s more to come given the possibility some VMware users are on the move.

Total quarterly revenue hit $17.1 billion, up five percent. IBM’s share price wobbled in after-hours trading and ended down a point or two.

‘Emergency maintenance’ for IBM Cloud

Another IBM matter The Register is watching today is a severity one incident in the Big Blue cloud, caused by “emergency maintenance” that meant users could not provision new resources or virtual private clouds for 90 minutes. Users of IBM’s Classic Cloud, the service it acquired from SoftLayer in 2013, have been told Big Blue will “restore new provisioning and resource management in our Classic environment” by 08:00 UTC on Thursday – meaning the incident will last at least 16 hours.

IBM commenced that work a handful of hours after the announcement of critical bugs in PHP. We’ve asked if fixing that flaw is the reason for the incident and will update this story if we get a response. ®

ai and mlibmopen weight modelsmainframesystems33

Make us preferred on Google

REG AD

asia in brief ## After threatening AI export bans, China allows DeepSeek to keep selling PLUS: AI datacenter bubble wobbles in Australia; E-commerce app suggests it can detect diseases; India’s top telco to IPO amid investor frenzy; and more!

Reg readers land on 'Binface' as name for UK space cleanup mission Intergalactic Space Warrior, or a bin in orbit?

Why the immediate future of liquid cooling is hybrid PARTNER CONTENT: AI is driving liquid cooling adoption, but hybrid architectures will help operators balance density, flexibility, and operational realities

Autonomous attack agent in 'Skynet mode' hijacks other models to carry out its evil deeds Designed to carry out 100,000 steps without human approval

## Torvalds: AI is fine for competent devs, but spare us the 'drive-by' patches Kernel maintainers stop asking who used AI, as long as the code is good

OpenAI disables internet access during model training while Anthropic issues another mea culpa We've totally got things under control

TOP STORIES

• #### Indian certification authority eMudhra’s private keys compromised – one of them was the number ‘4’
• #### OpenClaw creator says he spent $10 million in tokens last month
• #### PHP devs publish critical bug fix as site owners warned: Pull 'em offline if you're unpatched
• #### PHP blames supply chain security biz Chainguard for putting huge chunks of the web at risk
• #### PHP2Shell exploit bypasses protections designed to stop remote code execution
• #### Google Cloud's London zones takes a nap, multiple services disrupted

AI

• Torvalds: AI is fine for competent devs, but spare us the 'drive-by' patches Kernel maintainers stop asking who used AI, as long as the code is good

• Nvidia woos wary Europeans with more open weights models New embedding, parsing, and reranking models promise smarter agentic RAG workflows

• Jensen Huang sends in the robot ducks to sell Nvidia's AI-everywhere vision Containerized datacenters, androids, and telecoms all feature in Berlin keynote interrupted by environmental protesters

• Nvidia's grand plan for speeding up datacenter deployments: AI trailer parks GPUzilla boasts modular bit barns could speed up deployments from six months to just two weeks

• The spec is the new code, so brush up on your prose, says Kaseya CTO Domain expertise and clear writing will matter more than syntax, and yes, that means documentation

Infosec

• Security #### Russians are posing as Signal support to launch phishing attacks PLUS: US takes down Iranian propaganda sites; Marketing company asks 'Why Do We Have Your Information?' And more!
• Security #### Microsoft patches failed to fix on-prem SharePoint, which is now under zero-day attack PLUS: China upgrades smartphone surveillance tools; Ring eases anti-snooping stance; and more
• Black Hat and DEF CON #### DEF CON Franklin project enlists hackers to harden critical infrastructure Voting village reports have been so successful, says Jeff Moss, that the whole of DEF CON will now be included
• Security #### EQT buys majority share in Swiss cybersecurity biz Acronis Went at equivalent of $3.5B+ valuation for entire firm, though portion sold not specified
• Malware Month #### Ten years since the first corp ransomware, Mikko Hyppönen sees no end in sight On the plus side, infosec's a good bet for a long, stable career

FOSS

• SparkyOS: Lightweight general-purpose Arch, on a pendrive Interesting new project from the minds behind SparkyLinux

• Red Hat helps teach about tech writing But learns tough lessons about AI in the process

• NetBSD may have found a new role: The AI-free FOSS OS Increasing awareness and adoption, followed by increasing donations

• OpenAI dots inspire open source imitators amid technical difficulties Persistent agents appeal to AI companies, but demand isn't obvious

• Ubuntu 'Stonking Stingray' beta swims out: Deeper Rust 'oxidization' plus LLM speech-to-text GNOME 51 is here, and Ubuntu interim releases are getting interesting again

• Revive Raskin's 3 laws of software with humane work that considers a dev's duty of care All this, and angry NeoVim users

✕

Biting the hand that feeds IT

About Us

Contact us
Advertise with us
Who we are
Newsletter

Our Websites

The Next Platform
DevClass
Blocks and Files
Situation Publishing

Your Privacy

Cookies Policy
Privacy Policy
Ts & Cs
Do not share my personal information
Your Consent Options

Archives

27 years of articles

Copyright. All rights reserved © 1998-2026.