Intel puts its spin on an AI spin-off

Welcome to Runtime! Today: Intel makes another software move in hopes of denting Nvidia's AI advantage, why VMware partners are up in arms over Broadcom's new policies, and the latest moves in enterprise tech.

The entrance to Intel's headquarters building in Santa Clara, Calif.
(Credit: Intel)

Welcome to Runtime! Today: Intel makes another software move in hopes of denting Nvidia's AI advantage, why VMware partners are up in arms over Broadcom's new policies, and the latest moves in enterprise tech.

(Was this email forwarded to you? Sign up here to get Runtime each week.)

The opposite of hallucinate?

As Intel has started to regain its footing in the server CPU market, it has been turning its attention to the enormous gap between its AI CPUs and Nvidia's, which have taken the enterprise tech world by storm. Intel kicked off the year by announcing a new startup spun off from internal research that hopes to give enterprises an extra incentive to use its chips.

With financial backing from Intel and DigitalBridge, Articul8 launched Wednesday to continue the AI software-development work Intel did in partnership with Boston Consulting Group. The general idea is to convince companies struggling to deploy AI apps to add yet another tool to their tech budgets, and a lot of enterprise software buzzwords were shipped in service of that goal.

  • For example, according to its website Articul8 customers will be able to "rapidly develop and deploy enterprise GenAI applications with Articul8’s GenAI engine via elegant APIs, enabling effortless integration across development workflows."
  • Elsewhere on that site you'll find references to "transforming customer data into actionable insights" and "quick roundtrip responses."
  • Intel's press release says Articul8 is "a full-stack, vertically-optimized and secure generative artificial intelligence (GenAI) software platform."

Dr. Arun Subramaniyan, CEO of the new startup and a former Intel and AWS executive, shed a little more light on what Articul8 hopes to accomplish in a video.

  • As companies rush to articulate (!) a generative AI strategy amid pressure from executives and boards, they're rolling out a lot of proof-of-concept projects that they're struggling to turn into production-quality deployments, as Deloitte's Nitisha Henry explained in our interview last month.
  • According to Subramaniyan, the four most common deployment problems are speed, scale, sustainable cost, and security.
  • Companies often run into trouble trying to make proof-of-concept AI projects work reliably at scale, he said, and if they do manage to pull that off that trick, what seemed like reasonable expenses for running a small-scale project can skyrocket.
  • The aforementioned "elegant" APIs will help companies that lack a deep bench of data scientists deploy AI applications faster on the hardware or cloud of their choice using their existing security policies, he said.

That last part is an interesting component of Articul8's pitch.

  • While Articul8's software has been optimized for Intel's Xeon CPUs and Gaudi GPUs, it will work with Nvidia's chips and support "a range of hybrid infrastructure alternatives," which presumably includes Arm CPUs from Ampere, AWS, and Microsoft.
  • Nvidia's AI chips get all the headlines, but a vital component of its success in the data center has been its CUDA programming architecture.
  • Building on the launch of the Unified Acceleration Foundation last year, Intel is clearly trying to develop an alternative to Nvidia's software ecosystem, even if it means developing software that works on its rival's chips.

There's no doubt enterprises need help turning their generative AI dreams into reality. But is a software tool developed by a chip maker the answer?

  • Intel and Articul8 are betting that customers will prefer a tool that wasn't developed by a major cloud provider but still gives them the flexibility to use cloud services if they prefer, according to an interview Subramaniyan gave to Reuters.
  • That might be true for companies in highly regulated industries that run hybrid cloud infrastructure, but it might be a tougher sell outside that group.

It's looking like 2024 will be a comeback year for enterprise tech, and as you make your plans for the year please consider sponsoring Runtime and getting your message in front of the more than 20,000 enterprise tech industry leaders and decision makers that receive this newsletter each week. We also plan to roll out several new products next year, including special reports, sponsored content, and events, both virtual and live. If you're interested in learning more, contact us here.

With partners like these…

Broadcom's announcement late last month that it will force VMware channel partners to reapply for the right to resell its software was jarring right off the bat, and little clarity has emerged in the weeks since, according to CRN. It sounds like the hardline stance might drive a fair amount of VMware customers into the hands of competitors like Nutanix, which could really sour Broadcom's return on the $61 billion deal.

"Imagine 50,000 VMware partners with a chip on their shoulder,” one unnamed partner told CRN, summing up the situation inside the community of companies that have worked with VMware for years. Sales orchestrated by channel partners make up a surprisingly large amount of enterprise tech revenue, and usually vendors bend over backwards to accommodate anyone who wants to help them make money.

Broadcom, however, does things a little differently And longtime VMware partners are starting to understand that there's a new sheriff in town: “Our go-forward in the market is, ‘How do you replace VMware?’” Alan Dumas, CEO of Secberus, told CRN. 

Enterprise moves

Justin Hotard is the new executive vice president and general manager of Intel's Data Center and AI Group, joining the company in a crucial leadership role after several years at HPE.

Brett Lindsey is the new CEO of Involta, moving over to the data-center provider after eight years as CEO of Everstream.

Have an executive move to announce? Contact us!

The Runtime roundup

SentinelOne acquired PingSafe, a small security startup specializing in security for multicloud infrastructure deployments.

Can't make it up: Orrick, Herrington & Sutcliffe, a law firm that specializes in dealing with data breaches, was the victim of a data breach that stole personal information belonging to clients who had already been victims of data breaches.

Orca Security cut 15% of its employees, citing "macroeconomic conditions."

The Information had a nice profile of Takeshi Numoto, Microsoft's new chief marketing officer, who developed many of its bundling strategies during its shift to cloud services over the last decade.

Niklaus Wirth, the Swiss computer scientist who invented the Pascal programming language and played a significant role in the advancement of computing in general, died on New Year's Day at 89.

Thanks for reading — see you Saturday!

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to Runtime.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.