Crusoe wants to be the enterprise AI concierge

Welcome to Runtime! Today: Why Crusoe thinks enterprises want a catered AI experience, Jensen Huang tanks quantum -computing stocks, and the latest enterprise moves.

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

Package deal

As we discussed on Tuesday, right now a lot of businesses that want to build generative AI applications are struggling to get over the finish line and into production. The simple truth is that generative AI is still an evolving technology that requires early adopters to jump through a lot of hoops, and the people who understand how to get it right aren't nearly as evenly distributed across the corporate world as the people who know how to build traditional software applications.

After Crusoe raised $600 million in new funding last month, much of the attention focused on its unique approach to sourcing the energy required to power its GPU infrastructure cloud. But its plan for generating a return on that investment involves building a curated suite of cloud AI software and services for specific markets such as (for example) energy, agriculture, or finance that could help solve some of those last-mile problems, said Nadav Eiron, senior vice president for cloud engineering, in an interview this week.

Some quick background on Crusoe: the company was founded in 2018 to mine bitcoin using data centers built next to oil and gas fields in order to harness the energy from flared gas, which is a byproduct of energy production that releases methane into the atmosphere.
Like CoreWeave, Crusoe eventually saw more value in renting those GPUs for training AI models and has been building an infrastructure cloud service since 2022.
But researchers or companies interested in GPU infrastructure services have lots of choices, including (of course) The Big Three, and while their salespeople would insist otherwise there's not a whole lot of difference between renting an Nvidia GPU from one provider or another.

Crusoe is now in the process of expanding its "vertically integrated" approach to data-center construction and energy sourcing to software. "We believe we've built a great IaaS product, [but] we also realize that is only one rung on that vertical ladder of needs, and that the more interesting and the more innovative stuff comes further up the stack," Eiron said.

Cloud infrastructure providers (Eiron spent 18 years at Google) started taking a similar approach with late arrivals to the cloud over the last several years, creating packages of software and services designed to meet the unique needs of unique industries.
Early adopters excited by a new technology like to forge their own path and want all the knobs and levers that vendors can give them, but other companies can become easily overwhelmed with options.
Eiron thinks those companies are looking for help: "As you go further up the stack, you find more and more customers that are focused on their product with AI being an enabling technology, and not the thing they're really interested in," he said.

Even with a fresh $600 million in the bank, Eiron knows Crusoe can't build all that software itself. The company plans to partner with other startups working on AI software for vertical industries that "sometimes have difficulty getting the attention of the Big Three," he said, and it will start rolling out those market-focused services over the course of 2025.

Crusoe will be competing with some of the biggest companies on the planet, companies that are under pressure to show returns on the billions they have invested in generative AI infrastructure and services.
But the startup thinks it has an edge with its energy strategy — which could appeal to businesses looking for a cleaner source of computing power — and, to paraphrase former AWS CEO Adam Selipsky, believes that demand for higher-level generative AI infrastructure services is just getting off the ground.
"It didn't make sense two years ago because it was too soon. It might not make sense in three years because it'll be too late. But I think there is an opportunity right now," Eiron said.

No solace for quantum

In the minds of Wall Street professionals, Nvidia CEO Jensen Huang is arguably tech's greatest soothsayer entering 2025 thanks to the meteoric rise in Nvidia's stock price since the dawn of the AI boom. After this week, executives at quantum computing companies are hoping he keeps his mouth shut.

The price of shares in several prominent quantum computing companies — including D-Wave, Riggeti, and IONQ — fell sharply Wednesday after Huang quite sensibly pointed out at CES that we're probably two decades away from the arrival of quantum computers that could change the tech industry pecking order. His "grim but frankly realistic outlook on the future of quantum computing," as The Register put it, continued to impact those companies during Thursday's trading session.

Huang is no neutral observer, of course, given the enormous demand for his company's chips among those trying to push the leading edges of classical computing. But the companies trying to push quantum computing have done themselves no favors in recent years by overhyping every incremental advance as an imminent breakthrough, and a little radical honestly is probably good for everyone.

Enterprise moves

Fazal Merchant is the new president and chief financial officer of Wiz, joining the IPO-bound security company after three years as co-CEO of Tanium.

Francis deSouza is the new chief operating officer at Google Cloud, following decades of enterprise operational positions at Illumina and Symantec.

Kaylin Voss is the new executive vice president of Agentforce at Salesforce, after a year as chief revenue officer of the company's Slack division and six years of other leadership positions at the company.

Lisa Luscap is the new chief marketing officer at Starburst, joining the data lakehouse company following similar roles at Pluralsight and Snowflake.

Eric Wilcox is the new chief revenue officer at Precisely, a promotion from his previous role as senior vice president of sales, Americas.

Rehan Sheikh is the new vice president for global silicon chip technology and manufacturing at Google Cloud, following almost four years in a similar role for Microsoft.

The Runtime roundup

Ivanti disclosed that a zero-day vulnerability is currently being exploited across several of its enterprise VPN products, and while a patch is available for its Connect Secure product patches for Policy Secure and ZTA Gateways aren't expected to arrive for almost two weeks.

Microsoft laid off "less than 1%" of employees — including an unknown number of people working on security — for performance-related reasons this week, according to CNBC, but that's still a sizable number given that more than 220,000 people work at the company.

Arm is thinking about acquiring Ampere, one of the most prominent designers of Arm-based server chips, according to Bloomberg.

Thanks for reading — see you Saturday!