Woman working on laptop
Photo: Joe Giddens/PA Images via Getty Images

Who will get to use the federal AI cloud?

Protocol Enterprise

Hello and welcome to Protocol Enterprise! Today: the pros and cons of letting private companies access a federally funded AI cloud, Nvidia adds to its software arsenal and the latest funding for enterprise tech startups.

Spin up

It’s a good time to be a chipmaker, aside from the whole manufacturing shortage problem that continues to persist. Worldwide revenue from chip sales crossed $50 billion in January, according to the SIA, a 26.8% jump over the same month last year.

Silicon Valley wants in on NAIRR

Rima Seiilova-Olson wasn’t sure why she was the only startup founder on a panel full of academics.

“I feel a little puzzled,” said Seiilova-Olson, co-founder and chief machine-learning scientist at a mental health AI startup Kintsugi, talking to Protocol about her participation in a Feb. 16 federal task force meeting about how she might use a federally funded AI research cloud.

The National AI Research Resource, or NAIRR, would be a repository of data and tools for AI research combined with access to the computing power necessary to develop machine learning and other AI systems. But just who will get to use it remains in question.

  • The National Artificial Intelligence Initiative, established by Congress in 2020, envisions the NAIRR as a research hub “for AI researchers and students across scientific fields and disciplines” including from “communities, institutions and regions that have been traditionally underserved.”
  • That legislation, which established the task force planning the NAIRR, does not exclude private-sector researchers, but some believe they should not belong in that community.
  • Researchers from the Stanford Institute for Human-Centered Artificial Intelligence, an influential proponent of the project, have indicated that the government-funded research hub should focus on the needs of academic and nonprofit researchers.

Amid representatives from five colleges and universities, Seiilova-Olson was the lone speaker representing the private sector at the virtual panel discussion addressing the needs of various potential users of the NAIRR.

  • Kintsugi is developing machine-learning models to help detect clinical depression and anxiety based on voice data, and she said that means the company needs access to costly computing power to process large-scale, unstructured data.
  • But in addition to the nuts and bolts of building AI, Seiilova-Olson said the NAIRR also should provide training or community resources for people without access to traditional computer science education.
  • “There’s a big need for small players like myself to benefit from these resources, and I’m not talking about compute power and data,” she told Protocol last week.

Use of the cloud by private-sector researchers would pose legal and logistical issues, as well as distract from its core mission, said Jen King, privacy and data policy fellow at the Stanford Institute.

  • “Overloading this resource out of the gate to address very different sets of users — small business and academic researchers — may jeopardize its development and ultimately its effectiveness,” she told Protocol.
  • “Small-business AI may be expressing legitimate needs and constraints with competing against big AI, but the NAIRR may not be the right solution for addressing them,” King said.

The NAIRR task force already has a private-sector presence, which is by design according to the legislation that established it.

  • Google has said it wants researchers who don’t need computing power from the research cloud to be able to access data in the NAIRR. For one thing, that would ensure that researchers from commercial cloud providers like Google, Amazon and Microsoft would be able to take advantage of the data flowing through the system.
  • Health care data giant Cerner, recently acquired by Oracle, also indicated interest in the NAIRR. The company, which helps health care customers manage patient data and is increasing its use of AI for hospital administration and patient care, emphasized public-private partnerships when it comes to how data in the resource is handled.

Meanwhile, opponents of the project, including some advising AI policy inside the Federal Trade Commission, have raised concerns that the NAIRR will be designed primarily to enable massive-scale AI projects that, by default, would require assistance from big private-sector cloud providers.

Still, there’s no shortage of academics who see value in private cloud providers building and maintaining the NAIRR.

  • “I would strongly urge that the NAIRR would be based on the existing cloud providers in the commercial space,” said Tom Dietterich, distinguished professor emeritus in the Collaborative Robotics and Intelligent Systems Institute at Oregon State University, during the panel discussion.
  • However, said Emily Grumbling, a research staff member with the Science and Technology Policy Institute, “relying on these resources could ultimately increase dependence on the private sector and for-profit-based resources for the AI [research and development] ecosystem.”

— Kate Kaye (email| twitter)


Get inspired as we make the digital economy personal at Adobe Summit 2022. Explore the top industry trends and insights, expand your skills, discover the latest innovations from Adobe Labs, and connect with peers and experts from around the world.

Learn more

Nvidia places another bet on software, and enterprise storage

Long known for its prowess at designing graphics chips, Nvidia’s growing software business has been receiving more attention as of late.

Nvidia said Monday that it had acquired the Israel-based software-defined storage company Excelero, in a bid to add its technology to Nvidia’s enterprise offerings. Excelero is known for its NVMesh tech, which is software that creates a pool of networked solid-state drives that behave like one that is physically attached to a server.

Using software to organize storage drives gives customers the option to optimize them for design factors such as cost or performance. Nvidia said it plans to use Excelero’s team of engineers to help expand support for block storage in its high-performance computing software stack and the software that helps its data processors.

To date, Excelero had raised about $35 million from a range of investors including Qualcomm Ventures and Mellanox, which Nvidia also acquired. Nvidia did not disclose the terms of the deal.

— Max A. Cherney (email | twitter)

Financial corner

Payhawk hit a $1 billion valuation after raising $100 million to help enterprises manage payments and expenses.

Connecteam was valued at $800 millionfollowing a $120 million fundraising round aimed at bolstering its productivity apps for deskless workers.

Insider reached a $1.22 billion valuationafter raising $121 million to expand its marketing data analytics product.

Apollo.io is worth $900 millionafter raising $110 million for sales intelligence and enablement.

Luminous Computing raised $105 millionfrom Bill Gates and other investors to scale up production of its light-based AI chip.

— Aisha Counts (email | twitter)

Around the enterprise

Three security companies — Cloudflare, CrowdStrike and Ping Identity — teamed up to launch the Critical Infrastructure Defense Project, a free package of security services for potential targets like hospitals and utility companies.

Weeks after Mandiant was said to be in acquisition talks with Microsoft, The Information reported that Google has begun similar talks to acquire the security company.

Ukraine asked AWS to stop doing business with Russian companies.

Intel filed an S-1 registration statement for its Mobileye self-driving car division, but did so confidentially, which is boring.

A data center fire in Iran took out internet service across much of the country last Friday.


Get inspired as we make the digital economy personal at Adobe Summit 2022. Explore the top industry trends and insights, expand your skills, discover the latest innovations from Adobe Labs, and connect with peers and experts from around the world.

Learn more

Thanks for reading — see you tomorrow!

Recent Issues