Why AI companies are cozying up to federal agencies, and more tech news this week

Welcome back to LinkedIn News Tech Stack, which brings you news, insights and trends involving the founders, investors and companies on the cutting edge of technology, by Tech Editor Tanya Dua . You can check out our previous editions here .

Pitch me the interesting investors, founders, ideas and companies powering emerging technologies like AI to reach the inboxes of nearly 1 million subscribers plus thousands more on LinkedIn. Follow me for other tech updates.

A deep dive into one big theme or news story every week.

Elections aside — another story has quietly flown under the radar this week signaling a new era for the federal government, particularly as it relates to AI. Companies including Meta , OpenAI and Scale AI have all inked partnerships with a handful of government agencies in recent days that aim to harness AI for national security purposes.

Scale AI launched the large language model Defense Llama, which it says is “purpose-built for American national security” and was trained on “military doctrine, international humanitarian law” and Department of Defense guidelines. Defense Llama was built on top of Meta’s AI model Llama 3, and followed the latter’s announcement on Monday that it will open up its Llama LLM to private-sector defense contractors such as Lockheed Martin.

Meanwhile, OpenAI unveiled its first ChatGPT Enterprise partnership with the Defense Department’s Air Force Research Laboratory, which will be used on its unclassified systems and data.

The burst of activity comes as AI adoption has soared , and defense innovation has become a bigger priority for the U.S. and venture capitalists alike, amid heightened geopolitical tensions and rapid technological advancements. Over $70 billion was deployed into the sector in 2022 and 2023 — a number only poised to grow with a proposed $850 billion defense budget for 2025, including a significant allocation for research and development, according to J.P. Morgan research .

“What you’re really seeing is the beginning of the adoption cycle of a new tool in the computer science toolbox, in the same way you’re seeing in enterprises,” Seth Winterroth , a partner at the VC firm Eclipse Ventures, told LinkedIn News. “One that can help drive greater efficiency, higher velocity, reduction of costs and better one-to-one engagement.”

For Point72 Ventures ’ Sri Chandrasekar , who previously spent half a decade running an AI-focused lab for the CIA’s strategic arm, IQT (In-Q-Tel) , the recent government embrace of AI mirrors its rapid adoption of cloud over a decade ago.

“Everyone now understands that the majority of compute that people are going to pay for going forward is AI processing,” he said. “Everybody in the AI business wants to make sure that their models are the ones being delivered, because that’s what will be the next generation of infrastructure layer for the government.”

AI is indeed becoming a priority for the agency, with Pentagon leadership forming a temporary task force to accelerate the DoD’s grasp, oversight and deployments of generative AI last year. More broadly, the department and other agencies also received new directives on how to pursue the burgeoning technology in the National Security Memo on AI issued last month.?

This is not a surprise as government agencies seek to enhance operational efficiencies and combat growing cybersecurity concerns, Matt Johnson , a data system engineer at the US Army, wrote on LinkedIn : “AI's potential for efficiencies in threat detection, cybersecurity and logistics is increasingly recognized by government agencies, making it a valuable asset in decision-making.”

In addition to the government’s growing appetite, public-sector budgets represent a lucrative opportunity for AI companies looking to scale on the back of lofty valuations. Such partnerships, once set in motion, offer these companies sustained, locked-in revenue potentially amounting to hundreds of millions of dollars annually, Chandrasekhar said.

“The government is the largest IT spender in the world — so once the government makes a big decision, those tend to be decade-long decisions,” he said.

For companies like Scale, government contracts offer more than just revenue — providing the company an avenue to branch out, Chandrasekhar added.?

“Scale has a $7 billion dollar-plus valuation, mostly off government revenue,” he said. “They’re trying to move up-market from being a data-labeling company to a full-service AI provider.”

While the current AI partnerships are primarily being deployed for military and national security purposes, the scope could eventually extend to broader public services such as transit systems and benefits programs, according to both Chandrasekhar and Winterroth.

“Think about veterans support, or Medicare or Medicaid,” Winterroth said. “We’re limited in our ability to give world-class support by the number of humans in operation. That’s going to change materially, and the quality of service and support will go up.”

Still, public-sector AI initiatives come with their own set of unique challenges and requirements, with a greater onus on the AI companies to build relevant frameworks to ensure responsible and transparent AI use in high-stakes environments. This includes higher standards for oversight, regulatory adherence and data privacy — with higher stakes, especially national security and highly sensitive data.

“Safety, explainability and training is what these agencies would care the most about before putting any AI models into production,” said Chandrasekhar. “Most enterprises don’t think about what data went into building Llama 3 or GPT4, but if those models have been trained off proprietary information that they shouldn't have had access to, then that’s a problem. I don't think the government wants to be in the business of being sued by private companies.”

While the current AI push comes as regulators and policymakers have been weighing the technology’s risks, a new Trump administration next year means that the trend will likely continue to accelerate.

As Ayman Nadeem , founder and CEO of the startup Nuanced, wrote, “Trump’s ‘Make America First in AI’ plan focuses on speeding up innovation and reducing regulations, especially in military tech. While this could accelerate development, it might mean fewer checks and balances.”

See what other national security professionals are saying about the federal government’s AI push in this LinkedIn post — and join the conversation.

Here’s where we bring you up to speed with the latest advancements from the world of AI.

• Nvidia becomes the world's most valuable company. Overtaking 苹果 , 英伟达 is officially the world's largest company. The company's shares, which have risen more than 850% since the end of 2022, rose 2.9% Tuesday, resulting in a market capitalization of $3.43 trillion, versus Apple's $3.38 trillion. The shift demonstrates "just how dominant artificial intelligence has become on Wall Street," says Bloomberg . Bar Apple, Nvidia counts most of the tech giants as its clients. The company is set to report earnings later this month; its second-quarter earnings more than doubled from a year earlier, beating Wall Street estimates and signaling resilient demand from Big Tech customers. In related news, Nvidia will replace struggling Intel in the Dow Jones Industrial Average, a symbolic changing of the guard among semiconductor giants.
• OpenAI explores for-profit shift. OpenAI is in talks with the California attorney general's office to become a for-profit business, Bloomberg reports , citing unnamed sources. But restructuring could be a lengthy process, with regulators probing the value of OpenAI's "highly lucrative" intellectual property, including ChatGPT. The company has also discussed the matter with the attorney general of Delaware, where it's incorporated . OpenAI didn't comment on the discussions specifically, but told Bloomberg that "any potential restructuring would ensure the nonprofit continues to exist and thrive." The company faces questions about whether its "do-gooder" mission would be jeopardized by turning for-profit. In related news, OpenAI has acquired the domain chat.com , which now redirects to ChatGPT, TechCrunch said , citing an email from the startup. The domain name was previously owned by HubSpot founder Dharmesh Shah , who said in a LinkedIn post that he bought it "for $10M+" and "sold it for more than I paid." A domain sales database shows that it was purchased for $15.5 million in March 2023, which aligns with the timing of the post, according to The Verge .
• Perplexity valuation set to triple. Perplexity is set to raise $500 million in a new funding round that would value the AI company at $9 billion, The Wall Street Journal reports , citing unnamed sources. The increase — and fourth funding round this year — would triple Perplexity’s valuation from as recently as June. Perplexity, which uses online data to answer questions , faces legal challenges over its use of news content, along with growing competition from OpenAI and 谷歌 . Investors, which include Jeff Bezos and 英伟达 , appear undeterred, and the startup expects to generate $50 million in revenue this year.
• Super Micro hits more turbulence. Shares of Supermicro Computer have taken a dive after the tech manufacturer issued a lower-than-expected sales forecast , while also announcing that it does not know when it will be able to file its delayed annual financial report. The company has been reeling ever since questions were raised about its internal controls earlier this year. Super Micro's auditor, 安永 , resigned last week , saying it was "unwilling to be associated with the financial statements prepared by management." In a statement Tuesday, Super Micro said that a board committee investigation had found "no evidence of fraud or misconduct."
• Prime Video adds AI recaps. 亚马逊 ’s Prime Video is introducing a new generative AI feature that offers detailed descriptions of TV episodes and seasons, down to specific plot points and conversations. The new "X-Ray Recaps" will include guardrails to keep them spoiler-free, Amazon says , and will be customized to the moment you’re watching. Currently in beta, they're now available to users of Amazon’s Fire TV connected devices in the U.S., with wider access coming by the end of the year. It comes just after Google TV announced its own AI summaries . Meanwhile in TV news: Extremely large screens are selling like crazy as costs drop, per Circana . See Amazon exec Mike Hopkins ’ LinkedIn post here.
• AI spurs Palantir's record profit. Palantir Technologies reported record quarterly profit this week, driven by “unrelenting” demand for its AI products and services. Its shares surged more than 20% on Tuesday. Revenue soared 30% to $725.5 million, exceeding expectations. The company, whose clients include businesses as well as governments, is benefiting from the tech sector’s AI boom as it rolls out new software tools and services that leverage data analysis. About 70% of its revenue came from U.S. customers in the quarter. Palantir's stock, which was added to the S&P 500 in September, has advanced more than 140% this year.

Here’s a list of other notable AI developments from this week:

• Robotics "brain" startup Physical Intelligence has raised $400 million at a $2.4 billion valuation, backed by OpenAI , Thrive Capital , Lux Capital and Jeff Bezos.
• Canadian AI startup Cohere has announced a new embedding model called Embed 3 , designed for semantic search and to help enterprises build LLM applications. Embedding models help transform various types of unstructured data into numerical representations, which is essential to build generative AI applications.
• A new AI startup called Hopscotch Labs announced that it has raised a pre-seed round co-led by Alpaca VC and Lightspeed , with participation from Betaworks and Next Wave NYC .
• Data-center builder Crusoe is on track to raise as much as $50 million, led by Peter Thiel's Founders Fund , according to the Financial Times.

Catch up on the tech headlines you may have missed this week and what our members are saying about them on LinkedIn.

• Amazon CEO defends RTO mandate. After weeks of backlash to 亚马逊 's announcement that it will soon return to full-time in-office work, CEO Andy Jassy addressed employee concerns in a company meeting Tuesday, according to multiple news outlets citing transcripts of the remarks. In the meeting, Jassy denied accusations that the return-to-office mandate was intended as a "backdoor layoff" or coordinated with city officials, saying: "This was not a cost play for us. This is very much about our culture and strengthening our culture." Last week, more than 500 Amazon Web Services employees signed a letter blasting the policy .
• Berkshire unloads more Apple shares . Warren Buffett's Berkshire Hathaway sold more Apple stock in the third quarter, offloading about 25% of its 400 million shares, according to the conglomerate's latest quarterly report . Apple is still Berkshire's biggest holding, valued at nearly $70 billion. Berkshire also reduced its position in 美国银行 , and grew its cash pile to a record $325 billion. Buffett suggested at the company's annual meeting earlier this year that it planned to sell Apple shares for tax reasons . Apple reported an increase in iPhone revenue in its recent results, but faces challenges including weakness in China .
• Speaking of Apple, the tech giant is eyeing a move into smart glasses. 苹果 is running an internal study of similar products currently on the market, Bloomberg reports , citing unnamed sources. The effort, code-named Atlas, is meant to gather feedback that can inform Apple's future decisions on glasses features. Rival Meta's $299 smart glasses have proven to be a hit, while Apple's previous attempt at a face-worn device, the Vision Pro, has faltered , thanks in large part to its hefty $3,499 price tag.
• Lyft is teaming up with three autonomous-vehicle companies in a bid to compete with Uber 's AV momentum. Lyft is working with May Mobility and Mobileye to launch autonomous vehicles on Lyft's network. It's also partnering with dash-cam company Nexar Inc. to help create safer AV technology. May Mobility plans to roll out AVs on Lyft's platform in Atlanta starting next year. The robotaxis will initially use safety drivers, before transitioning to fully driverless technology. Lyft shares rose on the news, as well as better-than-expected sales numbers from its third-quarter earnings report on Wednesday.
• How Big Tech lost its perks culture. As the tech landscape shifts from one of abundance to relative scarcity, so too has its culture of employee perks. What began as quality free food at Google in the early 2000s eventually morphed into vouchers that could — and were — used for anything from wine glasses to hard cash. But as companies such as Meta began cutting costs post-pandemic, the permissiveness has given way to scrutiny. One example is its recent firing of employees found to be using free GrubHub vouchers for nonfood items, yet the change is industry-wide as companies compete less for workers.

Here’s keeping tabs on key executives on the move and other big pivots in the tech industry. Please send me personnel moves within emerging tech.

• OpenAI has hired former Twitter and 谷歌 executive and Pebble co-founder Gabor Cselle , he shared on LinkedIn .
• OpenAI has also snagged Caitlin (CK) Kalinowski ???? , Meta ’s former head of AR glasses hardware to oversee robotics work and partnerships as it ventures into hardware, she shared on LinkedIn .
• You.com has hired former IBM , LinkedIn and 1Password executive and GrokAI founder Chris Mann as head of enterprise AI products, he shared on LinkedIn .
• Netflix has appointed Mike Verdu , formerly its VP of games, as VP of GenAI for games, he shared on LinkedIn .
• Sprinklr has appointed former Dell Technologies and AMD executive Rory Read as president and CEO, outgoing CEO Ragy Thomas , who is taking on the role of chairman, shared on LinkedIn. Also see Read’s LinkedIn post here.
• eBay has hired former Ahold Delhaize and SymphonyAI executive Jateen Patel as chief analytics officer, chief growth officer Julie Loeger shared on LinkedIn .
• Cloud analytics and AI data platform 天睿 has promoted Louis Landry to chief technology officer, he shared on LinkedIn .

As always, thanks for reading. And click 'Subscribe' to be notified of future editions.