| Here’s a look at today's AI briefing: - Google DeepMind's new AlphaFold can model DNA and RNA.
- U.S. considering rules to restrict China's access to AI models.
- Microsoft to invest $3.3B in Wisconsin for AI infrastructure.
- Report: Employees keeping AI use hidden at work.
- Arm forecasts only modest revenue outlook amid AI boom.
Beth
p/beth-duckett | |
| 1 | | Google Deepmind announced AlphaFold 3, the third version of its AI model that could revolutionize disease research and treatments. AlphaFold 3 not only predicts protein foldings, as prior versions have, but also interactions with other biological molecules including DNA, RNA, and small molecules. More: - DeepMind's existing AlphaFold program predicts a protein's 3D structure.
- AlphaFold 3, developed by DeepMind and sister lab Isomorphic Labs, can also map behaviors for all life's molecules, including human DNA.
- With the new capabilities, DeepMind can design molecules that bind to specific protein sites and predict the binding strength.
- This could speed up and reduce the costs of developing transformative treatments in areas like drug design, genomics research, and biorenewable materials, DeepMind said.
- AlphaFold 3, described in a new Nature paper, is free for researchers to use for non-commercial research.
Zoom out: - DeepMind also launched the "AlphaFold server," a free online tool enabling scientists to validate hypotheses before conducting physical experiments.
- The new server allows researchers to create biochemical models without needing to install the system from open-source code, unlike earlier versions.
|     | |
| 2 | | The U.S. government is considering new rules to restrict China's access to advanced proprietary AI models, Reuters reports. The Commerce Department could ban exports of closed-source AI models, such as OpenAI's ChatGPT and Google's Gemini, though such a rule is still far from being finalized. More: - The rule could target China as well as Russia, North Korea, and Iran, a source told Reuters.
- To decide which AI models would be subject to the rule, the U.S. could rely on a threshold from Biden's AI executive order.
- The threshold is based on the amount of computing power used to train an AI model, sources said.
- No AI models are thought to have reached that level yet, so the rule would likely only affect future models.
Zoom out: | | | |
| 3 | | Microsoft is building an AI data center on the site of a failed Foxconn factory in Wisconsin. Microsoft said it will invest $3.3B by 2026 in AI and cloud infrastructure at the site. More: - Taiwan electronics manufacturer Foxconn was set to build an LCD factory at the site starting in 2018.
- The plans, touted heavily by then-President Trump, were later downsized and mostly unrealized.
- Microsoft bought a portion of the site for $50M in 2023.
- Its new data center campus will create 2,300 union construction jobs and 2,000 permanent jobs once built, the White House said.
Zoom out: - Microsoft said it also will develop an AI innovation lab at the University of Wisconsin-Milwaukee and a data center academy at the state's Gateway Technical College.
- The company will work with community partners to give generative AI training to more than 100,000 people in Wisconsin by 2030.
| | | |
| 4 | | 75% of all knowledge workers currently use AI at work, and most are using personal AI tools that are hidden from their employers, according to a Microsoft-LinkedIn report. 
What happened: LinkedIn and parent company Microsoft released their 2024 Work Trend Index Annual Report, which focuses heavily on AI. What the numbers show: The report shows three-quarters of knowledge workers, those whose main capital is knowledge, report using AI in their jobs. Nearly half started to do so within the last six months. Users report that AI helps them save time (90%), focus on critical tasks (85%), improve creativity (84%), and increase job satisfaction (83%). Zoom in: Nearly 80% of these AI users "are bringing their own AI tools to work," meaning they are "keeping AI use under wraps," according to the report. Knowledge workers of all ages use personal AI tools at work, from 73% of Boomers to 85% of Gen Z. Meanwhile, 52% of AI users at work hide its use on key tasks; 53% fear it makes them seem replaceable. What it means: The results raise concerns over employers' cybersecurity and privacy. Using AI tools like ChatGPT secretly at work could risk company data, according to Microsoft and LinkedIn. | | | |
| 5 | | Arm shares fell 7% today as the chip designer forecasted only modest revenue growth, disappointing investors who anticipated greater benefits from the AI boom. Amazon, Microsoft, and Alphabet use Arm's technology for their custom server farm chips. More: - While Arm's chip designs are used in most smartphones, the company has been pushing further into data center hardware needed for AI.
- Customers use Arm-based central processors that work alongside Nvidia chips for AI tasks in data centers.
- However, the British designer's full-year revenue forecast on Wednesday fell short of some analysts' expectations.
- The outlook differed from Arm's February earnings report, when Arm reported record sales and forecasts that eclipsed analysts' projections.
Zoom out: - In February, Arm wrote that the "most demanding AI applications are already running on Arm today."
| | | |
| 6 | | Fresh funding: - U.K. self-driving startup Wayve raised $1.05B in funding. SoftBank Group led the Series C round, with participation from Nvidia and Microsoft. Wayve is building a self-learning rather than ruled-based system for autonomous vehicles.
- French generative AI startup Holistic AI closed its first tranche of an initial funding round at $200M, which comprises $80M in equity and $120M in convertible debt. Accel Partners, UiPath, and Google’s former CEO Eric Schmidt participated in the round, which valued the company at $370M.
- Computer vision and AI-powered property inspection technology startup Honeycomb Insurance raised $36M in Series B funding. Ben-Zaken led the round, joined by Arkin Holdings, Launchbay Capital, Ibex Investors, Phoenix Insurance, and IT-Farm.
- Automotive AI solutions startup Stella Automotive SA bagged $19.7M in Series A funding from Reynolds and Reynolds, OREMOR Automotive Group, Mills Automotive Group, Flow Automotive, MBB Auto Group, Ford Family Investments, and other investors.
For the latest on AI startup funding and more in the VC community, subscribe to our Inside Venture Capital newsletter. | | | |
| |
 | AI and technology writer | | Beth is a contributing editor and writer of Inside's AI and Tech newsletters. She has written for publications including USA Today, the Arizona Business Gazette, and The Arizona Republic, where she received recognition with a Pulitzer Prize nomination and a First Amendment Award for collaborative reporting on state pension cost increases. You can reach her at Beth.Duckett@yahoo.com. | | This newsletter was edited by Beth Duckett | |
| |
| |
RE:
RE ALL:
FWD:
