Your evening roundup of AI news for November 2. Today's story is all about the movement of money. Marvell announced the acquisition of Lumentum at an enterprise value of $115 billion, pushing the consolidation of the components industry that underpins AI data centers one step further. Saudi Aramco also announced an investment in a state-backed AI company, accelerating the flow of Middle Eastern money. On the regulatory and social front, New York's new AI law and an investigation into the reality of deepfake apps stand out.
Marvell to Acquire Lumentum for $115 Billion — A Reshuffle in AI Connectivity
On November 2, Marvell Technology announced that it will acquire Lumentum Holdings, a data center optics company, at an enterprise value of $115 billion. In the all-stock deal, Marvell shareholders are expected to own 67% of the combined company. Bloomberg called it "one of the largest technology mergers in history."
Marvell's core business is data center semiconductors, while Lumentum specializes in optical communications devices. As AI training and inference move onto massive clusters, communication between servers—and between data centers itself—has become the bottleneck. It is a merger that plants the flag of "AI connectivity," bringing electrical silicon and optical devices under one roof, aiming to differentiate in a market where supply is dominated by NVIDIA.
Saudi Aramco Takes a 49% Stake in State-Backed AI Firm "Humain" for $2.6 Billion
Saudi Aramco, Saudi Arabia's state-owned oil company, is reportedly acquiring a 49% stake in Humain, the country's state-backed AI company, for $2.6 billion. Humain was founded to develop Saudi Arabia into a global AI hub and is regarded as the core of a national strategy to channel oil revenue into AI investment.
Major U.S. AI companies have been moving to procure power and capital from the Middle East, and now a new flow has emerged: Middle Eastern oil money becoming shareholders in AI companies. The back-and-forth of capital and energy looks set to become even more active.
New York's New AI Law — Will the Day Come When ChatGPT Demands Your ID?
An analysis in U.S. media is drawing attention to new AI-related legislation taking shape in New York State. The argument: through requirements such as age verification, the use of AI services like ChatGPT could effectively require proof of identity, potentially bringing online anonymity to an end in practice.
How to implement age verification for AI services is a puzzle regulators in every country are grappling with, and reconciling it with privacy remains a challenge. We will be watching closely how state-level regulation in the U.S. changes how platforms respond.
WIRED Investigation — Sexual-Abuse Deepfake Features Lurking in 400 "Ordinary AI Apps"
WIRED reports that it has uncovered "hidden features" capable of generating violent and sexual AI images and videos in roughly 400 apps and websites that look like ordinary AI tools. Many of the services are said to be linked to developers in the Chinese-speaking world.
The tactic of slipping past app review with innocuous front-facing features while hiding problematic ones inside makes detection difficult for store operators. Countering deepfake harm seems likely to extend beyond legislation to the management of distribution channels.
Why Is Postgres 19 Late? — An Analysis That Blames AI
LWN.net has published an analysis titled "Postgres 19 is late, and it's AI's fault," which is making waves in the developer community. It argues that behind the delayed release schedule lies the change that the spread of AI is bringing to open-source development.
As AI-generated code and patches pour into development workflows, how will the burden of review and maintenance change? This is becoming a shared theme across major open-source projects, not just Postgres.
Roblox Welcomes It, Steam Is Wary — Opinions Split on AI-Generated Games
According to PC Gamer, Roblox takes a positive view of the rise of AI-generated games, highlighting that 70,000 games were published on the platform in the last five months alone. Steam, on the other hand, has reportedly expressed concerns over the quality of AI-generated content.
Even under the same wave of AI-generated content, "volume" means something different on a UGC platform than on a game store. Does a flood of content drag down the quality of the experience, or expand the options? Platforms' differing stances seem likely to influence where developers go from here.
A 25-Year Payoff — Berners-Lee's "Semantic Web" and the Agentic Web
MIT Technology Review has published an article titled "The agentic web is here." It argues that the "Semantic Web" — the vision that Tim Berners-Lee, inventor of the WWW, laid out in Scientific American in 2001: an internet where autonomous software agents handle our requests on our behalf, without going through browsers or search engines — is finally becoming a reality after 25 years.
The article draws parallels between the vision of that era and today's AI agents, including an anecdote in which a brother and sister have an agent book a physical therapy appointment for them. As a web where agents interact with agents becomes the norm, the design philosophy of that era may well be worth revisiting.