Nvidia's $25 billion bond sale; Salesforce buys Fin for $3.6 billion

Nvidia sold $25 billion of investment-grade notes across seven tranches maturing as far out as 2056 — its first corporate bond deal since a $5 billion raise in 2021. Demand reached more than $85 billion, over three times the offering, and the company upsized the sale from an initial $20 billion. For a firm sitting on a roughly $5 trillion market value and ample cash, tapping the debt market signals just how much it plans to spend building out AI data-center capacity.

Salesforce agreed to acquire Fin — the AI support agent from the company formerly known as Intercom — for about $3.6 billion, with the deal expected to close in early 2027. Fin resolves customer questions end-to-end across chat, email, WhatsApp, SMS, phone, and Slack, and Salesforce will fold its technology and team into Agentforce, the platform it sells for building enterprise AI agents. The price is one of the largest yet for a pure-play customer-service agent, a sign of how aggressively incumbents are buying their way into agentic AI.

Bengaluru-based Sarvam raised $234 million at a $1.5 billion valuation, led by IT services giant HCLTech, which put in $150 million. Sarvam builds full-stack AI tuned for Indian languages — spanning model training, inference, and enterprise apps across banking, insurance, government, and defense. The timing is pointed: it lands days after Anthropic cut India off from its newest models on US government orders, sharpening the case for home-grown frontier AI in a market that consumes far more AI than it builds.

Loft Orbital says its YAM-9 satellite became the first spacecraft to run a vision-language model in orbit, using Google DeepMind's Gemma 3 to answer plain-English queries — classifying where wilderness meets development, or spotting infrastructure near rail hubs — without beaming raw imagery to ground analysts first. The model ran on an Nvidia Jetson edge chip paired with NAVI-Orbital software from NASA's Jet Propulsion Laboratory. Onboard triage like this could sharply cut how much data satellites need to downlink, and Loft says 50 to 100 such craft would give near-real-time eyes on Earth.

Meta rolled out AI Mode on the Facebook mobile app for US users, letting people ask questions in plain language and get answers synthesized from public posts across Facebook, including Groups and Reels. It mirrors Google's AI search overhaul and follows Meta's Reddit-style Forum app from May. Pulling answers from everyday user posts rather than vetted sources raises the same reliability worries that have dogged AI summaries elsewhere — outdated or misleading information can slip through — even as Meta bundles more AI into the app to keep users engaged.

In this week's Stratechery analysis, Ben Thompson argues that Anthropic has turned safety into a "superpower" by framing self-serving moves — restricting access, retaining user data, limiting who can build frontier models — as safety imperatives. He reads the US government's shutdown of Fable 5 and Mythos 5 as an inevitable collision between a lab convinced only it can be trusted with powerful AI and a state unwilling to cede that judgment. His unsettling conclusion: a team that genuinely believes its own safety story may be harder to check than a cynical one.