AI This Week

Anthropic said it has closed a $30 billion funding round at a $380 billion post-money valuation, more than double its value from its last raise in September. The Claude maker now sits behind OpenAI’s record private round, as top AI labs keep pulling in massive checks to cover the soaring cost of training models and buying compute, including Nvidia GPUs. Coatue and Singapore sovereign wealth fund GIC led the round, with D. E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQ and MGX joining. Anthropic said the financing includes part of previously announced commitments from Microsoft and Nvidia. The company reported $14 billion in annualized revenue and said the cash will fund infrastructure, research, and enterprise products.

Elon Musk’s artificial intelligence company xAI lost a second co-founder in two days after influential researcher Jimmy Ba announced on X that he has departed. Ba, a University of Toronto professor, wrote that he was “grateful to have helped cofound at the start.” His exit followed Tony Wu’s departure a day earlier. The departures add to a longer list of early leaders who have left xAI, including Igor Babuschkin, Kyle Kosic, and Christian Szegedy, while Greg Yang said last month he would step back to focus on treatment for Lyme disease. The moves arrive as xAI faces regulatory probes in Europe, Asia, and the U.S. tied to Grok tools used to create non-consensual explicit images.

ByteDance has suspended a Seedance 2.0 feature that could generate a “personal voice” from a person’s facial photos, citing potential risks. The move puts a sudden pause on a capability that blended face-based identity signals with voice creation, a combination that can raise privacy, consent, and impersonation concerns. The company has not detailed when the feature will return, what safeguards will change, or whether the suspension applies across all markets and products tied to Seedance. The decision lands as tech firms face rising scrutiny over synthetic media, including tools that can mimic real people in audio and video.

Artificial intelligence emerged as one of the Super Bowl’s biggest ad themes, with AI-related messaging accounting for about 23% of commercials aired during the game. Brands across categories used the broadcast to link products and services to AI, betting that mainstream audiences would respond to promises of smarter tools, faster results, and new capabilities. The spike also created a crowding effect. As more advertisers used similar language and claims, it became harder for any single message to stand out, and viewers received a rush of overlapping AI pitches in a short window. The surge put pressure on marketers to explain AI’s role clearly, quickly, and without technical phrasing, while competing for attention in the most expensive media slot of the year.

‍

Anthropic unveiled Opus 4.6, calling it its most capable model for enterprise and knowledge work. The release follows new Claude Cowork plugins aimed at productivity, legal, sales, and marketing tasks, pushing the company deeper into territory long owned by software vendors. Investors have already punished that sector amid concerns AI platforms will replace some software-as-a-service tools; Salesforce is down 25% this year, SAP 18%, Intuit 32%, and Thomson Reuters 30%. Anthropic says Opus 4.6 tops major benchmarks, beating OpenAI’s GPT-5.2 and Google’s Gemini 3 Pro, and returns more “production-ready” first drafts for documents, spreadsheets, and presentations. It can run agent tasks across dozens of tools and analyze filings and market reports.

‍

AI voice startup ElevenLabs said it raised $500 million at an $11 billion valuation, tripling its $3.3 billion mark from a $180 million Series C in January 2025. Sequoia Capital led the round, joined by Andreessen Horowitz, Iconiq, Lightspeed, Evantic Capital, and Bond; cofounder Mati Staniszewski said Nvidia invested in September. Founded in 2022 and based in London, ElevenLabs moved beyond text-to-speech into speech-to-text, sound effects, dubbing, music and conversation. It sells voice and chat agents for enterprises and tools for brands and creators, counting Time, Nvidia, Meta, and Salesforce as users. The company said it ended 2025 above $330 million in annual recurring revenue and is building toward an IPO.

Fitbit co-founders James Park and Eric Friedman have launched Luffu, a new app built around AI that pulls together a family’s health data and keeps a running watch for changes. The company says the system can spot trends across records, wearables, notes, and uploads, then flag potential issues and surface follow-up questions for a caregiver. It also works like a chat assistant, answering health-related prompts in everyday language while drawing from the household’s stored information. Luffu supports kids, adults, aging parents, and even pets, aiming at the person who manages appointments, medications, and paperwork. Users can add details with voice, text, or photos, and the app connects with Apple Health and Fitbit. Access starts via a waitlist.

Leaks point to a new Anthropic model branded Claude Sonnet 5, with an internal date string set to February 3, 2026. Whether that marks a public release or an internal milestone remains unclear, but it lands during the week of Super Bowl LX on February 8, as AI labs push for consumer attention against ChatGPT and Google’s Gemini. Early hands-on testing describes a non-thinking Sonnet 5 variant that looks competitive on math with frontier models and, in some workflows, produces stronger code than Claude Opus 4.5. A standout demo involved structured visual output, where an ASCII world map prompt generated an unusually complete, detailed result, alongside UI and rendering-focused tasks. The tested build lists a 128k context window.

Elon Musk’s SpaceX is acquiring his AI start-up xAI, the company behind the Grok chatbot, in a move that consolidates AI, rockets, space-based internet, and media under one roof. The companies did not disclose terms, but a source familiar with the deal said it values xAI at $125bn and SpaceX at $1tn, a level that would make SpaceX the world’s most valuable private company. Musk framed the merger as an “innovation engine,” with a near-term focus on launching AI satellites and a longer-term plan for space-based data centers to meet AI’s energy and compute demands. The deal follows Tesla’s $2bn investment in xAI and arrives as SpaceX considers an IPO.

OpenAI said it will retire several models from ChatGPT next month, including GPT‑4o, the model some paid users favored for its warm conversation style. The company launched GPT‑4o in May 2024 and faced backlash in August after briefly removing access following the release of GPT‑5, then restoring it and promising advance notice before any permanent retirement. OpenAI now says only 0.1% of users choose GPT‑4o daily and most people use GPT‑5.2, citing recent upgrades to personality, customization, and creative ideation. GPT‑4.1, GPT‑4.1 mini, and o4-mini will also exit ChatGPT. The API stays unchanged.