Waveform And Vapi Raise Capital For AI Voices, TikTok Goes The Clock, Samsung’s New XR Headset

Open AI Went Down Yesterday. It was highly inconvenient. ChatGPT has snuggled right into my workflow and now I’m dependent on the little bugger. The moment reminded me of AOL’s outage in 1996. The service was down for 24 hours and sent shockwaves throughout the world, demonstrating the Internet was a utility, like electricity and water. This sent AOL’s valuation on a rocket ride. The stock split over and over again for the next three years.

The clock ticks for Tik Tok. ByteDance and its subsidiary TikTok have filed an emergency motion with the U.S. Court of Appeals for the District of Columbia, seeking to temporarily halt a law mandating ByteDance to divest TikTok by January 19, 2025, or face a U.S. ban. The companies argue that without this injunction, TikTok’s 170 million American users will lose access to the platform. They also note that President-elect Donald Trump has expressed opposition to the ban, suggesting a delay would allow the incoming administration to reassess the situation. The Justice Department opposes the request, citing national security concerns over Chinese control of the app.

Waveform Raises $40M for voices that have emotional intelligence. Founded by former OpenAI researcher Alexis Conneau, WaveForms AI is developing AI voice models with enhanced emotional intelligence, enabling more empathetic and realistic interactions. Conneau previously contributed to OpenAI’s GPT-4o voice mode, noted for its real-time responsiveness and ability to handle interruptions. Funding came from mega VC Andreessen Horowitz.

Vapi Raises $20M to Bring Voice AI Agents to Enterprise. Voice AI platform Vapi has raised $20 million in a Series A funding round to expand its engineering team, scale real-time infrastructure, and support enterprise customers. Since launching in 2023, Vapi has scaled to millions in revenue, providing customizable AI voice agents for industries like healthcare, finance, and customer service. Bessemer Venture Partners led the round, with participation from Y Combinator, Abstract Ventures, and others.

OpenAI Releases Sora, Outages Ensue Amid Surging Demand. OpenAI has officially launched Sora, its long-awaited text-to-video AI model, as part of its “12 Days of OpenAI” event. Billed as a breakthrough in AI-generated media, Sora allows users to create high-quality video content from simple text prompts. However, the release was met with overwhelming demand, causing widespread outages and access issues. Users reported slow response times and platform crashes, underscoring the intense interest in Sora’s capabilities. Despite the hiccups, OpenAI’s strategic rollout aims to solidify its lead in generative AI. As access stabilizes, Sora is expected to compete with models from Google and Runway. Personally, Sora still won’t let me in.

Road to VR’s Ben Lang Got Hands On With Samsung’s Project Moohan XR Headset. Samsung has revealed Project Moohan, its new mixed-reality (XR) headset powered by Android XR, blending design elements from Meta’s Quest and Apple’s Vision Pro. Slated for a 2025 release, the device features a Snapdragon XR2+ Gen 2 chip and supports hand, eye, and controller-based input.

Google’s New Gemini 2.0 Can Open the Browser to Check Information Independently. Unveiled this week, Gemini 2.0 introduces autonomous AI capabilities that set it apart from its predecessors. The model’s most notable advancement is its ability to independently open a browser to verify information in real-time — a defining feature of Google’s so-called “Agentic Era.” The multimodal AI supports text, images, video, audio, and code, with applications spanning Google Search. Projects like Astra (AI assistant), Mariner (web task automation), and Jules (developer agent) demonstrate the model’s practical potential. Google emphasized its focus on safety and continues to maintain the strictest guardrails in the industry.

X Makes AI Chatbot Grok Free for All Users, Shifts to Freemium Model X (formerly Twitter) has made its AI chatbot Grok available to all users, eliminating the need for a Premium subscription. Users now receive 10 free prompts and 10 image generations every two hours, though image analysis is limited to three daily before a subscription is required. This move aligns X’s AI strategy with competitors like ChatGPT and Claude. Unlike Google Gemini, X’s Grok is uncensored.

China Launches Bold BCI Trials to Rival Elon Musk’s Neuralink. China is advancing brain-computer interface technology with plans for large-scale clinical trials of its Neural Electronic Opportunity (NEO) device in 2025. Developed by Neuracle Technology and Tsinghua University, NEO employs a semi-invasive approach, placing electrodes outside the brain cortex to avoid direct contact with brain tissue. In a recent procedure, a 38-year-old spinal cord injury patient regained control of a prosthetic hand, performing tasks like unscrewing a bottle cap. The operation took just 1 hour and 40 minutes, aided by a real-time brain localization system.

“Dino Hab” Brings Prehistoric Adventure to MR and VR. Film director Doug Liman’s 30 Ninjas, Meta, and Dark Slope have teamed up to launch Dino Hab, a Mixed Reality experience that lets players raise dinosaurs and restore prehistoric habitats in immersive 3D worlds. The company is touting its AI companion, powered by Inworld AI, that guides players between MR and VR environments. This unique cross-platform approach adds depth to the experience, blending exploration, care, and adventure.

This column, formerly called “This Week in XR,” is also a podcast hosted by the author of this column, Charlie Fink, Ted Schilowitz, former studio executive and co-founder of Red Camera, and Rony Abovitz, founder of Magic Leap. This week our guest is. We can be found on Spotify, iTunes, and YouTube.

What We’re Reading

How Gaming Built The Metaverse While Big Tech Wasn’t Looking (Catherine Henry/)