Too long. Need 60-90. Let's shorten. "OpenAI launches Sora as Apple debuts cheap iPad

OpenAI’s Sora, the company’s generative‑video model that debuted in limited beta last year, received a sweeping upgrade on 9 February 2026 that turns it from a novelty clip‑maker into a true storytelling engine. According to Indonesian outlet Uzone.id, the new “Extensions” feature lets any user—whether on iOS or the web—ask Sora to continue an existing sequence, preserving lighting, composition and narrative flow. A creator could upload a drone shot of a forest canopy and command the AI to “take the drone into the cave,” and Sora will generate seamless footage that matches the original aesthetic. The move is being hailed as the most‑anticipated AI‑video development of the week, because it shifts Sora from producing disjointed GIF‑like snippets to enabling end‑to‑end visual narratives without a human editor.

Apple, meanwhile, unveiled a stripped‑down iPad priced under $200 that ships with a custom‑designed neural‑engine chip capable of running on‑device inference for OpenAI’s models. The device, announced at the company’s spring hardware event, is marketed as a “learning tablet” for students and hobbyists, promising smooth performance for AI‑powered apps without relying on cloud compute. Gizmodo notes that Apple’s strategy mirrors Sam Altman’s recent vision of AI as a utility “too cheap to meter,” suggesting that the hardware‑software pairing could make high‑quality generative tools affordable at scale. By embedding the inference engine locally, Apple sidesteps the bandwidth and privacy concerns that have plagued cloud‑only solutions, while also positioning the iPad as a gateway for developers to embed Sora’s extensions directly into native iOS experiences.

Altman’s utility metaphor was delivered at BlackRock’s U.S. Infrastructure Summit, where he told board member Adebayo Ogunlesi that “intelligence will be a utility, like electricity or water, and people will buy it on a meter.” He added that OpenAI’s “fundamental belief in abundance of intelligence” drives the company toward a “too cheap to meter” future, even as the industry grapples with the finite nature of compute and the rising cost of token‑based billing. The timing of Sora’s upgrade and Apple’s low‑cost iPad underscores that push: if the hardware can run the models locally, the “meter” could indeed disappear for end users, shifting revenue pressure back onto the platform providers.

The convergence of these moves has immediate implications for creators across the spectrum. Film studios, which have long feared AI‑generated footage cannibalizing traditional pipelines, now face a tool that can extend footage in minutes rather than weeks, potentially reshaping post‑production budgets. At the same time, classrooms could adopt the cheap iPad as a sandbox for students to experiment with AI‑augmented storytelling, democratizing access that was previously limited to high‑end workstations or cloud credits. TechCrunch’s coverage of OpenAI’s broader enterprise rollout—such as the new AI‑agent management suite and the Prism workspace for scientists—reinforces that the company is building an ecosystem where the same models powering Sora can be deployed in everything from corporate automation to academic research, all while leaning on hardware that makes on‑device inference viable.

Analysts see the pairing as a test of Altman’s utility thesis. If Apple’s iPad can deliver a smooth, low‑latency experience for Sora extensions without incurring per‑token charges, the “too cheap to meter” narrative gains credibility; if not, the model may revert to cloud billing, reinforcing the very metered usage Altman warned against. For now, the market is watching how quickly developers integrate Sora’s extensions into iPad apps and whether the device’s neural engine can sustain the compute demands of high‑resolution video generation. The outcome will likely shape the next wave of AI‑driven content creation, and could determine whether intelligence truly becomes a utility that’s free at the point of use.