AI Video News This Week: Sora Shut Down, Grok Imagine Takes the Lead, and What It Means for DTC Brands

This was a genuinely seismic week in AI, especially for anyone running AI video ads for DTC brands. OpenAI officially shut down Sora on April 26, the AI video leaderboard had a new model take first place, Anthropic raised another $50 billion, and the enterprise AI venture wars heated up. Here is the AI video news this week that actually matters for DTC marketers, with a clear-eyed take on what it changes for your creative workflow.

OpenAI shut down Sora completely

The big one. OpenAI officially closed the Sora consumer app on April 26, 2026, and the Sora API will be discontinued on September 24, 2026. The Wall Street Journal reported Sora was costing OpenAI roughly $1 million per day to operate while active user growth had stalled. OpenAI cited a strategic pivot toward enterprise and productivity tools, plus ongoing concerns about deepfakes, copyright disputes, and the proliferation of "AI slop" content the app generated.

For DTC brands, this is more significant than it sounds. Sora 2 was the most-talked-about AI video tool in 2025. Marketing teams who built workflows around Sora are now scrambling to migrate. If you were testing Sora for your brand or still have credits sitting in the Sora API, you have until September 24 to migrate before the lights go out completely.

The good news: there are now multiple production-ready alternatives that arguably outperform Sora on every metric DTC brands care about. The Sora shutdown is not a creative crisis. It is a forced upgrade.

Grok Imagine takes the lead: the AI video news this week

This is the under-reported story of the week and the one that matters most for DTC brands focused on creative volume and cost.

The LLM Stats Video Arena leaderboard, which ranks AI video generators based on 2,204 blind human votes comparing real video outputs, now shows Grok Imagine leading with an arena score of 724. Veo 3.1 sits in second at 618, and WAN Video 2.6 third at 577.

Grok Imagine winning the blind-vote arena is significant because it is also one of the cheapest models in the field at around £0.07 per generation. For DTC brands running 30-50 ad variants per week, the cost-quality intersection is exactly where Grok Imagine sits in May 2026. You can run more variants at lower cost and the human-judgement quality holds up.

For brands using Tonic Studio, Grok Imagine has been part of our auto-routing model lineup since the recent refresh, alongside Veo 3.1 Standard, Veo 3.1 Fast, Sora 2 Pro (still available via API until September), Kling 3.0 Pro, Seedance 2.0, and Happy Horse. Our auto-router automatically picks Grok Imagine for fast iteration and creative testing where quality is "good enough" rather than cinematic, which lines up exactly with what the leaderboard data now confirms.

If you have been defaulting to Sora 2 Pro for everything, this week is your prompt to switch your testing volume to Grok Imagine. Same quality, fraction of the cost, and the user data is now strongly behind it.

Veo 3.1 still wins for cinematic and audio

While Grok Imagine takes the volume crown, Google's Veo 3.1 continues to dominate the high-end cinematic tier. The model is the only mainstream tool that generates synchronised audio (ambient sound, dialogue, sound effects) in a single pass, and outputs true 4K at 60fps.

For DTC brands, this matters most for hero campaigns and premium positioning content. Product launch videos, founder-led brand films, and aspirational lifestyle content where production polish drives perceived brand value are exactly where Veo 3.1 should be your default.

The pattern emerging in May 2026 is clear: use Grok Imagine for performance creative volume (£0.07 per video), use Veo 3.1 Standard for hero campaigns where quality matters more than cost (£0.40 per second of video). Most successful DTC brands are running both in a 70/30 split, with Grok handling testing volume and Veo handling polished pieces.

Anthropic raises $50B at $900B valuation

TechCrunch reported this week that Anthropic is in the final stages of raising $50 billion in new funding at a $900 billion valuation. For context, OpenAI raised $122 billion at the end of March against an $852 billion valuation, so the two leading AI labs are now within striking distance of each other on valuation while sitting at roughly $1.7 trillion combined.

What this means for DTC marketers: the AI infrastructure underneath your creative tools is being capitalised for at least another decade of aggressive development. If you were waiting for the "AI hype cycle to end" before adopting AI video for your brand, that wait is over. The capital deployed into these companies indicates they will keep pushing capability, lowering costs, and integrating deeper into marketing workflows for years.

The brands waiting to adopt are now competing against well-funded competitors who are. The cost of waiting is a creative volume gap that compounds quarterly.

Anthropic and OpenAI both launching enterprise joint ventures

Both Anthropic and OpenAI announced new enterprise services joint ventures this week. Anthropic's venture, valued at $1.5 billion, includes Blackstone, Hellman & Friedman, and Goldman Sachs as founding partners. OpenAI's parallel venture, called The Development Company, was reported by Bloomberg.

Both ventures are focused on deploying AI directly into enterprise workflows, with the Anthropic announcement specifically describing engagements where AI engineers sit alongside clinical and IT staff to build tools that integrate with existing workflows.

For DTC brands, this matters less directly than the leaderboard news, but it signals where the industry is heading. AI is moving from "tool you experiment with" to "infrastructure layer of your business". The brands treating AI video as a strategic capability (rather than a one-off creative experiment) will benefit from the enterprise-grade tooling, integrations, and reliability that this capital will fund.

The Anthropic Project Deal experiment revealed something important

This week's most quietly significant data point came from Anthropic's "Project Deal" experiment, where 69 employee AI agents navigated 500+ listings to close 186 transactions totalling $4,000 in trades across the company's San Francisco headquarters.

The headline was the experiment working at all. The buried lede: Opus 4.5 agents systematically out-negotiated Haiku 4.5 agents on price and selection, while owners of the weaker agents remained unaware of the disadvantage they were operating under.

For DTC brands using AI for marketing creative, the same dynamic applies. The capability gap between top-tier AI models (Veo 3.1 Standard, Sora 2 Pro for the months it remains available) and budget-tier models (Grok Imagine, Veo 3.1 Fast) is real and measurable. Brands using only the cheapest models for everything are losing performance against competitors using premium models for hero content. Brands using only premium models for everything are spending 5-10x more than they need to on testing volume.

The right strategy is the same in performance marketing as it apparently is in office negotiations: match the model to the task, and use premium capability where premium capability matters.

Publishers blocking AI crawlers see traffic drop

A separate but related data point this week: research found that news publishers who blocked AI crawlers from training models on their content experienced an average 7% decline in weekly traffic within weeks of implementation. The drop appeared in human browsing data, not just bot traffic, suggesting reduced visibility in AI-driven discovery channels.

For DTC brands, this is a warning shot about the shift in how customers discover products. AI-mediated discovery (people asking ChatGPT, Claude, or Perplexity for product recommendations) is becoming a real traffic channel. Brands that block AI crawlers from their websites are removing themselves from these recommendation surfaces.

The bigger implication: DTC brands need to think about being discoverable by AI models, not just by Google. This means clear product information, well-structured content (we are working on this with our SEO push at Tonic Studio), and authentic brand voice that AI models can confidently surface in user queries.

What the AI video news this week means for your brand

If you take only one action from this week's AI video news, switch your testing volume from Sora 2 to Grok Imagine. Same quality, fraction of the cost, leaderboard now backs it, and Sora is going away anyway in September.

If you have time for two actions, also revisit your hero campaign tooling. If you have been using whatever your team set up six months ago, Veo 3.1 has likely surpassed it for cinematic quality. Worth running a side-by-side test on your next big campaign.

The broader pattern from this week is consistent with everything else in 2026: AI video is consolidating around a few clear winners (Grok Imagine for volume, Veo 3.1 for cinematic, Kling 3.0 for narrative chaining, Seedance 2.0 for reference-image fidelity), prices are continuing to drop, and the brands moving fastest are pulling away on creative volume.

If your DTC brand is still relying on traditional UGC for the majority of performance creative, this week is another data point that the cost-quality gap with AI is now too large to ignore.

How Tonic Studio fits this week's news

Quick context for the brands considering tools. Tonic Studio is the AI UGC platform built specifically for DTC brands in regulated categories: supplements, skincare, fitness, food and beverage, wellness.

Why this week's news strengthens the case for our platform:

We auto-route between all the top models including Grok Imagine, Veo 3.1 Standard, Veo 3.1 Fast, Kling 3.0 Pro, Seedance 2.0, and Happy Horse based on what your brief actually needs. As the leaderboards shift (like Grok Imagine taking the top spot this week), your output quality stays current automatically without you needing to migrate workflows. Same brand voice, same compliance guardrails, same library, just better models behind the scenes.

We are not affected by the Sora shutdown because we never ran on Sora's API as the primary engine. Brands using us did not need to migrate anything this week.

Our compliance guardrails for supplement, skincare, and food and beverage claim rules mean DTC brands in regulated categories can adopt AI video without the manual review burden that other tools force on you.

For the broader picture on how AI UGC fits into DTC marketing in 2026, see our complete guide to AI UGC for DTC brands.

If you are a DTC brand running performance marketing and the news this week has you reconsidering your AI video stack, try Tonic Studio. Welcome credits ready in 60 seconds.

Start creating →