Midjourney's AI video model is here

👋 Good morning/evening (wherever you are). It’s Wednesday.

Midjourney’s video model (image-to-video) is finally here.

Specs:

• 480p output (use an upscaler like Topaz for better quality)

• 5-second videos, extendable 4 times

• Two animation modes: Auto or Manual prompts

• Motion levels:

  • Low motion = subtle movement

  • High motion = dynamic scenes (more error-prone)

• Pricing: ~8x the cost of an image job, but you get 4 videos (roughly 1 image per second of output)

Additional thoughts:

• Best suited for anime and abstract visuals (expect more crossover fan clips)

• Fashion, gaming, and ecomm brands can now test hundreds of visual ads per month with ease

• Combine with Suno or Udio for instant music videos (someone’s going to make a Gorillaz-style AI band soon)

• Spider-Man: Beyond the Spider-Verse drops June 4, 2027…until then, expect plenty of AI versions.

Switching gears:

Hailuo’s recent AI circus video should be getting way more attention.

Creatives are raving about its ability to handle action scenes, acrobatics, and wild physics better than anything we’ve seen yet.

I believe the rise of synthetic doubles is just the start.

Watch the video. Posted on LinkedIn.

Also, MiniMax M1 just dethroned DeepSeek as the top open-source model right now.

OK let’s keep going ↓

Here’s what you should know:

• Apple eyes using AI to design its chips

  Apple is exploring generative AI to accelerate the design of its custom chips, according to hardware chief Johny Srouji in a recent speech at Belgian R&D center Imec.

• OpenAI launches a podcast

  OpenAI launched its new podcast with Sam Altman discussing GPT-5, AGI, Project Stargate, and the future of AI in its first episode.

• Oracle Cloud Infrastructure becomes top choice for leading AI startups

  Oracle Cloud Infrastructure is now powering leading AI companies like Fireworks AI, Hedra, Numenta, and Soniox with high-performance, cost-efficient infrastructure for training and deploying next-gen AI models at scale.

• Nvidia-backed AI startup SandboxAQ creates new data to speed up drug discovery

  SandboxAQ released a massive AI-generated dataset to help scientists predict how drugs bind to proteins, aiming to accelerate drug discovery without relying on traditional lab experiments.

The numbers:

• Tennr raised $101M to automate patient referrals using AI-powered document processing.

• Animaj raised $85M to build AI-generated kids' entertainment IP from the ground up.

• Wix acquired Base44 for $80M to let anyone build apps and websites using natural language instead of code.

• Traversal raised $48M to reinvent site reliability with an AI-native observability platform.

• HOPPR raised $31.5M to transform medical imaging with foundation models trained on radiology data.

• Multiplier, founded by an ex-Stripe exec, raised $27.5M to roll up accounting firms and scale them using AI-powered automation.

• Mercanis raised €17.3M (~$18.6M) to automate procurement with agentic AI assistants.

• Grifin raised $11M for an investing app that buys stock in companies you shop at using AI.

• Typedef raised $5.5M to simplify building AI-optimized data pipelines for enterprise workflows.

Thought starters:

• OpenAI finds misaligned personas in fine-tuned models

• Stephen Wolfram & Joscha Bach: What if we had bigger brains

  ⏱️ 3 hrs (YouTube)

Meme of the day:

Thanks for reading,

Eddie

Are we friends on LinkedIn and X? Let’s connect — DM me or hit the follow button.

P.S. If this was valuable, forward it to a friend. If you’re that smart friend, subscribe here.

P.P.S. Interested in reaching other ambitious readers like you? To become a sponsor, reply to this email.