❌ OpenAI shuts down Sora — Hype is not a product

Two years of hype. Spectacular demos. Media covers. And in the end, zero viable product. OpenAI shut down Sora on April 26, its AI video generator, because costs were impossible to scale and they had no clear use case. It's not a defeat — it's a lesson in focus that hurts but is necessary.

❌ OpenAI shuts down Sora — Hype isn't always product

On April 26, OpenAI shut down Sora, its AI video generator. Two years of hype, spectacular demos, media covers — and in the end, zero viable product.

Impossible-to-scale costs, lack of a clear use case, and competitors that had already eaten their lunch. OpenAI chose to kill it rather than divert resources from what really matters: GPT.

But Sora isn't the only story. While OpenAI killed its video generator, the rest of the sector kept moving in opposite directions.

CoreWeave and the money that doesn't stop

CoreWeave, the GPU-focused cloud provider, went public on April 22. It's the first major AI IPO of 2026, and it marks the starting gun for what will be a record year in tech exits. The company, which started as a cryptocurrency miner, has become one of the most important infrastructure providers for AI startups.

And between shutdowns and IPOs, the common thread is maturity. OpenAI kills what doesn't work. CoreWeave proves AI infrastructure has a public market. Investors are beginning to distinguish between hype and real business.

Killing a product that doesn't work while watching another go public in the same week is the perfect summary of where AI is in 2026: less magic, more business. And that, though it sounds contradictory, is the best thing that can happen to those of us who actually want to build with this.

— Max