In partnership with
Shoppers are adding to cart for the holidays
Over the next year, Roku predicts that 100% of the streaming audience will see ads. For growth marketers in 2026, CTV will remain an important “safe space” as AI creates widespread disruption in the search and social channels. Plus, easier access to self-serve CTV ad buying tools and targeting options will lead to a surge in locally-targeted streaming campaigns.
Read our guide to find out why growth marketers should make sure CTV is part of their 2026 media mix.
AI Just Got Ridiculously Good in 2025(And Your Job Might Not Be Boring Anymore)
Remember when ChatGPT was the cool new thing that could write essays for you? Yeah, that's basically the model year 2020 of AI now. In 2025, the AI world went absolutely bonkers.
Here's what happened: OpenAI dropped GPT-5 in August and said it was like having a PhD on speed dial. No hyperbole—they literally called it "the first time you can ask a PhD-level expert anything." The thing can write actual working software from a text description (they call it "vibe coding," which is the most Silicon Valley term ever invented). Imagine telling your computer "make me a calculator app" and it just... does it. Wild, right?
But OpenAI wasn't alone. Google showed up in November with Gemini 3 and basically said, "Hold my coffee." Google's model can see and understand videos, process a million tokens at once (that's like feeding it an entire book series in one go), and it's so good at reasoning that it beats GPT-5 at complex math problems. The catch? It's cheaper.
Here's where it gets spicy: open-source AI—the free stuff you can download—got ridiculously good. Meta released Llama, which performed just as well as models costing hundreds of dollars per use, but basically free if you run it yourself. The performance gap between fancy paid AI and free AI collapsed from 8% to just 1.7% in a single year. That's like discovering the expensive coffee and the cheap coffee taste almost identical.
So what does this mean for actual humans?
Companies are going all-in. 82% of enterprise employees now use AI weekly (up from 37% in 2023), and 46% use it daily. Finance teams are using it to analyze investments, hospitals are using it to diagnose diseases faster, and lawyers are using it to sift through millions of documents without losing their minds. People are actually measuring whether AI is making money for them—72% of organizations now track AI ROI.
The wild part? These models still hallucinate sometimes (they make stuff up), they struggle with certain types of logic, and nobody's really figured out the full safety implications yet. But organizations are deploying them anyway because the benefits are too big to ignore.
We're at this weird moment where AI went from "interesting experiment" to "how do we actually use this without breaking anything?" It's like everyone's still figuring out the rules while playing the game.
The game, by the way, is moving fast.