Communications teams, strategists, and founders in consumer AI are being asked to pick a narrative to position against. That decision used to have an 18-month window. This data says it now has a 9-month window. Pick late and you are buying into a crowded story. Pick wrong and you are paying enterprise CPC for a consumer keyword.
The category itself is growing, not rotating.
Thirty-nine months of continuous media data, seven narratives, over 200,000 English-language news sources. Total volume in Q1 2026 reached its highest point on record. The standard assumption, that each wave of AI coverage cannibalizes the last one, is not what the data shows.
Consumer AI coverage hit a 359K-article quarter in Q1 2026, a category high.
Quarterly article volume · Jan 2023 – Mar 2026
Source: Perigon; Shadow analysis. n = 200,000+ sources; as of Mar 2026.
Every new narrative adds to the aggregate. Nothing dies cleanly. That matters because it changes the question from "which narrative do I ride this year" to "which narratives can I credibly hold at once."
So what: positioning is a portfolio problem, not a switching problem.
The successor arrived faster than the incumbent expected.
Agentic AI first appeared at meaningful volume in Q1 2024. It hit its inflection in Q4 2024, when quarterly growth reached 201.8%. By Q1 2026 it was at 180,822 articles per quarter, the single largest narrative volume in the dataset. March 2026 alone produced 98,484 articles.
Agentic AI overtook generative AI in Q1 2026, two years after first surfacing.
Quarterly article volume · Q1 2024 – Q1 2026
Source: Perigon; Shadow analysis. Crossover: Q1 2026.
The acceleration was anchored by specific corporate events: Salesforce Agentforce, Microsoft Copilot agents, and the broader enterprise AI agent ecosystem that consolidated in late 2024. But the mechanism under all of it is compression. Agentic AI covered the same ground generative AI covered, one quarter shorter per stage.
So what: the next cycle will arrive faster still.
Each narrative has its own shape, and the shape is the strategy.
Three narratives are surging: agentic, embodied, and personal AI. Three are cyclical: generative, safety, and companion. One is fading into a capability descriptor: multimodal. Laid side by side, the shapes tell you how to position.
Every narrative has its own shape. Three are surging, three are cyclical, one is fading.
Quarterly article volume · Jan 2023 – Mar 2026
No plateau signal yet. March 2026 alone produced 98,484 articles.
Resurgence paradox: Q1 2026 exceeded the 2023 peak by 11%.
Reactive pattern. Spikes after incidents, settles back to baseline.
Sawtooth, not a wave. Spikes on product news, retreats between.
Invisible until 2025. +187.7% QoQ in Q1 2026, fastest in the set.
Buyer intent forming ahead of media. +123% QoQ in Q1 2026.
Decelerating (7% QoQ). Being absorbed as a capability descriptor.
Source: Perigon; Shadow analysis. Dusk segment marks each narrative's largest QoQ jump.
A sawtooth shape (AI companion) says buyers are reacting to products, not narratives. A shallow wave with a reactive tail (AI safety) says coverage is event-driven, not campaignable. A steep, unbroken climb (embodied AI) says a corporate anchor has claimed the term and the cycle is live.
So what: pick the shape that matches the story you can actually tell.
Coverage is concentrating at the top.
Agentic AI and generative AI together accounted for 80% of all narrative-tagged coverage in Q1 2026. The long tail, safety, companion, embodied, personal, multimodal, split the remaining 20%.
Agentic AI captured more coverage than the next five narratives combined.
Q1 2026 article volume · Ranked
Source: Perigon; Shadow analysis. Articles per quarter, as of Mar 2026.
Concentration at the top creates two distinct plays. Fight for share of voice inside the crowded stories, where buyer intent is already priced in (agentic AI SDK keywords run $20 to $136 per click). Or build position in a thinner narrative before the crowd arrives.
So what: the tail is where differentiation is still cheap.
Embodied AI is the fastest mover of the quarter.
Ranked by quarter-over-quarter growth, embodied AI is the standout: 187.7% in Q1 2026. Personal AI is the next fastest at 123%. Agentic AI, despite its scale, grew at 85.5%. Growth rates invert the volume ranking.
Embodied AI grew faster than every other narrative, 88 points ahead of agentic.
QoQ percent change · Q4 2025 → Q1 2026
Source: Perigon; Shadow analysis. Percent change in quarterly article volume.
Embodied AI has the structural pattern of a dominant cycle in formation: multiple simultaneous corporate anchors (NVIDIA, the Chinese government, humanoid robotics companies), a logical semantic progression from generative to agentic to embodied, and active sub-narrative absorption. At the equivalent lifecycle stage, agentic AI was growing at roughly half this rate.
So what: the next 180K-article quarter is already visible in the data.
Generative AI did not die. It got repurposed.
Q1 2026 quarterly volume for generative AI reached 138,689. That exceeds the original 2023 peak of 124,440 by 11%. The standard narrative-lifecycle model does not predict this. The Gartner Hype Cycle does not predict this.
What happened is a term-level evolution. "Generative AI" in 2023 was a novelty label attached to ChatGPT and wonder. "Generative AI" in 2026 is an enterprise deployment category. Same words, different editorial weight. Coverage shifted from "what is this" to "how is it being implemented."
So what: no narrative needs to be abandoned, only reframed.
The forecast: embodied peaks in late 2026.
Applying the observed 33% compression pattern to the current inflection point, embodied AI reaches peak surge between Q4 2026 and Q1 2027. Agentic AI continues its climb through mid-2026 before decelerating. Generative AI holds as a resurgent baseline.
Embodied AI reaches peak surge in late 2026 if the compression pattern holds.
Projected quarterly volume · Q2 2026 – Q1 2027
Source: Shadow projection based on Perigon historical volume and observed 33% cycle compression. Dashed = projection.
Three things could break this forecast: a major robotics safety incident that kills embodied momentum, regulatory action on autonomous systems, or NVIDIA rebranding away from "physical AI." None look likely in the next two quarters. The positioning window for embodied AI is open now and closes in Q3 2026.
So what: if you are waiting for embodied AI to feel obvious, the window has already closed.