What Ten Years of Agtech Podcasts Reveal About Food and Agriculture

The shape of the conversation

Seventeen themes emerged from open coding with no predefined list. Hover a bar for each theme's scope and subthemes.

Figure 1. Share of the 689 episodes discussing each theme. Solid = the theme is a central (primary) focus of the episode; pale = discussed at any level.

Takeaway. The two largest themes are not production technologies — they are the money (57%) and the grower relationship (51%). When this industry talks to itself, the conversation is commercial and human first, technical second.

The seventeen themes at a glance

Each theme carries a stable identifier (T01-T17) and a color used consistently in every chart that follows. Click a card to open the theme's full evidence base: every verbatim passage coded to it across the corpus.

The seventeen themes, grouped by trajectory

The families below are a descriptive, post-hoc grouping computed from each theme's prevalence trajectory over 2016-2026; they played no role in the coding itself. Evaluative frames such as challenges versus opportunities are deliberately not applied: those categories are interpretive rather than observed, and the study's inductive design derives structure from the corpus alone.

A decade of conversation, as a river

Each band is a theme; its width is that theme's share of everything discussed that year (stacked to 100%). Hover to read values; click legend entries to hide themes, or use the buttons to reset.

✓ select all top 6 only

Figure 2. Themes stacked to 100% of each year's theme mentions. 2016 (n=7) and 2026 (n=30) are partial years; early widths are noisy.

Takeaway. The river's banks are stable — finance and commercialization hold a third of the conversation in every year — while the middle churns: climate & carbon swells through 2021-22 then narrows, and data & AI widens steadily toward the present.

Rising, cresting, fading

Six selected themes as lines. Click legend entries to isolate any subset; hover for the crosshair readout.

Figure 3. Share of each year's episodes discussing the theme.

Takeaway. Three different fates: finance climbs all decade and peaks late (the funding correction became its own story); climate & carbon crests in 2021-22 and gives back a third of its peak; data & AI accelerates after 2023 and leads 2026.

The codebook as a wheel

The full emergent structure at a glance: hover any slice for its name and scope. Full theme profiles, including the complete evidence base, are available in the theme gallery above.

Figure 4. The emergent codebook: 17 themes (inner ring, sized by episode prevalence) and 53 subthemes (outer ring).

Takeaway. None of this structure was specified in advance — it is what 689 conversations produce when grouped by what speakers actually said. The wheel is the study's first finding.

What travels together

Each link joins two themes discussed in the same episodes; thicker means more shared episodes. Hover a theme to isolate its companions.

Figure 5. Links shown for theme pairs sharing at least 60 episodes.

Takeaway. Finance is the corpus's connective tissue: it forms the single most common pair with commercialization (202 episodes) and strong links to nearly every other theme — whatever the topic, the money question is usually in the room. The other standout pair is soil & regenerative with climate & carbon (129 episodes): the soil-carbon nexus is where stewardship and climate talk meet.

Every theme, every year

The full picture: 17 themes by 11 years. Hover any cell for exact values and episode counts.

Figure 6. Share of each year's episodes discussing each theme. Blank cells are true zeros (no episode that year discussed the theme); values below 10% are shown in muted gray. 2016 (n=7) and 2026 (n=30) are partial years.

Takeaway. The business rows stay dark across the entire decade, while the sharpest single-row swings belong to climate & carbon (16% → 52% → 23%) and to consumer trust, which opened the decade as a headline concern (71% of a seven-episode 2016) before settling into a steady background role.

Hype and fade, in the corpus's own vocabulary

Raw word counts, independent of the coding. Dark chips are combined families — an episode counts if it mentions any term in the family ("AI (all terms)" = AI, artificial intelligence, machine learning, computer vision, large language, ChatGPT). Toggle chips to compare.

Figure 7. Share of each year's episodes mentioning the term (or any term in the family) at least once.

Takeaway. Even counting every AI-adjacent term, the AI family only overtakes the climate family in 2025 — and blockchain remains the cleanest boom-bust on record here: 25% of 2018 episodes, effectively zero by 2023.

Does the structure hold up?

Figure 8. Saturation: cumulative distinct new-code concepts as coding progressed. The curve flattens by roughly 30% of the corpus — the basis for confidence that 17 themes span the conversation.

Three independent checks

Triangulation. An unsupervised topic model over the raw transcripts, blind to the coding, independently recovered the themes — including both late additions (insurance; food waste).

Reliability. An independent second coding of 40 random episodes agreed substantially: 87.5% cell agreement, Cohen's κ = 0.72, and 97% of episodes share a primary theme.

Verification. A script confirms all 3,335 quote-bearing tags are exact substrings of their cited transcripts. It passes.

Limitations. Self-selected speakers close to the agtech/venture world; uncorrected ASR; AI-assisted coding with subjective salience calls; a secondary analysis — we did not conduct the interviews.

Nine things the corpus reveals

It's about commerce and adoption, not the technology itself

Finance (57%) and commercialization & grower fit (51%) top the themes — and pair together in 202 episodes, the corpus's most common conversation.

"Talk to the grower" is the decade's throughline

Trust is the scarce resource; adoption is measured in seasons; service beats software alone. The lesson barely changes across ten years.

AI is the defining recent inflection

"AI" grows from 14% of 2016 episodes to 63% of 2026 episodes; LLM vocabulary appears only from 2023. The framing matures from novelty to infrastructure.

The sustainability wave crested and re-framed

Climate & carbon peaks in 2021-22 then softens, shifting from measurement toward business case — and, from 2023, toward adaptation and resilience.

Fads rise and fall; fundamentals persist

Blockchain: 25% of 2018 episodes, gone by 2023. Vertical farming's hype outran its prevalence. Soil, data ownership, and grower economics endure.

Bottom-up coding surfaces the human themes

Labor & community, consumer trust, markets & policy are mainstream — and the escape hatch surfaced energy, food waste, insurance, and farmer mental health.

The two shows divide the labor

AgTech So What? leans commercial (commercialization 63% vs 45%; climate 47% vs 31%); Future of Agriculture leans toward data & AI, labor & community, and value-added. One sits near the boardroom, the other near the farm gate.

Data, not models, is named as AI's real bottleneck

Across the AI surge, guests repeatedly point at standards, interoperability, and the willingness to share data — the plumbing, not the intelligence — as what holds agriculture back.

The evidence structure holds

Saturation by ~30% of the corpus, independent statistical recovery of the themes, and substantial double-coding agreement (κ 0.72) — the 17-theme structure is a property of the corpus, not of one reading.