By
Vlad Shvets
The AI Answers About EVs Run on Automaker Sites, Not Reddit
We split EV questions into what a spec sheet can answer and what only owners can. Community sources never appeared in either. The automakers' own sites carry the spec answers, the generalist press still tops the specialists, and the plan changes accordingly.
We split EV questions into what a spec sheet can answer and what only owners can. Community sources never appeared in either. The automakers' own sites carry the spec answers, the generalist press still tops the specialists, and the plan changes accordingly.
We split EV questions into what a spec sheet can answer and what only owners can. Community sources never appeared in either. The automakers' own sites carry the spec answers, the generalist press still tops the specialists, and the plan changes accordingly.
The standing advice for EV marketers has one villain and one hero: the spec sheet is dead weight, and Reddit is where the range anxiety, the charging honesty, and the year-eight battery verdicts live. Budget follows the forums.
So we split the EV questions a buyer asks into two kinds, the ones a spec sheet can settle and the ones only owners can, ran both on ChatGPT and Google AI Mode, and read what the answers were built from.
The short version: the community layer never showed up, in either kind of question. The automakers' own sites carry the spec answers, the generalist car press still outranks every EV specialist, and the window sticker is having a better AI-search year than the forums.
One caveat up front: these answers move monthly; what follows is a compass, not a survey.
Reddit Never Showed Up
Across a targeted set of EV buying and ownership questions we ran on ChatGPT and Google AI Mode in August 2026 (US-weighted, with a UK, Canadian, and Australian mix), community sources appeared in 0.00% of both question types' answers: no Reddit, no owner forums, no Quora, on the spec questions and on the exact experience questions the forums were supposed to own.
Read that carefully rather than triumphantly: it is one month's answers, and in the broader automotive picture we published from June data, Reddit led the cited sources. That was a different month and a different question set; we are not pooling the two.
What this August read says is narrower and sharper: if your EV visibility plan is a Reddit plan, these answers are not currently listening there.
The Window Sticker Won
What replaced the forums is the least romantic source on the internet: the manufacturer. An automaker's own site was in 40.00% of the spec-question answers and 16.80% of the experience-question ones (share of each question type's answers).
Whole probe of these questions, the leaders were hyundaiusa.com at 19.20%, kia.com at 12.40%, and tesla.com at 5.60% of all answers. The engines treat a complete, readable spec and range page as documentation, and the brands with the most machine-readable estates are getting cited for the category's factual layer.
Nobody's forum post beats the automaker's own range page when the question has a number for an answer.
The government spec layer everyone assumes anchors these answers barely exists in them: the official fuel-economy and safety sites reached 2.00% of all answers. If the EPA figure shows up, it shows up quoted on someone else's page.
Car and Driver Still Outranks Every EV Specialist
The second surprise is who carries the judgment questions, and it is the generalist car press rather than the EV-native one. caranddriver.com was in 18.80% of all answers, above every EV-specialist site (insideevs.com 8.80%, evchargesavings.com 4.40%) and every community domain (none appeared). As layers: the generalist press was in 26.00% of all answers, the EV specialists in 17.20%.
The rest of the surface is a working middle: autoblog.com at 10.00%, recurrentauto.com, the used-EV battery-data site, at 7.20%, evcompared.co.uk at 5.60%, iseecars.com at 5.20%. Useful names for an EV brand's mention plan, none of them the community layer the playbook promised.
What an EV Brand Does With This
Our judgment given the pattern, not a proven sequence:
Treat your spec pages as the citation asset they already are. Range, charging curve, battery warranty, in prose the engines can lift. The 40.00% layer is the one you fully control.
Court the generalist press first, the EV press second. The review budget goes where the answers look: Car and Driver's EV verdicts are carrying more of the answer surface than the EV-native sites.
Get into the data layer nobody markets to. Battery-health and price-analysis sites are already in the answers; a data partnership beats a press release there.
Keep the Reddit work sized to the evidence. Community can matter for humans and for other months; in these answers it carried nothing. Measure before you budget.
Watch Which Layer Answers Your EV Questions in Qvery
This split will move; trigger behavior and source mixes have moved before. The way to bet on it is to watch it.
In Qvery, track the spec and ownership questions buyers ask about your models and Qvery runs them daily on ChatGPT and Google AI Mode, capturing every citation tied to the query and engine that produced it. Open Citations and read whether your market's answers lean on automaker estates, the generalist press, or the specialist and data layer, and whether your own pages are in the pile.
(One reading note for the citation lists: a share of Google AI Mode's citations resolve to Google's own domain rather than an external source, so external-source reads lean on the ChatGPT side of the record.)
Your model lineup's questions fit inside a free 7-day trial: sign up for Qvery and see whether the window sticker is winning your category too.
One move this week: read your own range and charging pages the way an engine does, next to Hyundai's. That comparison, not your subreddit strategy, is currently the fight.
The standing advice for EV marketers has one villain and one hero: the spec sheet is dead weight, and Reddit is where the range anxiety, the charging honesty, and the year-eight battery verdicts live. Budget follows the forums.
So we split the EV questions a buyer asks into two kinds, the ones a spec sheet can settle and the ones only owners can, ran both on ChatGPT and Google AI Mode, and read what the answers were built from.
The short version: the community layer never showed up, in either kind of question. The automakers' own sites carry the spec answers, the generalist car press still outranks every EV specialist, and the window sticker is having a better AI-search year than the forums.
One caveat up front: these answers move monthly; what follows is a compass, not a survey.
Reddit Never Showed Up
Across a targeted set of EV buying and ownership questions we ran on ChatGPT and Google AI Mode in August 2026 (US-weighted, with a UK, Canadian, and Australian mix), community sources appeared in 0.00% of both question types' answers: no Reddit, no owner forums, no Quora, on the spec questions and on the exact experience questions the forums were supposed to own.
Read that carefully rather than triumphantly: it is one month's answers, and in the broader automotive picture we published from June data, Reddit led the cited sources. That was a different month and a different question set; we are not pooling the two.
What this August read says is narrower and sharper: if your EV visibility plan is a Reddit plan, these answers are not currently listening there.
The Window Sticker Won
What replaced the forums is the least romantic source on the internet: the manufacturer. An automaker's own site was in 40.00% of the spec-question answers and 16.80% of the experience-question ones (share of each question type's answers).
Whole probe of these questions, the leaders were hyundaiusa.com at 19.20%, kia.com at 12.40%, and tesla.com at 5.60% of all answers. The engines treat a complete, readable spec and range page as documentation, and the brands with the most machine-readable estates are getting cited for the category's factual layer.
Nobody's forum post beats the automaker's own range page when the question has a number for an answer.
The government spec layer everyone assumes anchors these answers barely exists in them: the official fuel-economy and safety sites reached 2.00% of all answers. If the EPA figure shows up, it shows up quoted on someone else's page.
Car and Driver Still Outranks Every EV Specialist
The second surprise is who carries the judgment questions, and it is the generalist car press rather than the EV-native one. caranddriver.com was in 18.80% of all answers, above every EV-specialist site (insideevs.com 8.80%, evchargesavings.com 4.40%) and every community domain (none appeared). As layers: the generalist press was in 26.00% of all answers, the EV specialists in 17.20%.
The rest of the surface is a working middle: autoblog.com at 10.00%, recurrentauto.com, the used-EV battery-data site, at 7.20%, evcompared.co.uk at 5.60%, iseecars.com at 5.20%. Useful names for an EV brand's mention plan, none of them the community layer the playbook promised.
What an EV Brand Does With This
Our judgment given the pattern, not a proven sequence:
Treat your spec pages as the citation asset they already are. Range, charging curve, battery warranty, in prose the engines can lift. The 40.00% layer is the one you fully control.
Court the generalist press first, the EV press second. The review budget goes where the answers look: Car and Driver's EV verdicts are carrying more of the answer surface than the EV-native sites.
Get into the data layer nobody markets to. Battery-health and price-analysis sites are already in the answers; a data partnership beats a press release there.
Keep the Reddit work sized to the evidence. Community can matter for humans and for other months; in these answers it carried nothing. Measure before you budget.
Watch Which Layer Answers Your EV Questions in Qvery
This split will move; trigger behavior and source mixes have moved before. The way to bet on it is to watch it.
In Qvery, track the spec and ownership questions buyers ask about your models and Qvery runs them daily on ChatGPT and Google AI Mode, capturing every citation tied to the query and engine that produced it. Open Citations and read whether your market's answers lean on automaker estates, the generalist press, or the specialist and data layer, and whether your own pages are in the pile.
(One reading note for the citation lists: a share of Google AI Mode's citations resolve to Google's own domain rather than an external source, so external-source reads lean on the ChatGPT side of the record.)
Your model lineup's questions fit inside a free 7-day trial: sign up for Qvery and see whether the window sticker is winning your category too.
One move this week: read your own range and charging pages the way an engine does, next to Hyundai's. That comparison, not your subreddit strategy, is currently the fight.
The standing advice for EV marketers has one villain and one hero: the spec sheet is dead weight, and Reddit is where the range anxiety, the charging honesty, and the year-eight battery verdicts live. Budget follows the forums.
So we split the EV questions a buyer asks into two kinds, the ones a spec sheet can settle and the ones only owners can, ran both on ChatGPT and Google AI Mode, and read what the answers were built from.
The short version: the community layer never showed up, in either kind of question. The automakers' own sites carry the spec answers, the generalist car press still outranks every EV specialist, and the window sticker is having a better AI-search year than the forums.
One caveat up front: these answers move monthly; what follows is a compass, not a survey.
Reddit Never Showed Up
Across a targeted set of EV buying and ownership questions we ran on ChatGPT and Google AI Mode in August 2026 (US-weighted, with a UK, Canadian, and Australian mix), community sources appeared in 0.00% of both question types' answers: no Reddit, no owner forums, no Quora, on the spec questions and on the exact experience questions the forums were supposed to own.
Read that carefully rather than triumphantly: it is one month's answers, and in the broader automotive picture we published from June data, Reddit led the cited sources. That was a different month and a different question set; we are not pooling the two.
What this August read says is narrower and sharper: if your EV visibility plan is a Reddit plan, these answers are not currently listening there.
The Window Sticker Won
What replaced the forums is the least romantic source on the internet: the manufacturer. An automaker's own site was in 40.00% of the spec-question answers and 16.80% of the experience-question ones (share of each question type's answers).
Whole probe of these questions, the leaders were hyundaiusa.com at 19.20%, kia.com at 12.40%, and tesla.com at 5.60% of all answers. The engines treat a complete, readable spec and range page as documentation, and the brands with the most machine-readable estates are getting cited for the category's factual layer.
Nobody's forum post beats the automaker's own range page when the question has a number for an answer.
The government spec layer everyone assumes anchors these answers barely exists in them: the official fuel-economy and safety sites reached 2.00% of all answers. If the EPA figure shows up, it shows up quoted on someone else's page.
Car and Driver Still Outranks Every EV Specialist
The second surprise is who carries the judgment questions, and it is the generalist car press rather than the EV-native one. caranddriver.com was in 18.80% of all answers, above every EV-specialist site (insideevs.com 8.80%, evchargesavings.com 4.40%) and every community domain (none appeared). As layers: the generalist press was in 26.00% of all answers, the EV specialists in 17.20%.
The rest of the surface is a working middle: autoblog.com at 10.00%, recurrentauto.com, the used-EV battery-data site, at 7.20%, evcompared.co.uk at 5.60%, iseecars.com at 5.20%. Useful names for an EV brand's mention plan, none of them the community layer the playbook promised.
What an EV Brand Does With This
Our judgment given the pattern, not a proven sequence:
Treat your spec pages as the citation asset they already are. Range, charging curve, battery warranty, in prose the engines can lift. The 40.00% layer is the one you fully control.
Court the generalist press first, the EV press second. The review budget goes where the answers look: Car and Driver's EV verdicts are carrying more of the answer surface than the EV-native sites.
Get into the data layer nobody markets to. Battery-health and price-analysis sites are already in the answers; a data partnership beats a press release there.
Keep the Reddit work sized to the evidence. Community can matter for humans and for other months; in these answers it carried nothing. Measure before you budget.
Watch Which Layer Answers Your EV Questions in Qvery
This split will move; trigger behavior and source mixes have moved before. The way to bet on it is to watch it.
In Qvery, track the spec and ownership questions buyers ask about your models and Qvery runs them daily on ChatGPT and Google AI Mode, capturing every citation tied to the query and engine that produced it. Open Citations and read whether your market's answers lean on automaker estates, the generalist press, or the specialist and data layer, and whether your own pages are in the pile.
(One reading note for the citation lists: a share of Google AI Mode's citations resolve to Google's own domain rather than an external source, so external-source reads lean on the ChatGPT side of the record.)
Your model lineup's questions fit inside a free 7-day trial: sign up for Qvery and see whether the window sticker is winning your category too.
One move this week: read your own range and charging pages the way an engine does, next to Hyundai's. That comparison, not your subreddit strategy, is currently the fight.
The standing advice for EV marketers has one villain and one hero: the spec sheet is dead weight, and Reddit is where the range anxiety, the charging honesty, and the year-eight battery verdicts live. Budget follows the forums.
So we split the EV questions a buyer asks into two kinds, the ones a spec sheet can settle and the ones only owners can, ran both on ChatGPT and Google AI Mode, and read what the answers were built from.
The short version: the community layer never showed up, in either kind of question. The automakers' own sites carry the spec answers, the generalist car press still outranks every EV specialist, and the window sticker is having a better AI-search year than the forums.
One caveat up front: these answers move monthly; what follows is a compass, not a survey.
Reddit Never Showed Up
Across a targeted set of EV buying and ownership questions we ran on ChatGPT and Google AI Mode in August 2026 (US-weighted, with a UK, Canadian, and Australian mix), community sources appeared in 0.00% of both question types' answers: no Reddit, no owner forums, no Quora, on the spec questions and on the exact experience questions the forums were supposed to own.
Read that carefully rather than triumphantly: it is one month's answers, and in the broader automotive picture we published from June data, Reddit led the cited sources. That was a different month and a different question set; we are not pooling the two.
What this August read says is narrower and sharper: if your EV visibility plan is a Reddit plan, these answers are not currently listening there.
The Window Sticker Won
What replaced the forums is the least romantic source on the internet: the manufacturer. An automaker's own site was in 40.00% of the spec-question answers and 16.80% of the experience-question ones (share of each question type's answers).
Whole probe of these questions, the leaders were hyundaiusa.com at 19.20%, kia.com at 12.40%, and tesla.com at 5.60% of all answers. The engines treat a complete, readable spec and range page as documentation, and the brands with the most machine-readable estates are getting cited for the category's factual layer.
Nobody's forum post beats the automaker's own range page when the question has a number for an answer.
The government spec layer everyone assumes anchors these answers barely exists in them: the official fuel-economy and safety sites reached 2.00% of all answers. If the EPA figure shows up, it shows up quoted on someone else's page.
Car and Driver Still Outranks Every EV Specialist
The second surprise is who carries the judgment questions, and it is the generalist car press rather than the EV-native one. caranddriver.com was in 18.80% of all answers, above every EV-specialist site (insideevs.com 8.80%, evchargesavings.com 4.40%) and every community domain (none appeared). As layers: the generalist press was in 26.00% of all answers, the EV specialists in 17.20%.
The rest of the surface is a working middle: autoblog.com at 10.00%, recurrentauto.com, the used-EV battery-data site, at 7.20%, evcompared.co.uk at 5.60%, iseecars.com at 5.20%. Useful names for an EV brand's mention plan, none of them the community layer the playbook promised.
What an EV Brand Does With This
Our judgment given the pattern, not a proven sequence:
Treat your spec pages as the citation asset they already are. Range, charging curve, battery warranty, in prose the engines can lift. The 40.00% layer is the one you fully control.
Court the generalist press first, the EV press second. The review budget goes where the answers look: Car and Driver's EV verdicts are carrying more of the answer surface than the EV-native sites.
Get into the data layer nobody markets to. Battery-health and price-analysis sites are already in the answers; a data partnership beats a press release there.
Keep the Reddit work sized to the evidence. Community can matter for humans and for other months; in these answers it carried nothing. Measure before you budget.
Watch Which Layer Answers Your EV Questions in Qvery
This split will move; trigger behavior and source mixes have moved before. The way to bet on it is to watch it.
In Qvery, track the spec and ownership questions buyers ask about your models and Qvery runs them daily on ChatGPT and Google AI Mode, capturing every citation tied to the query and engine that produced it. Open Citations and read whether your market's answers lean on automaker estates, the generalist press, or the specialist and data layer, and whether your own pages are in the pile.
(One reading note for the citation lists: a share of Google AI Mode's citations resolve to Google's own domain rather than an external source, so external-source reads lean on the ChatGPT side of the record.)
Your model lineup's questions fit inside a free 7-day trial: sign up for Qvery and see whether the window sticker is winning your category too.
One move this week: read your own range and charging pages the way an engine does, next to Hyundai's. That comparison, not your subreddit strategy, is currently the fight.
© 2026 Qvery AI OÜ