January 8, 2026
CITAlliance/Depositphotos.com
Has Retail Seen ‘Near-Universal’ Adoption of AI, as New Nvidia Data Suggests?
According to a new report issued by tech giant Nvidia titled “State of AI in Retail and CPG: 2026 Trends,” nearly all retail businesses are engaged with AI solutions as of this writing. The data provided indicates that a vast majority, 91% to be precise, of retailers are either actively using or assessing AI tools in some way, shape, or form.
“Across the business landscape, companies are transitioning from running numerous AI pilots to selecting and scaling solutions with proven business value. This trend is especially pronounced in retail and CPG, where active AI usage has significantly increased, and the share of organizations assessing AI has fallen 14 percentage points year over year,” a portion of the executive summary read.
Among the shifts noted by Nvidia: The maturation curve appears to be deepening. More than half (58%) of respondents indicated that their operations were actively rolling out AI tools, up 16 percentage points from the 42% figure registered in 2024. Conversely, about one-third (33%) suggested their businesses were engaged in assessing the viability of AI integrations, a statistic which fell by nearly the same measure (down by 14%, from 2024’s figure of 47%).
Other notable data points pulled from the report:
- Retailers say AI is slashing costs while boosting income: An overwhelming majority of those surveyed (95%) stated that AI deployment was decreasing annual operating costs, and only a slightly smaller cohort (89%) said that it was driving increases in yearly revenue.
- Retail enterprises remain optimistic on AI-enhancements in supply chain: When it comes to streamlining the supply chain, retailers who responded to the Nvidia poll were near unanimous in stating that AI solutions were helping curtail supply chain costs on an yearly basis (91%). A little over half (51%) said that they were currently leveraging AI to address “operational throughput and efficiency” in their supply chains.
- Objective versus outcomes: Retailers suggested three major objectives tied to AI usage — creating operational efficiencies (45%), improving customer experiences (38%), and improving employee productivity (29%). However, those same respondents indicated that they were not only meeting, but beating their initial goals. On the creation of operational efficiencies, 52% reported progress due to AI; on the improvement of customer experiences, 41% said as much; and concerning employee productivity based on AI involvement, 54% noted improvements.
- Agentic AI was deployed by 20% of retailers, with an even greater proportion (21%) planning deployment at some time in 2026: Over half of those planning to do so, or whom had already done so, cited increased process and efficiency outcomes as the primary motivator (57%). The next motivating factors were an enhanced customer experience and personalization measures (40%), improved decision-making involving real-time data (40%), increased ability to maneuver to address volatility or demand shifts (30%), and the reduction of operational costs (27%).
The Top Barrier to AI Adoption in Retail, Per Nvidia? A Lack of Talent
Nvidia was quick to highlight one massive problem tied to the breakneck pace of AI adoption in the retail and CPG spheres: a distinct lack of skilled manpower at hand.
“The AI talent shortage has intensified, rising from 31 percent last year to 46 percent this year, and is now the primary implementation barrier. This increase reflects both the surge in AI adoption and the rise of specialized technologies like agentic AI,” the study authors wrote.
“The shortage has real business implications. While 92 percent of executives plan to increase AI budgets, the lack of skilled talent can create a bottleneck where investment appetite exceeds execution capacity. Organizations that address this gap through hiring, strategic partnerships, or upskilling will gain a decisive advantage, translating AI ambitions into operational impact,” they added.
BrainTrust
Scott Benedict
Founder & CEO, Benedict Enterprises LLC
Gary Sankary
Retail Industry Strategy, Esri
Georganne Bender
Principal, KIZER & BENDER Speaking
Recent Discussions
Discussion Questions
Has retail actually embraced ‘near-universal’ adoption of AI, as Nvidia suggests? Why or why not, in your opinion?
Which data points are most intriguing or surprising to you, and which results are you most skeptical of?
Given central Nvidia’s position within the AI industry, do you believe this survey data may be subject to bias?
Poll
Loading …
I’ll bet you good money that there are LOADS of retail enterprises that haven’t done anything with AI yet. The leading retailers are all over it, as is the secondary tier, but to suggest that toy store chains with a few stores are racking up AI winnings just isn’t true.
“The data provided indicates that a vast majority, 91% to be precise, of retailers are either actively using or assessing AI tools in some way, shape, or form.” Using ChatGPT to write copy isn’t the same as actively using AI. Independent retailers, whom everyone seems to forget, are not adapting AI the same way the big guys are.
Has Retail Seen ‘Near-Universal’ Adoption of AI, as New Nvidia Data Suggests?
No, but it seems to be the Near Universal topic in the RW universe …today, anyway.
I believe that “near universal” adoption is about AI consideration/prioritization, but not implementation. We are only scratching the surface of AI adoption both from the brand and the consumer sides. While many AI features are being built into software and systems, most retailers still need to leverage the technology features to maximize efficiencies and profit gains. What remains most interesting to me is AI enabled commerce (AI agent conversion) which is now being tracked and measured and will explode within 5 years.
“Near-universal” adoption is a big stretch. Though the claim may hold up technically when you consider that AI is being integrated into everyday tools, like POS, accounting and marketing solutions. And any retailer who has ever written a prompt could be classified as “using AI.” While I accept and acknowledge NVIDIA’s perspective, I doubt that their survey is truly representative of the broad retail industry of which the vast majority are smaller independent/regional operations, and these firms will be well behind the adoption curve.
Reading this article, it’s clear that 90% of retailers are “assessing,” “planning,” or “piloting” AI. That’s very different from actually using it. We’ll get there, but for now, it’s more accurate to say that a conversation about AI is happening in 100% of retailers today.
Retailers are eager to deploy AI and they generally know where they want to apply it. I’m just not sure they’re as clear on which capabilities can truly be enhanced or how to operationalize the technology. My guess is that “mass adoption” will come from AI embedded in the solutions they already use, not from native implementations.
And for the many companies outside Tier 1 that don’t have development talent, that matters. Plenty are still running their BI out of Excel — expecting them to stand up custom AI models isn’t realistic.
I don’t know if “near-universal” adoption is accurate. To me, that means 80-90% (or more) of retailers are using AI. Regardless of whether that is accurate, the outcomes of proper AI use are accurate: operational efficiency, higher employee productivity, and better customer experiences. That’s what AI is supposed to do, and in many use cases, it’s working.
From a technical architecture perspective, “near-universal” AI adoption mostly reflects retailers consuming AI through SaaS platforms and embedded capabilities rather than building or operating AI as a first-class system. What matters less is whether AI exists in the stack and more whether it is integrated into core data pipelines, decision engines, and operational workflows. The most meaningful signal in the data is the move from pilots to scaled deployments, which aligns with rising pressure to prove measurable ROI in areas like inventory optimization, search relevance, and personalization.
I am skeptical of uniformly high claims around cost and revenue impact, because without clean data, strong governance, and tight system integration, AI outcomes tend to plateau quickly. Given Nvidia’s position in the ecosystem, the numbers may skew optimistic, but the underlying shift toward AI as an architectural necessity in retail is very real.
Another day, another lot of AI exaggeration! Again, the devil is in the details. If by adoption we mean anything and everything to do with AI – including things like basic use of ChatGPT and such – then yes, the numbers will be quite high. Even so, we know from our own discussions and surveys that it is not universal. But this kind of thing is akin to saying that most retailers use Excel. It is just not meaningful. The numbers deeply using AI and fully integrating it into operation is smaller and mostly confined to the large enterprises. This area is interesting, but we need to move beyond the headline to understand the impact.
Another day, another lot of AI exaggeration!
We can nevr be sure the sun will come up, but….
I think the claim of “near-universal” AI adoption in retail, as framed in the Nvidia narrative, is ambiguous at best and overstated at worst. Terminology like “near-universal” sounds definitive, but when you dig into the reality on the ground it’s clear that retailers are far from uniformly advanced in their AI investments and deployments. Yes, many of the largest, digitally savvy organizations have been strategic and aggressive about incorporating AI — whether that’s for demand forecasting, personalized recommendations, dynamic pricing, or computer vision-enabled loss prevention — but that doesn’t translate into consistent, optimized, enterprise-wide AI usage across the industry. For every category leader pushing ahead with agentic AI pilots and AI-augmented workflows, there are countless mid-tier and smaller chains still wrestling with basic data hygiene, siloed systems, and pilot projects that haven’t yet moved into scale or measurable ROI. In that sense, “near-universal” could mean that virtually every retailer talks about AI, but it doesn’t necessarily mean everyone has meaningfully adopted and integrated it into core operations.
The data points that feel most credible — and intriguing — are those that reflect intent and early adoption, such as the prevalence of pilot programs and stated investment priorities in AI technologies. It’s significant that retailers are thinking about AI as part of a future roadmap rather than dismissing it outright, and that speaks to a real shift in mindset. But I’m much more skeptical of any results suggesting deep integration or optimized impact at scale industry-wide. Adoption statistics often blur the line between experimentation and operational maturity, and without clear benchmarking on outcomes — like measurable lift in conversion, reduction in shrink, or improvements in labor productivity — it’s hard to take broad claims at face value.
Given Nvidia’s central position in the broader AI ecosystem — including its dominant role in supplying AI accelerators and infrastructure — it’s reasonable to view the survey data with a grain of salt regarding potential bias in framing and interpretation. Providers naturally want to highlight momentum and broad adoption, and survey questions or reporting lenses can subtly push narratives that align with vendor interests. That’s not to say the data is without value — clearly the conversation around AI in retail has shifted dramatically — but “near-universal” should be read as promotional shorthand rather than a rigorous industry census. The real story is that retail is in transition, with pockets of significant progress alongside plenty of uneven investment, implementation challenges, and genuine work yet to be done.
There seems to be a bit of a disconnect (measured in light years) between the headline and the ‘data’ and anecdotes in the article. How can AI adoption be ‘near-universal’ if the AI talent shortage is 46%…??? I’m sure there is massive interest and questions and curiosity about the Why and How and When and Where to apply AI, but it’s pretty clear we are a long way from ‘near-universal’ adoption. Why not just say near-universal interest and curiosity? No argument there.
“Near-universal” adoption feels like a stretch. Many retailers are experimenting with AI or running pilots, but that’s different from having it embedded into day-to-day operations. The intent and need are there, but the challenge now is deploying AI in a way that fits naturally into everyday tools and workflows.
A lot of the discussion has centered on whether AI adoption is truly “near-universal,” but the more interesting question is how retailers are defining success in the first place.
When surveys report near-universal cost reduction or revenue lift, it is worth asking what is being measured and over what time horizon. Early AI initiatives often deliver visible gains because they are narrow, supervised and insulated from day-to-day operational complexity. Those results do not always persist once AI decisions intersect with store labor variability, inventory accuracy and exception management at scale.
The signal this survey reflects is early impact under controlled conditions, not whether outcomes endure once AI becomes part of routine operations. What ultimately matters is whether performance remains consistent as accountability shifts from experimentation to execution and systems are exposed to everyday operational pressure.
In retail, the hardest problems rarely show up in surveys. They tend to surface months after rollout.
Directionally the data seem plausible. However I would question the magnitude and the sample used. If the “responses from hundreds respondents” sample is comprised of Nvidia clients, well… they’ve spent the money. They’re in the club. Few will say they are investing poorly.
In speaking with those deploying AI, the lack of quality data to inform the AI continues to be an issue. So much of what an AI needs is in the heads of staff. Specialized models need specialized data that’s accurate and well classified. That’s a barrier in a lot of instances to full AI success.
More Discussions
Would an Amazon Shift of Prime Day to June Be a Good Move?
Would it be a wise move for Amazon to move Prime Day ahead of its current…Should Amazon Be Banning AI Shopping Agents From its Platform?
Amazon is clamping down on access to its platform from competing AI agents.Does Macy’s Own Celebrations?
Macy’s has made a name for itself in terms of high-profile celebrations. Will that drive cross-over…
I’ll bet you good money that there are LOADS of retail enterprises that haven’t done anything with AI yet. The leading retailers are all over it, as is the secondary tier, but to suggest that toy store chains with a few stores are racking up AI winnings just isn’t true.
“The data provided indicates that a vast majority, 91% to be precise, of retailers are either actively using or assessing AI tools in some way, shape, or form.” Using ChatGPT to write copy isn’t the same as actively using AI. Independent retailers, whom everyone seems to forget, are not adapting AI the same way the big guys are.
No, but it seems to be the Near Universal topic in the RW universe …today, anyway.
I believe that “near universal” adoption is about AI consideration/prioritization, but not implementation. We are only scratching the surface of AI adoption both from the brand and the consumer sides. While many AI features are being built into software and systems, most retailers still need to leverage the technology features to maximize efficiencies and profit gains. What remains most interesting to me is AI enabled commerce (AI agent conversion) which is now being tracked and measured and will explode within 5 years.
“Near-universal” adoption is a big stretch. Though the claim may hold up technically when you consider that AI is being integrated into everyday tools, like POS, accounting and marketing solutions. And any retailer who has ever written a prompt could be classified as “using AI.” While I accept and acknowledge NVIDIA’s perspective, I doubt that their survey is truly representative of the broad retail industry of which the vast majority are smaller independent/regional operations, and these firms will be well behind the adoption curve.
Reading this article, it’s clear that 90% of retailers are “assessing,” “planning,” or “piloting” AI. That’s very different from actually using it. We’ll get there, but for now, it’s more accurate to say that a conversation about AI is happening in 100% of retailers today.
Retailers are eager to deploy AI and they generally know where they want to apply it. I’m just not sure they’re as clear on which capabilities can truly be enhanced or how to operationalize the technology. My guess is that “mass adoption” will come from AI embedded in the solutions they already use, not from native implementations.
And for the many companies outside Tier 1 that don’t have development talent, that matters. Plenty are still running their BI out of Excel — expecting them to stand up custom AI models isn’t realistic.
I don’t know if “near-universal” adoption is accurate. To me, that means 80-90% (or more) of retailers are using AI. Regardless of whether that is accurate, the outcomes of proper AI use are accurate: operational efficiency, higher employee productivity, and better customer experiences. That’s what AI is supposed to do, and in many use cases, it’s working.
From a technical architecture perspective, “near-universal” AI adoption mostly reflects retailers consuming AI through SaaS platforms and embedded capabilities rather than building or operating AI as a first-class system. What matters less is whether AI exists in the stack and more whether it is integrated into core data pipelines, decision engines, and operational workflows. The most meaningful signal in the data is the move from pilots to scaled deployments, which aligns with rising pressure to prove measurable ROI in areas like inventory optimization, search relevance, and personalization.
I am skeptical of uniformly high claims around cost and revenue impact, because without clean data, strong governance, and tight system integration, AI outcomes tend to plateau quickly. Given Nvidia’s position in the ecosystem, the numbers may skew optimistic, but the underlying shift toward AI as an architectural necessity in retail is very real.
Another day, another lot of AI exaggeration! Again, the devil is in the details. If by adoption we mean anything and everything to do with AI – including things like basic use of ChatGPT and such – then yes, the numbers will be quite high. Even so, we know from our own discussions and surveys that it is not universal. But this kind of thing is akin to saying that most retailers use Excel. It is just not meaningful. The numbers deeply using AI and fully integrating it into operation is smaller and mostly confined to the large enterprises. This area is interesting, but we need to move beyond the headline to understand the impact.
We can nevr be sure the sun will come up, but….
I think the claim of “near-universal” AI adoption in retail, as framed in the Nvidia narrative, is ambiguous at best and overstated at worst. Terminology like “near-universal” sounds definitive, but when you dig into the reality on the ground it’s clear that retailers are far from uniformly advanced in their AI investments and deployments. Yes, many of the largest, digitally savvy organizations have been strategic and aggressive about incorporating AI — whether that’s for demand forecasting, personalized recommendations, dynamic pricing, or computer vision-enabled loss prevention — but that doesn’t translate into consistent, optimized, enterprise-wide AI usage across the industry. For every category leader pushing ahead with agentic AI pilots and AI-augmented workflows, there are countless mid-tier and smaller chains still wrestling with basic data hygiene, siloed systems, and pilot projects that haven’t yet moved into scale or measurable ROI. In that sense, “near-universal” could mean that virtually every retailer talks about AI, but it doesn’t necessarily mean everyone has meaningfully adopted and integrated it into core operations.
The data points that feel most credible — and intriguing — are those that reflect intent and early adoption, such as the prevalence of pilot programs and stated investment priorities in AI technologies. It’s significant that retailers are thinking about AI as part of a future roadmap rather than dismissing it outright, and that speaks to a real shift in mindset. But I’m much more skeptical of any results suggesting deep integration or optimized impact at scale industry-wide. Adoption statistics often blur the line between experimentation and operational maturity, and without clear benchmarking on outcomes — like measurable lift in conversion, reduction in shrink, or improvements in labor productivity — it’s hard to take broad claims at face value.
Given Nvidia’s central position in the broader AI ecosystem — including its dominant role in supplying AI accelerators and infrastructure — it’s reasonable to view the survey data with a grain of salt regarding potential bias in framing and interpretation. Providers naturally want to highlight momentum and broad adoption, and survey questions or reporting lenses can subtly push narratives that align with vendor interests. That’s not to say the data is without value — clearly the conversation around AI in retail has shifted dramatically — but “near-universal” should be read as promotional shorthand rather than a rigorous industry census. The real story is that retail is in transition, with pockets of significant progress alongside plenty of uneven investment, implementation challenges, and genuine work yet to be done.
There seems to be a bit of a disconnect (measured in light years) between the headline and the ‘data’ and anecdotes in the article. How can AI adoption be ‘near-universal’ if the AI talent shortage is 46%…??? I’m sure there is massive interest and questions and curiosity about the Why and How and When and Where to apply AI, but it’s pretty clear we are a long way from ‘near-universal’ adoption. Why not just say near-universal interest and curiosity? No argument there.
“Near-universal” adoption feels like a stretch. Many retailers are experimenting with AI or running pilots, but that’s different from having it embedded into day-to-day operations. The intent and need are there, but the challenge now is deploying AI in a way that fits naturally into everyday tools and workflows.
A lot of the discussion has centered on whether AI adoption is truly “near-universal,” but the more interesting question is how retailers are defining success in the first place.
When surveys report near-universal cost reduction or revenue lift, it is worth asking what is being measured and over what time horizon. Early AI initiatives often deliver visible gains because they are narrow, supervised and insulated from day-to-day operational complexity. Those results do not always persist once AI decisions intersect with store labor variability, inventory accuracy and exception management at scale.
The signal this survey reflects is early impact under controlled conditions, not whether outcomes endure once AI becomes part of routine operations. What ultimately matters is whether performance remains consistent as accountability shifts from experimentation to execution and systems are exposed to everyday operational pressure.
In retail, the hardest problems rarely show up in surveys. They tend to surface months after rollout.
Directionally the data seem plausible. However I would question the magnitude and the sample used. If the “responses from hundreds respondents” sample is comprised of Nvidia clients, well… they’ve spent the money. They’re in the club. Few will say they are investing poorly.
In speaking with those deploying AI, the lack of quality data to inform the AI continues to be an issue. So much of what an AI needs is in the heads of staff. Specialized models need specialized data that’s accurate and well classified. That’s a barrier in a lot of instances to full AI success.