GM. This is Milk Road AI PRO, where we call the shots before the agents pump the charts.
Retail is entering its most important transition since the rise of e-commerce.
Not because websites are improving. Not because ads are getting smarter. But because AI is moving out of software layers and into the physical and economic core of commerce.
In the last report, we looked at AI adoption across industries and which sectors will be leading the race.
Retail is one of the most promising sectors, with economic pressure driving change across the industry.
Today, we unpack what that change actually looks like. And why it adds up to a $3-5T opportunity.
What weβre covering today:
- Why the traditional retail model quietly broke.
- How AI becomes the operating system of modern retail.
- What AI actually does inside stores and warehouses.
- How e-commerce is evolving into agent-driven commerce.
- Where durable investment leverage sits.
RETAILβS OLD MODEL NO LONGER WORKS
For decades, retail was supply-driven.
Produce in a standardized fashion at scale β push inventory through distribution β hope demand absorbs it.
That model worked when customers looked similar and distribution channels were limited. Success meant better purchasing, tighter logistics, and incremental efficiency gains.
That world no longer exists.
Consumer demand is now fragmented across lifestyles, income levels, cultures, and contexts. Channels have multiplied. Expectations are personalized by default.
Retail as a consequence didnβt just get harder. It became structurally different.
From mass production to mass personalization
Modern shoppers increasingly expect prices, offers, and experiences that reflect their situation, not that of an average customer.
The implication is uncomfortable but simple: one-size-fits-all retail no longer creates value.
Where retailers once focused on getting a raspberry to the customer before it spoiled, the core challenge now is knowing which customer wants which berry, through which channel, at which time.
Doing that well requires millions of micro-decisions across pricing, assortment, inventory, and promotions.
AI AS THE OPERATING SYSTEM OF MODERN RETAIL
Granular decision-making leads to enormous complexity. Solving this with humans in the loop explodes costs - undermining a core pillar of retail profitability.
AI breaks that link.
For the first time, retailers can increase decision density without increasing labor at the same rate.
Pricing can be updated hourly instead of weekly. Assortments can adapt store by store. Promotions can react to real demand, not last monthβs averages.
Thatβs why AI is not a feature upgrade in retail. It is the new operating system.
By 2030, agentic AI alone is expected to create:
- Over $1T in potential value creation in U.S. retail.
- Roughly $3-5T globally.
However, this value will not accrue evenly, and in practice is likely to be significantly higher, as the value of physical retail is not yet fully accounted for.
The impact of AI will unfold across two distinct yet deeply interconnected arenas: physical retail and e-commerce.
Physical retail: margins force innovation
Despite years of digital hype, physical retail still generates more than 80% of total retail revenue. But gross margins vary dramatically:
- Groceries: 1-5%.
- Consumer electronics: 5-20%.
- Fashion: 40-60%.
In low-margin categories, inefficiency is lethal. Every percent matters. This is where AIβs impact becomes concrete.
50-60% of all working hours across the retail industry can be transformed through AI.
Agentic dashboards can deliver automated morning briefings that proactively flag critical KPI deviations.
Instead of manually designing campaigns, sales and marketing teams can review, compare, and approve AI-generated recommendations, such as testing targeted ad placements across selected stores.
And customer service can be handled by calm, always-available chat or voice agents rather than overburdened support staff.
This can help companies both significantly reduce costs and free up human labor for more value-creating activities.
The reclaimed time can be redirected toward real decision-making rather than hours of preparatory work.
Implementing this in the real world requires three core pillars:
- Physical AI enabling machines to understand the real world.
- Physics AI to optimize physical systems and infrastructure.
- Data & AI to improve decision-making.
Physical AI: Enable real-world action
Physical AI combines computer vision, IoT sensors, robotics, and physics-informed models to enable real-time monitoring, predictive maintenance, and operational optimization across the retail value chain.
Next-generation physical world foundational models, such as the NVIDIA Cosmos, can perceive physical environments in real time, anticipate outcomes, and autonomously take or recommend actions.
If a box falls in a warehouse aisle, AI can detect the obstruction instantly and reroute incoming forklifts to maintain production flow, preventing downtime without any human intervention.
This is no longer theoretical. Early-stage physical AI is already operating in production and store environments:
- Trigo tracks in-store behavior to enable frictionless checkout while automatically detecting theft and fraud.
- Schwarz Group, Europeβs largest retail group, uses AI-powered computer vision to monitor bakery production in real time, detecting defects and ensuring consistent quality for up to 100,000 products per hour.
This is not digital innovation theater. It is AI replacing human sensory labor at industrial speed.
As these systems mature, Physical AI will systematically eliminate downtime, shrinkage, and inefficiencies that were previously invisible β or simply too complex for humans to manage at scale.
Physics AI: Scenario simulation
Going hand in hand with physical AI adoption, retailers increasingly simulate physical reality before changing it.
Physics AI refers to AI systems that understand and respect the laws of the physical world, such as space, volume, weight, and motion, allowing them to model and simulate real-world environments before acting on them.
This enables retailers to build digital twins of stores, warehouses, and distribution centers.
With these digital twins, retailers can:
- Optimize store layouts for customer flow, bottleneck reduction, and product visibility.
- Simulate staffing, replenishment, and shelf placement decisions.
- Increase warehouse and distribution center throughput before making physical changes.
Crucially, these simulations increasingly feed directly into Physical AI systems, closing the loop between prediction, simulation, and execution.
Uh, Ohβ¦ π§ The rest of this report is exclusive to AI PRO & PRO All Access members!
Already an AI PRO & PRO All Access member? Log in here.
WHATβS LEFT INSIDE? π
- The implications of AI on retail data and retail media as an emerging revenue stream.
- The rise of agentic commerce and how this will transform online shopping.
- The investor takeaway; what retail-driven investment plays exist.
Upgrade your subscription today to unlock access to all of the milky insights above, PLUS:
- More where that came from! Get two AI PRO reports every month, that boil big-picture shifts down to specific investment plays across chips, cloud, energy infrastructure and more.
- Monthly AI AMAs: Your direct line to our lead analysts. Join our monthly "open mic" sessions to grill us on everything from the latest Tesla FSD data to picking the right energy stock. ποΈ
- Access to the PRO community: Where you can chat directly with our analysts and fellow AI PRO members in our private Discord. Ask questions, share ideas, and stay plugged in between reports. π«
- The AI portfolio (coming soon): A live look at exactly where we are putting our capital to work across the AI ecosystem. π
Already an AI PRO & PRO All Access member? Log in here.
WHAT AI PRO MEMBERS SAID LAST WEEK:
