Eleven 2026 Agentic Commerce Predictions🔥🚀

What are the big moves we expect in 2026 from ChatGPT, Google, Microsoft, Stripe, Amazon, Xai, Anthropic, and Meta? It all kicks off at NRF next week, buckle-up Retailgentic readers!

🔥🚀Happy New Year! We’re back after a great break and some much needed battery re-charging after the most intense ecommerce news Holiday peak season I can remember after being in the industry for 26yrs. We thought about easing into it with a retrospective, but in the Agentic Era, we don’t look back we only look ➡️, so forget that - we’re going to jump right into the deep end with our 2026 Predictions.

The Challenge Making Agentic Commerce Predictions

Thanks to the Jason and Scot show (10+yrs of annual prediction shows) and being an entrepreneur in this space I have a lot of experience and scars making annual predictions about digital retail. Over that time I’ve gotten progressively better at it (I think I’m 5 to 1 vs. Jason the last 6yrs).

Then the agentic era began on 11/18/24 when Perplexity launched ‘Buy with Pro’. Ever since then I’ve been making predictions, and many. have come true, but what’s particularly hard is the timing. I used the break to reflect on this and research and what I discovered is this space is moving at about 4X speed from what I got used to since around 2005.

That’s my approach heading into these ‘26 predictions. I think there’s a 60% probability we stay at this speed. There’s a 30% probability we accelerate from here and my 4X becomes a 6X. If we burn through these predictions by May-July, we’ll know that the treadmill has accelerated several notches.

That leaves a 10% probability that we hit a slow-down. I score this very low probability because underneath our Agentic Commerce world at the ‘bottom of the stack’ are the state-of-the-art foundation models. According to Goldman Sachs, heading into 2026, there is already $527b committed to GenAI data centers and GPU/TPU/etc. compute buildout. That number is probably going to go up. On top of that bullish signal, we haven’t even seen the first Nvidia Blackwell-trained models come out yet, and they are already shipping the next-gen Rubin. That spend plus the unknown growth of GPU capability multiplied into that is why I think there’s a pretty material chance we could stay on the log-scale here which would be the 6X acceleration. Welcome to predicting Agentic Commerce in 2026 - exciting times 🚀!!

Retailgentic’s 11 Agentic Commerce Predictions for 2026

Without further adieu, here are our eleven 2026 Agentic Commerce Predictions:

1. Death of clicks.

2. Explosion of Agentic Commerce Ad Experiments

3. Cross-agentic checkout/payment protocol bridge emerges

4. Retail Media Networks feel agentic pain

5. Google Agentic 2.0

6. Perplexity/PayPal 2.0

7. Claude/Anthropic come out of nowhere

8. Grok/xAI make a move

9. Apple stays out of the game

10. Meta makes it to ‘find’

11. ChatGPT Instant Checkout++ first player it to ‘Advanced Buy’

Through the year we’ll track how we do against these and also log any ‘out of left field’ surprises.

In the rest of this post, we’ll go through each prediction with some detail laying out the thinking and breadcrumbs that led to the prediction. The biggest source of breadcrumbs is already hear next week: NRF.

The NRF Big Show’s Influence on the Predictions

Because of it’s roots in physical retail going back to January 1911, in my era, the NRF Big Show has always been 70-90% store and 10-30% digital. This year Agentic Commerce is flipping that on it’s head.

For the first time ever, the Google CEO is doing a keynote at NRF. This is huge because Google has benefited immensely from our vertical and now that they have a challenger they are waking up and coming to talk to us. Recently NRF updated the keynote section and imagine my surprise to see that Sundar is sharing the stage with non-other than John Furner, the incoming CEO of Walmart. This session is 9:45-10:15 on Sunday morning so if you aren’t coming to NRF, you may want to check your favorite Agentic Commerce substack in the 10-11 window to see what news broke.

My best guesses as to what’s going on here→

• Google announces an ‘upgrade’ or 2.0-ification to their Buy for Me functionality. We’ve documented the ‘percent-of-a-percent-of-a-percent-of-a-percent’ issues in the current implementation. This would include Walmart joining the new Buy for me functionality.

• Walmart announces they are joining AP2 protocol alliance (putting them in both the ChatGPT and Google camp).

• Walmart announces they and Google are partnering on an expansion of their RMN (Walmart Connect).

Also at NRF, several of the Agentic players are having private workshops, meetings, dinners, fireside chats including Perplexity, Stripe, Microsoft, Google, PayPal and ChatGPT which you’d expect. The one unexpected one that caught me off guard, is Anthropic.

Over the break, I spent a good bit of time with Opus 4.5 (released 11/23/25) on some unrelated to Agentic Commerce and that model is insane. If Anthropic adds a product catalog component along with a MCP-based protocol for checkout, it could be a serious threat to the current top players - ChatGPT, Google and Perplexity.

While not NRF, this came out this week - a Stripe/Microsoft webinar set for Jan 27 that could be an interesting development:

Prediction 1: Death of Clicks

According to Similarweb, across ecommerce retailers, in 2025 10% of clicks evaporated as consumers adopted Answer engines and Google converted more and more traffic to ‘answers vs. clicks’. As we talk to brands and retailers, there us a ‘pull’ from agentic commerce as an incremental new opportunity, but just as strong, maybe even more is the ‘push’ of the evaporation of both paid and organic traffic - particularly from Google. We’re hearing whisper numbers for 2026 predicted impact that are 30% organic/seo and 15% paid. This leaves a huge hole in the 2026 forecast to fill and the obvious choice to fill that hole, and potentially more than fill it is Agentic Commerce. If you do that the loss of clicks isn’t felt because it basically converts to up-stream transactions.

One brand used the phrase - this is the year clicks go away.

Prediction 2: Agentic Ad Experiments

With that previously mentioned $527b in infrastructure spend comes increasing pressure to keep growing revenue at a crazy pace. Here’s the Information’s latest revenue chart with the ‘Winter 2025 Projection’ that ChatGPT put in front of investors to raise their last round:

Focusing on 2026, ChatGPT is at a $13b run-rate at the end of 2025. You can see here the prediction is for 2026 to more than 2X that to $29b - it’s not clear if this is ARR (end of year run-rate) or total year’s revenue. In any case, they need $16b of new revenue and given what we saw in 2025, I’m sure they would like to crush this number.

The most likely place to look for this revenue in addition to Agentic Commerce is rolling out an ad product, particularly to monetize their free tier. That’s the ChatGPT view, everyone, but Google, faces this same challenge. Therefore I predict we’re going to see 10-20 different innovative ad formats

Prediction 3: Cross Agentic Commerce Protocol Emerges

In 2025, we had MCP (and derivatives) gain steam as a protocol for Agentic Commerce, then Google and PayPal launched AP2, then ChatGPT and Stripe came out with ACP. There’s a ‘high ground strategy’ where someone connects all of this together and builds the ‘one protocol to rule them all’. Nobody wants to have to build and test against 2, 3, 4 different systems, they want to be ‘one and done’.

Who could it be? Let’s think about potential candidates.

Potential Candidates

To build this, it needs to be one of the players in Agentic Commerce

Need to be both a player and in payments as my first choice, otherwise they need to be one of the players that’s established a ‘agnostic plays well with everone’. Out of the players in the Agentic Commerce horizontal game, Google is the only one that checks the boxes of being a player and in payments.

Then on the agnostic side, Stripe is committed to ChatGPT/ACP, so they are out, that leaves PayPal. While they are partnered tightly with Perplexity, they also were heavily involved in the Google/AP2 alliance. Another two agnostic firms in the mix are Shopify and Walmart.

Why is the cross-agentic commerce position the high ground?

If (and it’s a big if), a company can pull this off, if they can even get 10-20% of the transaction volume going through either ACP and/or AP2 they will have a unique view of the data - what’s selling better via ACP vs. AP2, what merchants are available where, and so on. In the land of AI, data is King.

Prediction 4: Retail Media Networks (RMN) Feel Agentic Pain

Amazon is by far the largest, but I think it will actually take share because the lion’s share of their traffic is ‘direct to Amazon’ - thanks to Prime and their position in the heart and wallets of hundreds of millions of consumers.

That being said, the tier-2 RMNs: Walmart, Target, BestBuy, Kroger, etc. I believe are going to (because of Prediction 1 - the death of clicks) start to really feel that in the form of decreased revenues. If we see traditional clicks decrease 25%, it goes to say that the RMN will be nearly 100% correlated to that - 20-25% seems very likely.

Unfortunately, I don’t see a World where this revenue is replaced ‘up-stream’ at the GenAI layer - they are going to want to keep 100% of that unless there’s some clever shared network move (could be what Walmart/Google announce?), but I think that’s low probability.

Predictions 5-11 - Players make Moves Up the Agentic Autonomy Ladder

These predictions are all moves by the players on the race to Level-5 ‘full autonomy on the Agentic Commerce. Let’s start the analysis of these 7 predictions by reviewing where we ended 2025:

Agentic Commerce Player Positions - End of 2025

Here’s our latest tracker:

One surprise of 2025 was Microsoft not moving to Level-3. While Perplexity and Paypal ‘leveled up’ Perplexity’s offering, it really got them to Level-3 parity with ChatGPT (once more merchant’s roll out).

Player positions by Holiday 2026

Putting together Predictions 5, 6, 7, 8, 9, 10, 11 and updating the table, here’s where I think we end 2026:

Google Agentic 2.0

This is influenced by both the NRF keynote and the fact that objectively Google’s ‘buy for me’ functionality is several steps behind ChatGPT’s Instant Checkout, they need to catch up. The timing is tricky, it may not be at NRF, but historically Google has had their I/O conference in May, if we don’t get ‘Google Agentic 2.0’ at NRF, I feel confident it’s coming in May. Google simply can’t lose this to ChatGPT.

Perplexity/PayPal 2.0

The Perplexity/PayPal ‘Instant buy’ implementation cuts in half the distance inside Level-3 between ChatGPT/Instant Checkout and Perplexity’s ‘Buy with Pro’, but they have a ways to go to catch up to, and potentially surpass ChatGPT in 2026.

Claude/Anthropic Enter the Arena

Anthropic being at NRF hosting meetings makes me feel like they wouldn’t do that just to stimulate some enterprise usage and it’s a signal they are considering or about to enter the Agentic Commerce Arena. They birthed agents and MCP and with the solid foundation of Opus 4.5 and their strong design aesthetic, I think they could come out at parity with ChatGPT Instant Checkout and maybe surpass in some clever/unique ways that would add a ton of value to the industry.

Grok/xAI Enters the Arena

X now has their license to transmit money in 41 states - 9 states to go! Elon is so active and controversial, I think everyone forgets that 27yrs ago in 1999 he started this business X that merged into PayPal. Elon saw digital wallets 27yrs ago. He saw the need for GenAI and founded OpenAI - regardless of what you think of him, his track record predicts they will be a force to be reckoned with in Agentic Commerce. I think 2026 is the year, probably on the back of Grok 5.

Apple stays out of the game

As a long-time Apple fanboy, it pains me to say this, but aside from a slightly ok upgrade to Siri this year, I don’t see Apple getting into the Agentic Commerce game in 2026. This is the one I hope I’m wrong about the most. Come on Apple, prove me wrong!

Meta Makes it to Level-2 (Find)

Meta made a big move over the holiday and acquired one of my favorite AI agentic tools, Manus. It’s more in the business productivity category than Agentic commerce. I think we’ll have to wait until 2H of 26 to see movement from Meta. They have a ton of wood to chop. They need to:

• Get a new foundation model build, trained and released and iterated at least to 2.0, 3.0 would be better.

• Build out the shopping infrastructure inside of there.

• Iterate on THAT a turn or two before November 1, 2026.

AI speed makes that possible, but I think they run out of time and make it only to Level-2, which is better than being stuck on Level-0 and sets them up for a bigger move in 2027.

If Facebook gets serious and continues their M+A strategy, I would acquire Perplexity if I were them - before Apple takes a run at them.

ChatGPT Instant Checkout++ - The First Competitor at Level-4 (‘Advanced Buy’)

While all the other players push to catch up to ChatGPT’s position at the top of Level-3, I think they continue to pour on the heat and add features and functionality that pushes them to Level-4. My top candidates for those enhancements are:

• Increased consumer personalization and predictive capabilities

• Integration with merchant loyalty programs

• Instant Checkout goes multi-item AND multi-merchant

• Expanded partnerships with PSPs outside of Stripe

ACP further shared across ChatGPT Agent (virtual and inside Atlas), ChatGPT Atlas, ChatGPT Apps and the in-line experience.

Want to read some hints on where this could go? Checkout Fidji’s Substack post about ChatGPT becoming a highly personalized personal assistant in 2026→

Closing the capability gap between frontier AI and everyday use in 2026

Bonus Prediction: Amazon Promotes Rufus

Thanks for making it this far! As a reward for hanging in there on a long post, here’s a bonus prediction. This one is based on personal usage. I’ve been increasingly using Rufus on Amazon and reading the tea leaves on Jassey’s Q3 gushing over Rufus and the third-party data from Sensor Tower, Rufus is clearly a better discovery mechanism than the legacy search experience. I think that by Holiday 26 Amazon makes a bold move an replaces all or a 50%+ portion of the search experience with Rufus. To make this move, they’ll need to figure out the Rufus ad models so they don’t lose all of that revenue, but if they can even get to within 10-15%, I think they make the move.

You’ll also see them (Prediction 2) experimenting like crazy in Q1-Q3 with new ad units inside of Rufus. If you loop in Amazon, we could see on the order of 20-30 new ad unit experiments this year. It’s going to be wild to watch and try and keep track of.

The timing of this one is tricky, it’s 51%/49% probability for me and, I believe, the riskiest prediction of the 12.

What Do You Think?

What predictions to you agree with and which ones do you strongly disagree with? Do you have any predictions to throw in the mix? Sound off in comments→

Extra Bonus!

I hear frequently from subscribers that you would like more downloadable content from Retailgentic you can mix into your internal presentations and communications. Here’s a PDF with the predictions on one handy page along with the 2025 and 2026 levels of autonomy charts - feel free to use these internally, our only ‘ask’ is that you keep the Retailgentic attribution in there and send some subscribers our way if you find the content helpful.

After NRF, I’ll be doing a big drop of some of our best content so you can leverage this for any internal strategic planning sessions to get everyone up to speed on Agentic Commerce.

Retailgentic Eleven Agentic Commerce Predictions For 2026

706KB ∙ PDF file

Download

Download