Srinivas Ramesh

Srinivas RameshSep 5, 2026

PepsiCo just moved a global media account estimated at roughly $1.7B - $1.9B to Publicis Groupe after more than two decades with Omnicom. The number is big. What says more to me is how it happened: no traditional pitch, just a capabilities review. PepsiCo is already simplifying how the business operates, reducing handoffs and pushing data, automation and AI deeper into the organization. Now it wants strategy, planning, activation, connected identity, data and technology working together across 200+ markets. I don’t read this as Publicis Groupe simply having “better AI.” Omnicom has Omni, Acxiom, Flywheel and the capabilities coming together through Interpublic Group (IPG). Publicis Groupe has Epsilon, Publicis Sapient and its broader technology ecosystem. At this level, capability ownership tells me very little. What matters is whether the client still has to stitch it all together. I’ve seen this problem for years. Media, commerce, identity, measurement and data sit in different places. Now AI gets added on top. Everyone calls it integrated while the client is still managing the gaps. That’s not integration. And this is why the PepsiCo move says something bigger about where agency competition is going. If the client is simplifying internally, the agency cannot show up with more complexity externally. The part that matters now is interoperability. PepsiCo already has Google Cloud and Gemini deeper inside the business. Publicis has Epsilon and Sapient while operating across a much broader technology environment. None of this is going to live neatly inside one agency-owned stack. The advantage is no longer owning more technology. It is being able to operate across technology you don’t own without forcing the advertiser to reorganize around you. As identity, commerce data, measurement, media platforms and AI move closer to the advertiser’s own infrastructure, closed architecture starts becoming a constraint. The stronger agency model is one that can integrate deeply while the advertiser keeps control of its data, technology choices and underlying architecture. That is a very different kind of agency moat.

Srinivas RameshSep 6, 2026

🚨 AdTech thinks it knows Amazon. Amazon Leo shows how much of the picture it still isn’t looking at. We already know the obvious Amazon map: Amazon Ads, AWS, commerce, measurement, streaming, devices, AI, logistics. That’s not the part that matters. The part that matters is what happens when the same enterprise sits across several of those businesses at once. That’s why Amazon Leo is important. The satellite story is easy to see. The structure underneath it is much more consequential. Apple can keep the device and customer relationship while Amazon supports part of the satellite layer. Mobile operators can keep the subscriber while Leo provides connectivity. Delta Air Lines already has a major AWS relationship and is adding Leo across 500 aircraft. That structure tells you something important: Amazon can become more relevant inside the same enterprise without every new relationship needing to look like an Amazon-owned customer relationship. That changes how Amazon should be read. A media team can understand Amazon Ads perfectly. A CTO can understand AWS perfectly. Commerce can understand marketplace economics perfectly. Procurement can know every contract. Finance can know every dollar. And the company can still be missing what the full Amazon relationship actually looks like. That is the part I would put in front of a CEO, CMO, agency leader, board or investor. Not: How much do we spend with Amazon? A better question is: How much of the business touches Amazon somewhere, and which of those relationships only look separate because different teams own them? That is where the org chart starts hiding the real picture. For agencies, it changes how you read a client’s Amazon exposure. For CMOs, media diversification does not necessarily mean company-level diversification. For CEOs and boards, it shows where relationships are being split across functions. For investors, it raises a harder Leo question: if connectivity deepens an existing Amazon relationship or broadens an enterprise account, where does that value actually show up? That is why I would not look at Leo only through subscribers, capacity, ARPU and margins. Some of the most important value sits between businesses, not inside one business. Put Amazon Ads, AWS, commerce, streaming, devices, AI, logistics and Leo on one page. Remove the department names. Then ask yourself one uncomfortable question: Are these actually separate relationships, or have we just made them look separate because different executives own them? Because if the answer is the latter, then most companies are not looking at their Amazon relationship. They are looking at fragments of it. And that means the strategy conversation probably needs to happen one level higher than it does today.

Srinivas RameshSep 7, 2026

🚨 Google kept AdX. Now comes the experiment I’ve wanted to see in AdTech for years. Take Google demand out of Google’s ad server and show me the economics. The remedy is headed toward giving competing publisher ad servers real-time access to AdX bids. Fine. Access was never the hard part. Can somebody else take that demand, run it through a different stack and keep the publisher economics intact? Publishers don’t stay with an ad server because they love the ad server. They stay because nobody wants to discover halfway through a migration that more choice came with less revenue. The publisher reaction this week makes that even clearer. Some large publishers are relieved AdX wasn’t forced into a sale. They still want competition. They just don’t want to replumb critical monetization infrastructure while traffic and revenue are already under pressure. When removing a dependency feels riskier than living with it, the dependency itself has economic value. So six months from now, don’t show me how many integrations went live. Show me yield after the switch. If publishers can run AdX demand through competing infrastructure and hold or improve their economics, independent publisher platforms have a real opening. If they can’t, we’ll have learned something more important than whether the connection worked: You can make demand portable without making the market portable. And this is where I’d be paying attention if I ran an independent ad server or SSP. “We connect to AdX” won’t be much of a differentiator once others can say the same thing. Prove that a publisher can move more of its stack away from Google without giving up the economics that kept it there. There’s another piece here. While everyone is focused on opening AdX, Google is also bringing large buyers closer to GAM through Buyer Direct. The route demand takes through the publisher stack is already changing while the industry is still debating the exchange. So the bigger question isn’t just who gets access to Google demand. It’s who becomes best at deciding where that demand goes, what it’s worth and how much more money the publisher makes because of that decision. That’s when this stops being an antitrust story and starts becoming a market-share fight. #AdTech #Programmatic #Publishers #DigitalAdvertising