What emerged wasn’t a market bracing for collapse or charging toward a boom but one inching toward a structural reset, where 2026 looks less like a rebound and more like a reckoning over how advertising is bought, measured and paid for. 

Read on to see our big takeaways from the event.

AI and the CFO firewall

The real brake on AI adoption is no longer technology, it’s finance. In Sir Martin Sorrell’s view, large legacy companies will not materially change how they operate until CFOs feel real economic pressure. Right now, that pressure does not exist. As it stands, companies are too profitable and too comfortable to force change, leaving AI “interesting” rather than urgent and largely confined to test phases and incremental workflow tweaks instead of driving truly autonomous operations. 

“Until CFOs push these companies to do things — so unless there is an external economic tightening of the screw, which there isn’t at the minute — it’s not going to happen in my view,” said Sorrell. 

Search Everywhere Optimization is the new SEO

One of the quieter shifts surfacing at CES was the blurring of discovery and conversion inside conversational and agentic surfaces. Consumer journeys are increasingly moving fluidly across AI-mediated environments — from chat to commerce-enabled social and search the traditional funnel as product questions sit closer to transactions and bypass familiar landing page and retargeting loops. In practice, retail media, social and search are beginning to behave less like separate channels and more like adjacent layers in a single commerce flow.

LLM buying is being publicly demoed — but privately distrusted

Despite the noise around autonomous media buying, Yahoo is drawing a clear line between workflow and decisioning. Its agentic layer is built to compress time — speeding setup, trafficking and optimization — rather than replace the machine-learning systems that actually move dollars through auctions.

Indeed, under the hood the bidding logic remains ML-led with agents sitting above it as a service layer that makes the Yahoo DSP platform easier to use, easier to move between and easier to plug into other systems. In other words, the Yahoo DSP is becoming less a buying dashboard and more of a data spine with productivity layers attached.

“Agentic is cool but it’s really still about the data and the machine learning,” said Adam Roodman, GM of Yahoo DSP. 

Agencies are trying to turn DSPs into interchangeable pipes 

For most of the past decade, DSP competition was governed as much by inertia as by performance. Certification programs, proprietary workflows and platform-specific algorithms made it costly — operationally and culturally — for agencies to move spend once teams were trained and pipelines were in place. 

What’s emerging now is a thinning of the moat. Agencies are increasingly building custom decisioning, filtering and scoring layers upstream of the DSP, allowing their proprietary logic to travel with them — and making execution across platforms more portable and competition among DSPs more fluid. 

“There’s no question agencies will continue to still use agencies but they are saying upstream of those platforms ‘what kind of decisioning can I introduce and apply to them’,” said Michael Richardson, vp of product marketplaces at Index Exchange. “Whereas before they’d be relying on whatever tech the DSP provides now advertisers and agencies are trying to customize that for themselves.”

2026 will be a modest growth recalibration year, according to ad execs at CES. 

After a year shaped by tariff anxiety, geopolitical noise and ad spending cuts, spending drifted back to the market rather than accelerated, leaving advertisers heading into 2026 with cautious optimism — preparing for downside while quietly penciling in slight upside. The result is a market where topline spend inches up but allocation logic shifts underneath with dollars continuing to move out of linear into CTV, performance budgets concentrating into fewer data-credible platforms, Amazon pulling share from both DSP rivals and linear and upfront commitments thinning in favor of flexibility. 

“We’re starting to see that people are working independently of what is happening with the economy,” said Tom Swierczewski, vp of media investment at Goodway Group. “They’re just preparing for the worst, but assuming the best. And so I’m anticipating a slight shift in general media spend this year in the U.S.”

The story isn’t that marketers spent more. It’s that they spent smarter, and with less anxiety about having all the answers upfront. That’s how Paul Boruta, CEO and founder of ad tech platform Slingwave put it: “What we’re seeing heading into 2026 is cautious optimism, but it’s a different flavor than past cycles. It’s not optimism rooted in stability or certainty. It’s optimism rooted in adaptability. Brands have spent the last few years getting comfortable with uncertainty, learning to plan in shorter cycles, test faster, move budgets more fluidly, and stop waiting for the perfect signal before making a call.”

Numbers to know

46.3%: Percentage by which Reddit grew its U.S. ad spend YoY last November, topping Instagram’s 22% and TikTok’s 9% YoY growth

$20 billion: The total amount raised in xAI’s latest funding round, despite controversies and backlash against Grok

40%: Percentage of marketers globally that use AI for social media management

200,000: Number of U.K. small businesses that have signed up to TikTok Shop

What we’ve covered

OpenAI’s countdown: monetization, ads, and a Google-shaped threat

Last year, with Google preoccupied with antitrust trials, OpenAI had a free run at the market. But with one out of the picture, and the other coming to an end soon, Google’s sights are firmly back on track. And with it burning through so much cash, OpenAI needs to fast track launching an ad business to fund its projects.

Pitch deck: How Amazon is emerging as the proof layer for TV spend

While it’s not the default buying layer for TV, Amazon is positioning its CTV offering as a “first-stop shop” for advertisers, making it the easiest place to make TV spend inside organizations — and its recent pitch deck provides that picture.

Virality is no longer just a vibe at MrBeast’s Beast Industries

MrBeast is hiring a head of viral marketing, who will be responsible for launches, customer acquisition, conversion, retention and marketing efficiency, across the creator’s portfolio, signalling that this is no longer just a nice bonus, but a formal way to grow his company.

‘This is what the future will look like’: Accenture Song has moved upstream of advertising

Accenture’s marketing arm is now comparable to the largest ad agency holdcos and what’s emerging is a control layer in marketing services — one that increasingly determines how marketing labor, data and tech are orchestrated inside large enterprises.

What we’re reading

OpenAI reserves $50 billion for stock grant pool

Last fall, the AI company set aside a $50 billion employee stock grant pool (worth 10% of the company’s $500 billion value) — far exceeding compensation at the likes of Google and Meta, but indicates that OpenAI will continue to spend heavily on and reward talent, according to The Information.

YouTube has a firm grip on daytime TV

The reason behind YouTube’s lead over streaming competitors like Netflix, is thanks to its dominance during daytime hours, according to The New York Times.

Reddit overtakes TikTok in U.K. thanks to search algorithms and Gen Z

Reddit has now overtaken TikTok as the fourth most visited site in the U.K., as a result of changing search algorithms and Gen Z, according to The Guardian.

Pinterest announces new CTV show

Following its acquisition of tvScientific last month, Pinterest is launching a six-episode series on Roku called Bring My Pinterest to Life on Roku in March, according to Social Media Today.