Michael Barrett told tipsheet on July 16, 2026 that the most optimistic industry forecasts put agentic advertising spend at $600 million to $700 million by 2027, a figure he described as modest against the wider programmatic market, while artificial intelligence has already absorbed contractor work inside Magnite itself.

The interview, conducted by John Ebbert and published on tipsheet.ai on July 16, 2026, covered the sell-side platform's positioning for AI-driven buying, the economics of connected television, the company's Walmart partnership, recent leadership changes, and the metrics Barrett expects the company to be judged on by 2030. The full conversation is available at tipsheet.ai.

What emerged is a set of claims that are unusually specific about timing. Agentic buying is not, in Barrett's account, arriving at scale in 2026. Nor is it arriving through a rebuild of existing infrastructure.

A trusted layer between two sets of agents

Asked what role he wants Magnite to play as AI reshapes advertising over the next five years, Barrett described a position rather than a product. "We'd love to be positioned as the trusted infrastructure layer where AI-driven advertising is executed across the open internet," he said.

The premise underneath that ambition is that agents proliferate on both sides of the transaction. "We fundamentally believe that every publisher will have an agent and every buyer will have an agent," Barrett said. "The key for us to stay relevant on the sell-side is to remain the trusted layer between them."

Scale is the argument for an intermediary. Even capable agents, in his framing, still have to reconcile tens of thousands of global media companies with tens of thousands of buyers. That reconciliation is what Magnite Orchestration, launched on June 11, 2026 with dentsu and DIRECTV Advertising as first beta partners, was built to handle.

Barrett's forecast for the supply-side category is blunt. "I think there will be far fewer of them, and they'll need to be highly sophisticated technically to help publishers and buyers navigate this transition," he said, adding that platforms built around inventory arbitrage, undifferentiated demand and shallow publisher relationships "will struggle."

The SSP label, and why he keeps it

Recent product announcements have pushed Magnite well past the description that attached to it after the Rubicon Project and Telaria merger. Barrett said the terminology does not concern him. "I'm completely comfortable with the SSP label. We work for publishers, we're paid by publishers, and our responsibility is to create a transparent, open and fair transaction layer on their behalf."

What has changed is the shape of the product. According to Barrett, the streaming business "looks nothing like the traditional Rubicon Project model built around open web header bidding." It is modular. Most streaming publishers, he said, select the capabilities they need rather than adopting the full stack, with some using primarily the ad server and others taking different combinations.

That modularity has a commercial history. Magnite merged its SpringServe ad server with its SSP technology in April 2025 and unified curation and activation inside ClearLine on October 1, 2025, drawing on technology from its September 2025 acquisition of streamr.ai.

Two forces acting on connected television

Barrett's strongest claim concerns television economics. "AI is going to usher in a golden age for television," he said, predicting that within five years television will effectively mean connected TV, because consumers have already chosen streaming for entertainment, sports and news.

He identified two mechanisms. The first is outcome-based buying, visible across this year's upfront presentations. If agencies are compensated on outcomes, Barrett argued, they stop buying Nielsen ratings and CPMs alone, and media companies have to become partners in delivering results. That, in his view, is not achievable without AI in the middle of the system. Research published on May 14, 2026 found that 78% of advertisers planned to factor addressable TV into 2026 upfront negotiations, with outcome-based accountability a central negotiating lever.

The second mechanism is supply arithmetic. Even if streaming absorbed every dollar that once went to linear television, which Barrett said it will not, there would still be more inventory than that spending could fill. New advertisers have to enter the market for the economics to work.

AI, he said, removes several of the barriers that kept them out: creative costs, campaign complexity and measurement difficulty. Businesses that could not previously advertise in streaming media "are signing up with credit cards, launching campaigns and becoming increasingly sophisticated." Magnite's acquisition of streamr.ai was aimed squarely at that segment, and Barrett described the barriers at the time as complexity and high cost.

Nielsen's 2026 Upfront Planning Guide, published March 12, 2026, put streaming at 66.7% of ad-supported television time among adults 18 to 49, a figure consistent with the direction Barrett described.

What an orchestration layer has to settle

Pressed on how mediation, agentic execution and Orchestration fit together, Barrett described a layer sitting above the transaction itself. "It's not just about agents discovering inventory," he said. "The orchestration layer also handles settlement, policy enforcement, pricing, yield management, identity, payments and compliance. Those responsibilities can't simply be left to agent-to-agent interactions."

His analogy was the New York Stock Exchange, where floor trading gave way to electronic execution but the clearing function remained. "You still need a trusted exchange to ensure that when someone buys a share of General Motors, the seller actually owns it, the buyer receives it and everyone gets paid," Barrett said. The infrastructure built around trust and transparency, in his phrasing, becomes the rails on which AI operates.

Magnite is not alone in claiming that layer. June 11, 2026 produced four separate agentic launches in a single day, including Mediaocean's NIVO AI, Teads EngageOS and the Walmart Connect partnership with Google DV360.

ClearLine's changing purpose

ClearLine began as a self-service platform for midsized agencies without demand-side platform relationships, converting insertion orders into programmatic campaigns automatically. Most of those agencies now have DSP access, and Barrett said the product's role has shifted.

It now serves two functions. The first is a buying tool giving advertisers access exclusively to Magnite supply, which Barrett described as useful for fee-sensitive spend that would otherwise stay outside programmatic. Because the company's economics come from the sell side, he said, there is no need to maximise economics on the buying side.

The second function is the one he considers strategically more important: an entry point for large agencies reaching premium private marketplaces on Magnite. "People often think of ClearLine as a low-cost DSP alternative," Barrett said. "It certainly can serve that purpose, but it's much more strategically important because it's embedded across many of the workflows we provide." That embedding has been extended repeatedly, including through the April 15, 2026 AMC Global Media agreement that brought linear television inventory into programmatic buying through a single ClearLine entry point.

Protocols: still in the discovery phase

Magnite was an early supporter of the Ad Context Protocol, the open standard built on Anthropic's Model Context Protocol that launched on October 15, 2025 with six founding companies. Asked where the industry sits on adoption, Barrett was direct: "We're still in the discovery phase."

Buyers and sellers want proof points, he said, "even if it's only a $100 transaction." Once a counterparty completes one successful agentic transaction, in his account, it is committed.

What he has not encountered is appetite for demolition. "What we haven't seen is any pressure from our partners to throw away today's infrastructure and rebuild everything from scratch," Barrett said. "Publishers trust our rails. Buyers trust our transparency. Everyone knows transactions settle correctly and people get paid." The prevailing question he hears is why the industry would not layer agentic capability onto working infrastructure instead of starting over.

The numbers he cited are the sharpest part of the interview. Asked when protocol-based buying reaches meaningful scale, Barrett said the answers he collected at Cannes ranged from essentially no meaningful DSP adoption by 2027 to $600 million or $700 million in spend, and that the outcome sits somewhere between the two. "Even at the high end, that's still relatively modest within the context of the overall programmatic market," he said.

Measured performance so far supports caution rather than acceleration. A DataBeat report published June 22, 2026 found conventional programmatic buyers holding a 13.4% CPM advantage over AI agents, running against earlier vendor-reported results.

The one-to-many threshold

Magnite built a seller agent into SpringServe in December 2025 and added a buyer agent on April 27, 2026 with Disney Advertising, Spectrum Reach, Kepler and MiQ. What that infrastructure currently executes, Barrett said, is one-to-one activity: a buyer working directly with a seller, an insertion order that used to require manual handling now running through an agentic workflow.

He described that as already meaningful, and as an expansion of the addressable market for programmatic, because direct insertion order deals that historically sat outside the channel can now move into it. The threshold he is watching is different. "When a buyer's agent can simultaneously discover, negotiate and transact across many sellers, that's when the industry will know the technology has truly scaled."

Internally, the productivity effect is already recorded. Tasks that once required large numbers of contractors are now automated, Barrett said, and those roles have not been backfilled: "We haven't needed to replace those people because AI is doing much of that work."

Conversational media and third-party demand

On advertising inside AI-native environments, Barrett avoided naming companies but described a structural change in sequencing. Historically, platforms built direct sales operations first and opened to third-party demand later, if at all. "Now we're seeing companies begin with third-party demand instead," he said, calling it encouraging for the programmatic ecosystem because building an advertising business from scratch against incumbents is difficult.

That pattern is visible in the market. OpenAI's ChatGPT advertising pilot, which began on February 9, 2026, opened a self-serve Ads Manager to all US businesses on May 5, 2026 with cost-per-click bidding, and extended to the United Kingdom on June 6, 2026. Technology partners including Criteo, Kargo, Adobe, Pacvue and StackAdapt were admitted early.

Magnite's involvement, Barrett said, typically comes later, once a platform moves from one or two DSPs toward hundreds of demand partners. He cited Pinterest and Roku as precedents for the same progression from closed approaches to broader third-party demand.

Walmart, Vizio and sell-side decisioning

Barrett referenced two Walmart announcements, on April 28 and May 28, and said the more consequential development discussed at Cannes was the evolution of Walmart's thinking about its own first-party data.

His description of retail media history is unsentimental. The first inning, he said, involved handing valuable data to a single DSP and requiring advertisers to use that platform for access. The data was valuable enough that advertisers accepted the constraint.

What changed is the recognition that the constraint is optional. Data can be moved out of the DSP, placed with a trusted sell-side partner, positioned closer to premium inventory such as Vizio, which Walmart owns, and activated by multiple DSPs. On May 28, 2026, Walmart Connect made its first-party shopper audiences available inside Yahoo DSP routed through Magnite's supply-side technology and covering Vizio connected television inventory. Walmart Data Ventures launched the Scintilla Media Data Feed on April 28, 2026, opening up to 500 retail data elements to agency and technology partners.

"I think the public markets are gradually recognizing that SSPs aren't simply dumb pipes bringing undifferentiated demand," Barrett said. "They play a much more strategic role than that."

The same logic, he argued, extends past connected television. Buyers do not want to manage six or seven specialised exchanges for different inventory types, and combining channels in one platform produces a stronger offering than treating them as silos.

Where publisher value accrues

Large streaming platforms can strike direct data partnerships. Barrett named Netflix, Disney and Roku as companies with sufficient scale to combine their own data with retailers such as Walmart.

For most publishers, the benefit is indirect and arrives as pricing. Inventory activated with valuable first-party data becomes more valuable, and in bidding environments prices can rise further, allowing publishers to monetise inventory that previously went undervalued because the relevant data was not attached to it.

Barrett expects retail media to keep pushing outward from owned-and-operated properties. Most retail media advertisers, he said, are television advertisers at heart, and brands such as Gillette ultimately want placement on Netflix, Disney, Paramount and other premium streaming services. Walmart's ownership of Vizio, in his reading, is where that expansion becomes visible.

Leadership, and the test Barrett sets for 2030

Asked about recent executive departures, Barrett reframed the question around durability. Magnite was assembled through the acquisitions of Rubicon Project, Telaria, SpotX and SpringServe, and leadership teams built that way, he said, do not usually hold together for a decade. Most employees have only ever worked for a company called Magnite.

Against the current phase of the industry, he said the company is creating scope for Katie Evans and Sean Buckley to take on additional responsibility. Both were promoted to president-level roles in December 2024, Buckley to president, revenue and Evans to president, operations.

The closing exchange produced the interview's clearest accountability standard. "There's a lot of discussion around AI today, but many companies are talking about AI while their costs continue rising and their revenue growth slows," Barrett said. "If we're still talking about AI in 2030 but our growth is flat and our headcount has ballooned, then we've missed the point."

He put a number on what passing would look like, without committing to it: "Whether it's 50% growth with 20% margins or some other combination, AI should improve both growth and operating leverage." His summary was terse. "Until that happens, AI is just a talking point. It's not reality."

Why this matters for the marketing community

For media buyers, the interview functions as a schedule check. A sell-side chief executive with direct visibility into agentic transaction volume is placing meaningful protocol-based spend somewhere below $700 million in 2027, which means the operational case for rebuilding buying workflows around agents in the next eighteen months rests on efficiency gains rather than reach.

For publishers, the consolidation warning carries a practical edge. Barrett expects a smaller field of supply-side platforms and a widening gap between those with deep publisher relationships and those reselling undifferentiated demand. That expectation is being tested in public financials: Magnite reported Q1 2026 revenue of $164.4 million and net income of $4.4 million on May 6, 2026, with connected television crossing 51% of contribution ex-TAC for the first time, while PubMatic has made louder agentic claims on a smaller revenue base.

For retail media operators, the sell-side decisioning argument is the one with immediate budget consequences. If first-party data delivers more value sitting next to premium supply than locked inside a single DSP, the exclusivity arrangements that defined the category's first phase become harder to justify.

Timeline

Summary

Who: Michael Barrett, chief executive officer of Magnite (Nasdaq: MGNI), interviewed by John Ebbert of tipsheet. The conversation references Magnite executives Katie Evans and Sean Buckley, partners including Walmart, dentsu, DIRECTV Advertising, Disney, Netflix and Roku, and platform precedents including Pinterest.

What: An interview setting out Magnite's positioning as an infrastructure layer for AI-driven advertising, in which Barrett forecast far fewer surviving sell-side platforms, described protocol-based buying as still in a discovery phase with 2027 spend estimates ranging from negligible to $600 million or $700 million, said AI has absorbed contractor work internally without backfilling, and argued that first-party data creates more value stored on the sell side than inside a single DSP.

When: The conversation took place on July 15, 2026 and was published on July 16, 2026. It references Walmart announcements on April 28 and May 28, 2026, the June 11, 2026 launch of Magnite Orchestration, and a 2030 horizon for judging the company's AI claims.

Where: Published by tipsheet.ai. Magnite is headquartered in New York with principal offices in Los Angeles, Denver, London and Sydney, and the discussion covers connected television, retail media and programmatic display across global markets.

Why: Agentic and protocol-based advertising has attracted a large volume of vendor claims with limited disclosed transaction volume. Barrett's estimates place meaningful scale beyond 2027 and locate the near-term benefit in workflow automation and expansion of the advertiser base rather than in displacement of existing infrastructure, a position that carries direct consequences for how publishers, agencies and retail media operators sequence their own investment.