Roku's DSP Spend Grew 40% YoY. The IO Premium Just Got Negotiable.

Roku reported $613M in Q1 2026 advertising revenue on April 30, with ad spend through third-party DSPs jumping more than 40% year-over-year. Programmatic now drives the majority of Roku video ad delivery, including expanded integrations with DV360, The Trade Desk, Amazon DSP, Yahoo, and FreeWheel. CTV buyers can match Roku's direct-sales pricing through their existing DSP, which removes most of the negotiating leverage Roku's IO desk has historically held.

The number that actually matters

Roku's headline numbers are clean. Total revenue of $1.25B (up 22%), platform revenue of $1.13B (up 28%), advertising revenue of $613M (up 27%), and an advertising gross margin of 60.5%, up 400 basis points YoY according to the Q1 2026 shareholder letter. Wall Street ate it: EPS came in at $0.57 against a $0.32 forecast, and the stock popped on the print.

The buried line is the one CTV buyers should actually care about. Ad spend routed through third-party DSPs grew more than 40% YoY, and Roku itself wrote that the majority of its video ad delivery now flows through programmatic partners rather than direct insertion orders. PPC Land's coverage pulled this out of the call. Compare that 40%+ growth to total ad revenue growth of 27%. The math is uncomfortable for Roku's direct-sales team: programmatic is expanding roughly 1.5x as fast as the overall business, which means direct IO is decelerating in absolute share even while the headline keeps climbing.

I think this is the part most coverage missed. Variety, Deadline, and The Wrap all led with the EPS beat and the 100M-household milestone. Those are buyer-friendly numbers, sure. But the 40% DSP figure is the buyer-friendly negotiating number, and it deserves to be the lead.

Why your DSP can now buy what your Roku rep used to gatekeep

For most of CTV's existence, Roku ran a soft two-tier inventory pricing model that buyers were too polite to call by name. Tier one was direct IO through Roku's sales team, which got you premium placements (Marquee Ad Video, the home screen, branded Streaming Stores) and the audience signal that came with first-party Roku data. Tier two was whatever scraps showed up on the open exchange after Roku's reps pulled inventory back to feed the higher-CPM direct deals.

That bifurcation is mostly gone. Roku's own shareholder letter confirms expanded integrations with Amazon DSP, The Trade Desk, Yahoo, FreeWheel, and now DV360 with Google's Confidential Publisher Match. According to The Desk's coverage of the Google partnership, Roku is the first streamer to participate in Publisher Match, which lets DV360 buyers activate Google's first-party signal against Roku inventory.

Charles Collier, Roku's media president, said on the Q1 earnings call that Campaign Manager 360 "can now measure Roku Media regardless of where the advertiser's buy lands."

Read that carefully. The pitch isn't "buy direct because the data is better." The pitch is "buy wherever you want, because the measurement is the same."

That's a meaningful concession from a publisher whose margin used to depend on being the only door into the inventory.

The press-release version vs the buyer's version

Roku's framing is that this is a story about openness and platform reach. CFO Dan Jedda described the 60.5% advertising gross margin as "sustainable and potentially capable of moving higher." They want investors to read the DSP growth as additive: more demand sources buying the same inventory at the same effective rate.

The buyer's version is different. When the same inventory is biddable through five DSPs and one direct desk, prices tend to converge toward the lower of those auction floors, not the higher of the IO rate cards. Roku's gross margin held up in Q1 because direct sales still moved decent volume during the post-Super Bowl quarter, but the trajectory is unambiguous. From what I've seen on the buy side, holdco trading desks have been moving Roku spend to DV360 since the Google announcement in March, and independent agencies started routing through TTD or Amazon DSP a quarter earlier.

If you're a CTV buyer who has been paying Roku's direct desk a 15-20% premium for "Premium" inventory, that premium is a negotiating chip now, not a fixed cost.

Other Q1 numbers media planners should pull out of the letter

A few details worth flagging:

• Marquee Ad Video doubled advertisers and tripled ad spend YoY. That's the home-screen takeover unit, and it's still mostly direct-IO. Demand pressure is real, but the substitution argument is also real. Pinterest just opened CTV inventory it doesn't own through TVScientific, which our Pinterest CTV piece covered last week.
• Non-media/entertainment brands now drive ~30% of Roku ad revenue, an all-time high. CPG, retail, finance, and auto are pulling Roku away from its old streamer-promo base. Buyers in those categories should expect Roku reps to push direct deals harder this quarter.
• Streaming hours hit 38.7B, up 8% YoY, against household growth that's well above that. Per-household engagement is roughly flat, which is the stat the analyst calls didn't dwell on and probably should have.
• Trailing-twelve-month free cash flow of $539M is an all-time high. Roku has the runway to keep undercutting direct-IO rates if it decides to defend programmatic momentum at the expense of margin.

Three moves to make this week if you buy CTV

Ranked by how fast you can actually execute them.

1. Pull your Q3 Roku IO and re-quote it through DV360 or TTD. If your direct rep is quoting $40-45 CPMs on Marquee or branded Streaming Stores, run the same buy through The Trade Desk's PMP and DV360's Confidential Publisher Match and see what fills. Based on the public Q1 data, you should expect roughly a 10-20% CPM gap closing, depending on how much premium signal your campaign actually needs. Don't sign anything new this quarter without that comparison.

2. Re-test DV360 specifically on Roku inventory before June 1. The Confidential Publisher Match integration is new enough that DV360's reach extension models haven't fully optimized against it yet. Early adopters tend to get cleaner audience matches before the auction dynamics tighten and bid density catches up. We covered the broader DV360 reach overlap update earlier this month, and the new dimensions stack well with Roku's audience signal.

3. Ask your Roku rep what the direct-IO premium actually buys you in Q3. If the answer is "Marquee Ad Video, branded Streaming Stores, and the home-screen rotator," fine, those are still genuinely scarce inventory. If the answer leans on "premium placements and audience data," that's now mostly available programmatically, and you should treat the IO commitment like any other negotiable CPM.

A small qualifier before anyone over-rotates

I'll caveat all of this. Direct IO still wins for sponsorship-style takeovers, City Stream-style branded-channel work, and live event inventory like Olympics or NFL Sunday Ticket placements that Roku gets through deals with NBC and Disney. Those aren't biddable through an open DSP, and they're priced like the scarce executive-buyer products they are.

But for the boring 80% of a CTV media plan (audience-targeted reach, mid-funnel video, retargeting), the DSP growth number from this quarter says the negotiating geometry just changed underneath everyone. Roku's investors heard a beat-and-raise. The buy side should hear a permission slip.

Notice Me Senpai Editorial