Viant Survived Myspace. Now It Wants to Prove Your TV Ad Was Actually Watched

At first glance, Viant Technology sells the least romantic object in advertising: a demand-side platform. It is the software buyers use to choose audiences, bid for ad inventory, manage frequency, and decide whether a campaign produced anything more useful than a lovely spreadsheet. The Irvine company calls itself exclusively buy-side. In plain English, it works for the advertiser, not the publisher selling the slot.

That distinction matters because programmatic advertising is a bazaar of exchanges, resellers, data vendors, verification firms, and platforms taking turns touching the same impression. Viant's sharper pitch is to collapse more of the decision into one place. A marketer can plan and buy connected TV, streaming audio, digital billboards, mobile, and desktop; bring in first-party customer data; ask the system to optimize; and measure online sales, store visits, or other outcomes afterward. The buyer can drive manually, hire Viant's managed-service team, or let its AI products take more of the wheel.

The three-signal machine

Viant becomes interesting when you stop thinking about it as a generic ad-buying terminal and see the three proprietary signals it is assembling. Household ID tries to answer who is reachable. IRIS_ID tries to explain what is on screen. TVision tries to measure whether a person paid attention. ViantAI and Outcomes then use those signals to decide what to buy and how to adjust.

Household ID is the foundation. Viant says its identity graph turns signals from more than 1.8 billion IP addresses and other identifiers into roughly 125 million homes linked to an estimated one billion devices. The unit is the residence, not a browser cookie. That helps a brand reach a household across a television, phone, laptop, or audio stream, control frequency across those screens, and connect an ad exposure to an online or offline result.

IRIS.TV, acquired for $10 million in November 2024, added a different layer. Its content identifier lets streaming publishers attach standardized information to individual video files and scenes. Instead of buying broadly against “news” or an app, an advertiser can use contextual, emotional, and brand-suitability signals closer to the actual program. In 2026, Viant paid $40 million for TVision, whose panels measure person-level, eyes-on-screen attention across linear and connected TV. Viant began testing those attention signals before the bid - not merely as a report after the money was gone.

“An impression tells you the television played the ad. Attention asks whether a human joined the meeting.”The practical leap in Viant's TVision bet

The Myspace inheritance

This neat signal stack arrived by an untidy route. Tim, Chris, and Russ Vanderhook founded the business in 1999 as AdvertisementBanners.com, later known as Specific Media. In 2011, Specific Media and partners including Justin Timberlake bought Myspace from News Corp. for a reported $35 million. News Corp. had paid $580 million six years earlier. The bargain price was not magic. A redesign and entertainment push did not return Myspace to the center of online life.

What failed first? The consumer comeback. Myspace remained famous but stopped being the strategic center. The liabilities were less nostalgic: an old-platform data breach disclosed in 2016, a long-running Federal Trade Commission consent order, and a 2019 server migration that lost some users' older media. Viant still owns the subsidiary and still lists it as a risk. The lesson is awkward but useful: a cheap cultural artifact can remain expensive in attention, compliance, and maintenance.

No executive announced a cinematic change of heart. The shift is visible in what Viant kept funding. It co-founded ad-supported streamer Xumo, sold to Comcast in 2020. It acquired self-service DSP Adelphic in 2017. The founders bought back Time Inc.'s controlling stake from Meredith in 2019. Then Viant went public in 2021, selling 10 million shares at $25 and raising $250 million before underwriting discounts. The durable business was not running a social network. It was helping advertisers identify, buy, and measure audiences as media splintered across screens.

How the money moves

Viant's customers are advertising holding companies, independent agencies, mid-market media shops, and marketers buying directly. Public case studies span hotels, restaurants, universities, casinos, pharmacies, nonprofits, and health care. Some customers use Viant as their primary DSP. Others keep multiple platforms and move budgets campaign by campaign, which is easy because contracts generally offer limited commitments and campaigns can often be reduced or canceled.

Viant earns platform and service fees in two principal ways: a percentage of advertiser spend or a fixed cost per thousand impressions. It also charges for managed execution and related services. That model scales when self-service software and automation let more spending pass through without an equal rise in headcount. It also creates a blunt vulnerability: buyers can turn the tap down quickly.

The recent numbers show why management is leaning into television. Viant reported $344.2 million in 2025 revenue, up 19 percent. In the second quarter of 2026, revenue rose 34 percent to $104.3 million and CTV advertiser spend grew nearly 50 percent, passing half of all spend on the platform. More than 80 percent of that CTV spend moved through Direct Access, Viant's program for connecting buyers closer to publishers and cutting non-value-adding resellers out of the supply path.

Where Viant wins - and where it does not

The alternatives are formidable: The Trade Desk, Google, Amazon, Yahoo, StackAdapt, Basis, Simpli.fi, and other DSPs. Google and Amazon bring enormous first-party ecosystems. The Trade Desk brings greater scale and international reach. Viant's counter is focus. It is independent, buy-side only, built heavily around CTV, and willing to bundle identity resolution, content context, reporting, and hands-on support in the same relationship. Its 2025 customer survey put satisfaction above 90 percent.

That does not make it universal. Viant cannot buy inventory locked inside social platforms or every walled garden. Its household model works best where address signals are reliable and permissioned data is available. Attention panels remain modeled samples, not omniscience. AI optimization is only as sensible as the objective and conversion data supplied to it. A poorly defined “outcome” can teach a system to chase cheap proxies with great efficiency.

There is also regulatory gravity. Household identity lives close to privacy law, consumer opt-outs, data-broker rules, and changing platform policies. Direct supply connections help only when the publishers a buyer needs participate. Smaller advertisers may not have enough conversion volume to train useful optimization or justify onboarding. Global buyers may prefer platforms with broader reach. And when a brand needs Amazon shopping behavior, Google search intent, or closed social inventory, independence becomes a limitation rather than a virtue.

A company built to keep changing

Viant had about 380 employees across 10 North American offices at the end of 2025. Its annual report calls the culture “Founder Mode” and gives it a less polished core value: “Figure It Out.” That sounds like a slogan until you scan the corporate archaeology - banner network, social network, publisher ownership, management buyback, IPO, CTV, AI. Reinvention is not a workshop exercise here. It is operating history.

The next test is whether Viant can turn the three-signal story into better economics for buyers. Context should reduce unsuitable placements. Attention should distinguish a valuable impression from background noise. Household identity should connect fragmented screens without leaning on a browser relic. Outcomes should coordinate the work. If the stack produces lower acquisition costs or higher sales lift, Viant has something sturdier than another AI label. If it merely creates more dashboards and modeled metrics, advertisers will keep multiple DSP tabs open and move the budget.

For now, the company occupies a useful middle ground: smaller than the giants, older than most challengers, independent of the media it buys, and unusually concentrated on the living-room screen. Viant's peculiar journey offers one clean conclusion. The valuable part of advertising technology is not owning yesterday's audience. It is making tomorrow's decision with better evidence.