Let's not be precious about this.

Integral Ad Science has one product, and it isn't software. It's the promise that somebody in this business isn't on the take. Brands pay IAS because they don't trust the platforms, don't trust the exchanges and, if they've been paying attention, don't entirely trust their own agencies. So they hire a referee. IAS is the referee. Its own SEC filings list being an "independent third party" as a core competitive advantage. That's the whole pitch. Everything else is a dashboard.

Now here's the part nobody on a stage at Advertising Week wanted to say out loud.

The referee is in hock.

On December 23, 2025, Novacap, a Montreal private equity firm, closed its purchase of IAS for about $1.9 billion, or $10.30 a share. That's 43% below the $18 IPO price IAS set in 2021, which means a lot of people who bought the independence story at the top got cashed out for considerably less of it. To make the deal work, Novacap lined up commitments for a $1 billion term loan and a $150 million credit line from RBC, Bank of America and Barclays.

Then the banks went to sell the loan to the ordinary leveraged-loan market. These are not sentimental people. Pricing risk is what they do for a living. They passed.

Bloomberg reported on November 17, 2025 that the syndicated market didn't bite, and that concerns about AI's impact on digital advertising were part of the reason. RBC took the deal to private credit, where money costs more and the lenders don't have to tell you their names.

Sit with that for a second. IAS markets itself as the company that keeps advertising safe in the age of AI. The professional money looked at the exact same AI story and decided it was the risk, not the moat.

So here's where we are. The company your brand pays to watch the ad market now owes a billion dollars to people who thought it was a shaky bet. Novacap isn't paying that loan back out of its own pocket. Private equity doesn't work that way. The loan gets paid out of IAS's cash. IAS's cash comes from you.

That's bad for IAS. It's worse for the brands writing the checks.

Before Novacap, IAS was for sale for about a year. The merger filing reads like the guest list for a party where the host kept checking the window.

The bankers started making calls in September 2024. A handful of strategic buyers, the companies that might actually want a verification business and could pay for it with something other than borrowed money, signed confidentiality agreements and looked at the books. Not one of them made a formal bid.

The best offer IAS ever got was $12.00 a share, from a bidder the filing calls Party B, in October 2024. Party B did the diligence. Then Party B walked, citing "slowing growth, competition, pricing headwinds and low cash conversion." That's not a short seller on X. That's a buyer who read the numbers and decided it didn't want to own them.

Another bidder, Party R, floated $10.50 to $11.50 in March 2025. It walked too.

That left Novacap, which opened at $9.30, moved to $9.80 and landed at $10.30. In August 2025, IAS's board told Novacap it would need to line up its financing before it could pay more.

Read that one again. The seller was worried the buyer couldn't afford the purchase. That's not a bidding war. That's a landlord asking the only applicant for proof of income.

Shareholders never voted. Vista Equity, the firm that owned IAS before its IPO, and Atlas Venture together held 52.6% of the vote and approved the deal by written consent. Everyone else got an information statement. Not a ballot. A notice.

On paper, the financing looked like this: $1 billion in term debt, $150 million in revolving credit, and up to $857.8 million of Novacap equity. RBC ended up as administrative agent when the deal closed.

In the market, it looked like this. PitchBook reported the private-credit pitch at 5 points over SOFR, issued at 98.5. In plain English, the lenders wanted a fat interest rate and a discount on day one. That's what money costs when the people lending it are nervous.

Who actually holds the loan today isn't public. The private-credit funds that took the paper haven't been named anywhere we could find. The final pricing, maturity and covenants weren't filed. IAS's most important financial relationship is with lenders nobody outside the deal can identify.

For a company whose product is transparency, that's a hell of a thing.

On August 6, 2026, Nielsen agreed to buy DoubleVerify for roughly $2.15 billion in enterprise value, $13.60 a share in cash, with closing expected by the first quarter of 2027.

DoubleVerify was the counterargument. DV was the proof that a verification company could stand on its own as a public company, answer to shareholders and keep everyone honest. That proof just got bought by a measurement conglomerate that is itself owned by private equity. Nielsen's announcement promised to "preserve independent verification standards," which is the kind of sentence you only write once independence is no longer the default.

Two years ago, the ad business had two big independent verification companies. By early 2027, it may have none. It will have a Nielsen-owned DoubleVerify and a debt-financed IAS owned by a fund that, by definition, intends to sell it.

If you're a brand, your choice of referee just went from two to one and a half.

Keep reading free with your email: why the debt is bad for IAS, the cracks the bidders already saw, and why the company went dark the day it went private.