SoundHound AI paid roughly $43 million in stock for LivePerson. In return it got a customer roster that includes 25 of the Fortune 100, a digital messaging network carrying about a billion consumer conversations a month, and a business whose revenue is shrinking by roughly a fifth every year.
The acquisition closed on September 4, 2026, four and a half months after the two companies announced it on April 21. LivePerson stockholders approved the merger on September 2 with roughly 97% support — a number that says less about enthusiasm than about the alternative. LivePerson common stock stopped trading on Nasdaq the day the deal closed, ending a run as a public company that began in 2000.
"This merger represents a defining moment for the new agentic AI era. Together, we are delivering the most complete AI platform to the most comprehensive enterprise customer base in the industry," said Keyvan Mohajer, CEO and co-founder of SoundHound AI, in the closing announcement. "Now global brands have a single, unified engine to power intelligent customer interactions."
What SoundHound actually paid
The headline equity number is small: $43 million, about a 22% premium to LivePerson's 30-day volume-weighted average price at announcement. Almost none of the value here flowed to common stockholders, because by April there was very little equity value left to flow.
The more meaningful figure is the implied enterprise value of $250 million, which SoundHound arrived at after applying significant discounts to LivePerson's remaining debt. SoundHound also expected to receive $74 million of LivePerson's cash at closing, before repaying the 2026 convertible senior notes. At close it retired LivePerson's outstanding debt with a mix of cash and equity at its own discretion, leaving what the company calls a debt-free combined balance sheet.
That structure is the deal. SoundHound did not really buy a company from shareholders; it negotiated with creditors and cleaned up a capital structure.
Two companies moving in opposite directions
SoundHound arrived at the closing table with momentum. Second-quarter 2026 revenue hit a record $61.9 million, up 45% year over year, with non-GAAP gross margin of 58.4% and an adjusted EBITDA loss narrowed to $9.6 million from $14.3 million. GAAP net loss was still $42.8 million. The company held $203 million in cash at June 30 with no debt, and raised full-year 2026 guidance to $230–$260 million, promising an update for LivePerson before year-end.
LivePerson arrived from the opposite direction. Full-year 2025 revenue was $243.7 million, down 22%, and 2026 guidance called for $195–$207 million — another 15% to 20% decline. First-quarter revenue was $57.0 million against $64.7 million a year earlier, and the company carried a stockholders' deficit of about $51.5 million.
The debt was the real constraint. LivePerson issued $517.5 million of convertible senior notes due 2026 in December 2020, then spent years restructuring them — exchanging $341.1 million of principal for a package of cash, 10.0% senior subordinated secured notes due 2029, preferred stock and common stock, cutting debt by roughly $226 million along the way. By June 30, 2026, only $15.0 million of the original notes remained. The company survived its maturity wall; it did not survive the revenue decline that came with it.
Distribution, not model quality
The strategic logic here cuts against how most people talk about AI competition. SoundHound did not buy LivePerson for its technology. It bought incumbency.
Contact-center AI is not won by having the best model. It is won by being already installed — wired into the CRM, the ticketing system, the telephony stack, the vendor security questionnaire that took eleven months to clear. LivePerson's Conversational Cloud sits inside hundreds of enterprises, many of those relationships spanning more than a decade. That is the asset. The combined company now works with customers across 30-plus countries, including 12 of the top 15 global banks, four of the top five global airlines, four of the top five global automakers, and more than ten leading telecom providers.
"The artificial boundaries between talking and typing are disappearing," said John Sabino, CEO of LivePerson, when the deal was announced. "Consumers expect to start a complex request over the phone and finish it seamlessly via text or web messaging, without ever repeating themselves or losing context."
He is describing a real product gap, and also why a struggling messaging vendor was worth buying: SoundHound had voice and no digital footprint at that scale, LivePerson had the footprint and was losing the technology race. LivePerson's platform is being folded into OASYS, SoundHound's orchestration layer, and the combined patent portfolio now exceeds 750, up from the 400-plus SoundHound cited in August.
The risk of buying something that is shrinking
None of this makes the acquisition safe. SoundHound projects at least $350–$400 million in 2027 revenue, including a minimum $100 million contribution from LivePerson customers, and a $500 million opportunity from the combined base alone. Those numbers require LivePerson's churn to stop. Acquisitions rarely stop churn; they often accelerate it, because renewal season is exactly when a customer reconsiders a vendor that just changed owners.
There is also an integration burden. LivePerson is SoundHound's fifth acquisition, following Amelia and Interactions among others, and SoundHound has not yet shown it can turn accumulated acquisitions into profitable growth — the $42.8 million quarterly GAAP loss is the evidence. Buying distribution is cheap. Retaining it is not.
One telling detail: SoundHound's new CFO, John Collins, comes from LivePerson, where he served as CFO, COO and interim CEO and led the debt restructurings that captured $227 million of debt discount. SoundHound hired the person who knows exactly how thin the acquired business is.
What to watch
Three things. First, the updated guidance SoundHound has promised before year-end — that is when the market learns how much LivePerson revenue management actually believes it will keep. Second, LivePerson's renewal cohort over the next two quarters, which will show up as either stabilization or a steeper decline. Third, dilution: SoundHound has already filed a prospectus supplement covering resale of Class A shares issued to former LivePerson noteholders, and a stock-funded debt retirement means new supply hitting the market.
SoundHound bought a seat at 25 of the largest companies in America. Whether it keeps that seat is a 2027 question.
“The artificial boundaries between talking and typing are disappearing. Consumers expect to start a complex request over the phone and finish it seamlessly via text or web messaging, without ever repeating themselves or losing context.”— John Sabino, CEO, LivePerson