From $4.7 Billion to $43 Million: SoundHound Closes Its LivePerson Acquisition Today

One of the internet’s oldest customer-service brands quietly disappears from Nasdaq today. LivePerson — the company that pioneered live web chat in the late 1990s and rode the conversational-commerce wave to a peak market value of roughly $4.7 billion — is closing its acquisition by SoundHound AI on September 4, 2026, ending its run as an independent public company. The equity price tag: approximately $43 million, with a total enterprise value around $250 million once LivePerson’s convertible debt is factored in.

It is a sobering full-circle moment for the AI industry. LivePerson wasn’t a company that missed the technology shift — it was one of the earliest to brand itself as a conversational AI company. It simply couldn’t outrun the economics of a new generation of AI-native competitors, and its shareholders ultimately approved a sale that values the business at a fraction of one percent of its former worth.

How the deal is structured

SoundHound first announced the acquisition on April 21, 2026. The consideration is predominantly stock: most LivePerson (LPSN) shareholders will receive SoundHound (SOUN) common shares under a formula-driven exchange ratio. Because the final ratio depends on SoundHound’s VWAP near closing, the agreement includes a collar — filings reference a $7–$12 VWAP band — to bound the dilution on either side.

The July 2026 amendment reshaped the deal further. Cash payouts for LivePerson shares traded on the Tel Aviv Stock Exchange were capped at $7.5 million, and a $5 million termination fee was added, an acknowledgment of how much both sides wanted certainty as the closing date approached. LivePerson stockholders approved the transaction on September 2, and the companies targeted September 4 for the close, subject to customary conditions.

At first glance, a $43 million equity value for a company that still powers roughly one billion customer messages per month looks almost absurd. The enterprise value of ~$250 million reflects LivePerson’s 2026 Convertible Senior Notes — and that is where the deal’s financial engineering gets interesting. At closing, SoundHound expects to receive approximately $74 million of LivePerson’s cash balance before repayment of those notes, and the combined company is expected to emerge debt-free.

A cautionary tale 26 years in the making

LivePerson’s history is effectively a history of digital customer service. Founded in 1995 by Robert LoCascio and headquartered in New York, the company pioneered web chat technology around 1997–1998, went public in 2000 at $8 per share, and spent two decades layering on proactive chat, AI-driven routing, and eventually a full conversational cloud used by major brands.

The decline was as dramatic as the rise. Revenue peaked around $514.8 million before collapsing by more than half, and filings around the merger show LivePerson still shrinking, with revenue down roughly 22% year over year at the time of the vote. By early 2026 the market cap had fallen to the tens of millions. LoCascio himself departed years earlier, and successive pivots — from chat, to messaging, to “AI-powered conversational cloud” — never restored growth. Investors who bought the 2021 peak lost essentially everything: a 99% collapse, as Calcalist bluntly put it.

The uncomfortable lesson is that rebranding around AI is not the same as having AI economics. LivePerson’s cost structure was built for an era of human agents supplemented by automation. When large language models arrived, enterprises discovered they could get credible conversational automation from a new wave of vendors without the legacy platform premium.

What SoundHound gets for its stock

For SoundHound, the acquisition is the third leg of an aggressive consolidation strategy. The company acquired enterprise AI firm Amelia in 2024, completed its purchase of voice-AI competitor Interactions in September 2025, and now adds LivePerson’s digital messaging footprint. The strategic logic is omnichannel completeness: SoundHound’s strength is voice — restaurant ordering, automotive assistants, call-center automation — while LivePerson brings text-based channels: web chat, SMS, in-app messaging, and the bot orchestration layer that connects them.

The numbers SoundHound is guiding to are ambitious: combined 2027 revenue of at least $350–$400 million, with at least $100 million of “growable” contribution from LivePerson’s long-tail customer base, and a path toward $500 million based on the existing combined pipeline. Those targets make SoundHound one of the few pure-play conversational AI companies attempting to scale to half a billion dollars in revenue — territory where it will compete directly with the agentic offerings of much larger platforms.

Not everyone is convinced. SoundHound’s stock is down roughly 29% over the past six months, and analysts — Wedbush among the bluntest — have questioned the wisdom of using equity to absorb a shrinking asset. The counterargument is the balance-sheet arithmetic: $74 million of incoming cash, a debt-free exit, and a customer list of enterprise brands that would cost far more to acquire through sales and marketing.

Consolidation season in conversational AI

The LivePerson close lands amid a broader wave of conversational-AI consolidation. As voice agents and chat agents converge on the same enterprise budgets — and as buyers increasingly want one platform that handles a customer whether they call, type, or tap — mid-sized specialists are being absorbed at a rapid clip. SoundHound has been the most serial acquirer, but the same forces are driving deals across the sector, from neoclouds buying inference capacity to model labs buying distribution.

The symbolic weight is hard to miss. The company that put live chat on e-commerce sites in the dot-com era is being folded into a voice-AI firm that didn’t exist when LivePerson went public. Whatever SoundHound makes of the asset, today’s close is the industry’s starkest reminder yet that in the LLM era, being early is not the same as being positioned — and that a once-$4.7-billion franchise can be had, debt and all, for less than the price of a single AI training cluster.