The Redemption Near-Miss: SoundHound AI's $2.1B deSPAC With Archimedes Tech

Every figure below comes from Archimedes Tech SPAC Partners Co.'s 8-K disclosing the business combination closing (April 27, 2022), the original deal announcement (November 16, 2021), and SoundHound AI, Inc.'s S-1/A financial statement tables. Full sources on SEC EDGAR.

Deal snapshot

Why a deSPAC is structurally different from an IPO

Every deal in this series so far has been either a straight M&A transaction or a traditional IPO. A deSPAC is neither, it's a reverse merger into an already-public shell, and the mechanics that matter are almost entirely different from a normal IPO roadshow-and-price-discovery process.

A SPAC raises cash in its own IPO, parks it in a trust account earning interest, and has a limited window (typically 18-24 months) to find and merge with an operating company or return the cash to its shareholders. When a merger target is identified, the SPAC's own public shareholders get a choice at the shareholder vote: stay in as shareholders of the newly combined company, or redeem their shares for a pro-rata slice of the trust account (roughly $10.00 per share, since that's what the trust was funded at) and walk away in cash before the merger even closes.

That redemption right is the single most important structural difference from an IPO, and it's what makes SoundHound's deal worth studying: the "up to $244 million" headline figure at announcement was never a guaranteed number. It was a ceiling, contingent on how many Archimedes shareholders chose to stay versus cash out, and the market gave a very clear answer.

What actually happened to the money

At announcement, the deal was structured around up to $133 million sitting in Archimedes' trust account, plus a committed $111 million PIPE (private investment in public equity, priced at $10.00/share, sold to bring in cash on top of whatever the trust delivered).

By the time the deal actually closed in April 2022, the redemption math had turned sharply against the SPAC side: an aggregate of $127,679,500 was paid from the trust account to shareholders who exercised their redemption rights, leaving remaining trust proceeds of only approximately $5,356,628, roughly 4% of the original trust survived the vote. That's a redemption rate north of 95%, consistent with the broader wave of SPAC redemptions that hit nearly every 2021-vintage deSPAC once markets turned in early 2022 and the arbitrage math (redeem at ~$10 versus hold shares trading near or below $10 in a falling market) made cashing out the obvious move for most SPAC-IPO investors, who were never long-term believers in the target company to begin with.

That's also exactly why the PIPE mattered so much, and why Archimedes added to it late. On April 9, 2022, roughly two weeks before the vote, Archimedes added a further $2 million to the PIPE, bringing the deal's total PIPE proceeds to $113 million, with the original $111 million already covering the deal's $111 million minimum cash condition, and the additional investment providing extra cushion in a high-redemption environment. That single detail is the whole story in miniature: the SPAC sponsors could see redemptions running hot into the vote and topped up the one funding source they still had full control over, specifically to protect the deal's minimum-cash closing condition from a trust account that was clearly about to be gutted.

The PIPE ultimately delivered its full $113,000,000, sold as 11,300,000 shares of Class A common stock at $10.00 per share to subscribers under agreements entered into before the vote, meaning that of the roughly $118 million total gross proceeds SoundHound actually received at closing, the overwhelming majority came from the PIPE, not the SPAC's own trust. The trust, which was the entire original premise of the shell vehicle, ended up contributing a rounding error.

The strategic investor roster is doing PR work the trust couldn't

One detail worth flagging for anyone building comps on SPAC PIPEs: SoundHound's PIPE wasn't just financial investors backfilling a redemption gap. The PIPE was anchored by Oracle, Qatar First Bank, Koch Industries, and MKaNN Ventures, alongside Cota Capital, VIZIO, HTC, FIH Mobile (a Foxconn Technology Group company), Structural Capital, Provco Group, Sompo, and Pejman Nozad, and the deal materials explicitly framed this mix of strategic and financial investors as validation of the underlying technology.

That framing is worth reading skeptically rather than at face value, a heavily strategic-investor PIPE is genuinely useful for a story-driven pre-revenue-scale company that needs commercial credibility, but it's also a common way to paper over a redemption-battered trust with logos that look better in a press release than "95% of our SPAC investors took the cash and left."

The post-close reality check

Once trading began, the market's read on SoundHound was immediately more cautious than the November 2021 announcement framing. Coverage around the close pegged the completed combination at an estimated $2.0 billion valuation, noted management's expectation of a $30 million revenue midpoint for 2022, and concluded the stock was priced for perfection with a hold rating pending clearer evidence of revenue growth and improved operating results, a valuation resting on a company still generating well under 2% of that valuation in annual revenue and running significant operating losses, which is the standard bear case leveled at nearly every 2021-vintage AI/voice-tech deSPAC once the sector's multiples came back down to earth.

AI/deSPAC-specific deal checklist this case illustrates

"Up to" proceeds figures in SPAC deal announcements are ceilings, not commitments, always check the actual closing 8-K for the real redemption and PIPE numbers, which can differ enormously from the headline announcement figure. Redemption rates are the single most important number in any deSPAC closing filing, a trust account can be reduced to a small fraction of its original size and the deal can still close, provided the PIPE and minimum-cash conditions are structured to survive that. Late-stage PIPE top-ups in the weeks before a shareholder vote are a redemption-risk signal, sponsors add PIPE capacity specifically when they can see the trust is about to be gutted. And a strategic-investor-heavy PIPE roster is partly a financing decision and partly a PR decision, it's worth separating "this validates the technology" framing from the more mechanical reality that the SPAC's own trust largely failed to deliver on its original premise.