SoundHound AI (SOUN) ☆

AI applications & software · Industry Index 7 (rank #175) · $4.36 (-1.80%) · mkt cap $1.9B

ⓘ Data freshness

price: as of 2026-10-29 index: as of 2026-10-29 fundamentals: latest filing 2026-08-10 Reports quarterly (10-Q XBRL)

SoundHound AI develops independent voice and conversational-AI technology for automotive, restaurant, and customer-service applications.

Industry exposure 0.9 (A pure-play voice-AI company whose entire business is conversational AI across automotive, restaurant, and customer-service markets.)

Signal

signal: as of 2026-10-13

Short (~2 weeks)

42%

Bearish

Conviction 42

Options positioning— Delayed per-contract options — call premium 1% / put 99% (net skew -0.99), unusualness 28/100. Context: Consumer Price Index in 1 day.[options]
Price momentum— Technical downtrend across short+medium horizon(s) — vote -0.81, confidence 0.56 from 11 indicator sub-signal(s).[price] · derived internally
Structural— Industry Index 7.0 (falling -11.0 over window) — structural headwind.[index] · derived internally

Medium (~3 months)

Low-confidence · Neutral

Conviction 0

Growth trend— Revenue 45% YoY (decelerating), net margin expanding.[filing]
Price momentum— Technical downtrend across short+medium horizon(s) — vote -0.81, confidence 0.56 from 11 indicator sub-signal(s).[price] · derived internally
Structural— Industry Index 7.0 (falling -11.0 over window) — structural headwind.[index] · derived internally

Long (~6 months)

Low-confidence · Neutral

Conviction 0

Growth trend— Revenue 45% YoY (decelerating), net margin expanding.[filing]
moat— Editorial prior (subjective, NOT backtested): narrow moat / behind / weak pricing ⇒ narrowing/lagging long-horizon tilt -0.15. Bounded, low-confidence input.[moat: Editorial moat profile — soundhound — narrow moat, behind, weak pricing (as of 2026-03-02)]
Structural— Industry Index 7.0 (falling -11.0 over window) — structural headwind.[index] · derived internally

Show Details

Relationship network

Loading relationship graph…

NVIDIA — partner

Key relationships

NVIDIA — partner · SoundHound AI integrated NVIDIA AI Enterprise (NIM and NeMo microservices) to accelerate its voice-AI stack. source: sec.gov

Options positioning

delayed · aggregate: as of 2026-10-29

Call vs put volume: 9,997 calls · 1,999 puts — tilt +66.7% toward calls
Vol/OI turnover: 0.08×
Premium ($ notional proxy): $161920 calls · $133111 puts — tilt +9.8% toward calls
OTM tilt: +78.9% toward OTM calls
IV skew (put − call): +49.9%

Moat

moat: latest change 2026-03-02

SoundHound sells independent, white-label voice and agentic AI to carmakers, TV/IoT makers, restaurants and enterprises. Its 10-K pitches this as the alternative to big-tech assistants that take over a brand's users and data. Its defensibility rests mainly on automotive design wins, which the 10-K says customers are very unlikely to change until a product model is revamped. It layered on the Amelia (Aug 2024) and Interactions (Sep 2025) acquisitions, and it completed the LivePerson acquisition on 2026-09-04 per that day's 8-K, whose press release cites a strengthened IP portfolio of over 750 patents. Against that, the 10-K lists rivals with far greater resources and data that may be able to offer competing functionality at lower cost or free, customers that may develop competing products, competitors that may build their own solutions on open-source software, and possible OEM pricing pressure. The 10-K reports substantial net losses and negative operating cash flows to date, and GAAP gross margin fell from 75.4% (FY2023) to 42.4% (FY2025), a decline the company attributes mainly to acquisition mix and amortization. The advantage looks narrow and not yet proven economically.

Rating: narrow

The FY2025 10-K describes a real but limited edge: after an automotive design win a competing product may not be able to displace the winner because it is 'very unlikely that a customer will change complex technology until a product model is revamped', and it claims technical barriers to entry and cites over 359 granted and over 102 pending patents. The same filing concedes intense, price-sensitive competition from larger rivals that may be able to offer competing functionality 'at lower cost or free of charge' within larger offerings, possible pricing pressure from OEMs, master service agreements customers can terminate on six months to one year of prior written notice, and, to date, substantial net losses and negative cash flows from operating activities. The stickiness is real, but it has not yet produced profits.

Type: switching costs

The stickiest advantage in the 10-K is design-win lock-in: a product that did not win an OEM or Tier-1 design may not be able to displace the winner until the customer begins a new selection process, because the 10-K says customers are very unlikely to change complex technology until a product model is revamped. Private domains make SoundHound a long-term "partner" (the 10-K's quotation marks) because subscribing customers retain the ability to iterate and update their content on its platform. Patents are the weaker source: the filing warns that competitors 'may independently develop technologies that are substantially the same or superior' without infringing its rights.

Leadership: behind

The 10-K calls SoundHound 'a global leader in conversational intelligence' with 'best-in-class speed and accuracy', but these are the company's own claims and it gives no share figure. Apart from the company's own positioning, such as its belief that some legacy vendors offer dated technologies, the 10-K's comparisons with rivals run against SoundHound: some current or potential competitors are large technology companies that have 'significantly greater financial, technical and marketing resources than SoundHound does', they may be able to respond more rapidly to new or emerging technologies, and its large competitors may have greater access to customer data, which the 10-K says provides them with a competitive advantage in developing new products and technologies.

Pricing power: weak

The 10-K flags price sensitivity among customers, rival alternatives offered 'at significantly lower costs or free of charge', and the strong purchasing power of major OEMs, under which SoundHound may be expected to quote fixed prices or be forced to accept annual price-reduction commitments. GAAP gross margin, computed from the 10-K's revenue and cost of revenue, fell from 75.4% (FY2023) to 48.9% (FY2024) and 42.4% (FY2025). The company attributes both declines primarily to acquisitions: the FY2024 10-K cites Amelia and SYNQ3, including SYNQ3's lower-margin call-center agent business and amortization of acquired intangibles, and the FY2025 10-K cites Amelia and Interactions, including $14.1M and $0.8M of amortization of acquired intangibles. Neither passage attributes the decline to pricing. The Q2 2026 release (8-K filed 2026-08-05) reports GAAP gross margin up 6.1 points year over year to 45.1%, with non-GAAP gross margin flat at 58.4%.

Curated moat judgment grounded in the cited source. source: sec.gov

Compare with…

Value-chain position

Layer: AI applications & software

Upstream (suppliers): —
Downstream (customers): —
Competitors: —

Voice and agentic AI layer that OEMs, restaurants and enterprises embed under their own brand. Per the 10-K it combines its domains with third-party LLMs such as OpenAI's ChatGPT and runs on hosting facilities and third-party public clouds.

Derived from this company's mapped supply-chain edges and its industry layer.

Signal contribution editorial prior

The moat prior voted -0.15 on the long (~6-month) horizon at weight 0.20 — the swarm call there is neutral (score 0.12, conviction 0).

Editorial prior — subjective, not backtested. It is bounded (±0.6), low-confidence, and never votes on short/medium. See exactly how it voted →

Product lineup

Products & services with an editorial market-share band and technical barrier band — qualitative judgments tied to a cited basis, never invented figures.

Product	Market share	Technical barrier
Houndify Voice AI platform (automotive, TV, IoT) · Embedded and cloud voice assistants White-label speech recognition, Speech-to-Meaning NLU, wake words, TTS and 100+ content domains, licensed to product makers on a royalty basis across edge, cloud and hybrid deployments.	Unknown The 10-K sorts most Voice AI participants into 'big tech' companies and 'legacy vendors' and positions SoundHound as the independent alternative to both. It discloses no market-share figure, and it concedes some rivals have significantly greater resources.	Moderate The 10-K says that, in its view, it 'takes many years and significant investment to build all the components of Voice AI', and that after a design win a competing product may not be able to displace the winner until the customer begins a new selection process, because a customer is very unlikely to change complex technology until a product model is revamped. It also warns that large competitors may be able to offer competing functionality at lower cost or free of charge within larger offerings, and that customers may decide to develop competing products.
OASYS enterprise agentic AI platform · Enterprise customer-service AI agents SoundHound's agentic platform. The Q2 2026 release says: 'Its agentic platform, OASYS, is a self-learning, orchestrated AI system where organizations can build and deploy conversational AI agents to handle transactions, tasks, and workflows on behalf of customers and employees.' The FY2025 10-K, which does not name OASYS, separately says that after integrating the acquired Amelia platform in 2025 it enhanced its AI agent platform with systems that can reason through IT-support or customer-service requests.	Unknown The Q2 2026 release credits 'major enterprise AI deals attributed to OASYS' and lists enterprise renewals in banking, insurance and telecom, including Rakuten Securities. The leadership language in SoundHound's releases is the company's own, and no share figure is given.	Low The barrier call rests on the FY2025 10-K (filed 2026-03-02), which does not name OASYS: it says it is possible for competitors to develop their own solutions using open source software, that customers may decide to develop competing products, and that large rivals may be able to offer competing functionality at lower cost or free of charge within larger offerings. It also says the company believes generative AI and large language models let better quality models be created faster and with fewer tedious engineering requirements, which it presents as an opportunity for fast movers.
Smart Ordering and Smart Answering (restaurant voice AI) · Restaurant phone, drive-thru and kiosk ordering Subscription voice agents that answer restaurant calls and take phone, kiosk and drive-thru orders, learning each menu and integrating with POS systems.	Unknown The 10-K gives no market-share figure for restaurant voice AI. The Q2 2026 release (8-K filed 2026-08-05) lists new deals (Ruby Tuesday for Smart Answering and Smart Ordering, a large seafood QSR for drive-thru ordering), renewals (Habit Burger, Red Lobster, Torchy's Tacos) and a prominent pizza brand with SoundHound technology live in more than 75% of its total locations, also without a share figure.	Low The 10-K describes Smart Ordering as learning the menu of each business and seamlessly integrating with multiple POS systems, with a Gateway for custom POSs, but no passage describes a switching cost. Its risk factors describe a market marked by price sensitivity on the part of customers and alternatives from other vendors which may be offered at significantly lower costs or free of charge.
SoundHound Chat AI (generative voice assistant) · Generative-AI voice assistant A voice assistant that combines dozens of real-time knowledge domains with third-party large language models such as OpenAI's ChatGPT (10-K).	Unknown The 10-K gives no share, unit or customer figure for Chat AI, and the Q2 2026 release (8-K filed 2026-08-05) does not mention it.	Low The 10-K says Chat AI combines its knowledge domains with third-party large language models like OpenAI's ChatGPT, so the model layer is not proprietary, and the company says it believes generative AI and large language models let better quality models be created faster and with fewer tedious engineering requirements. No passage in the 10-K ties Chat AI to a design win.

What you're looking at

These charts plot SoundHound AI's reported income statement, cash flow, and balance sheet from its SEC filings through 2026-08-10. Every figure is as-filed XBRL — none of it is an estimate.

Over the reported window, net income at a loss of $14006 in FY2025.

Analyst consensus — third-party estimates from Yahoo Finance, as of 2026-10-26. Not SEC filings and not a TradingPilot forecast. source: finance.yahoo.com

Period	EPS (est.)	Revenue (est.)	YoY

Each figure shows the consensus mean, its (low–high) range where available, and · the number of analysts behind that estimate.

Next report ~2026-11-05.

Our model leans Low-confidence · Neutral on the stock over the medium horizon — a market-direction call, not an earnings forecast. See the Signal tab.

Income statement

Cash flow & liquidity

Balance sheet

Filings

2025Q2 · filed 2026-08-10 source: sec.gov
2025Q4 · filed 2026-08-10 source: sec.gov
2026Q2 · filed 2026-08-10 source: sec.gov
2025Q3 · filed 2026-08-10 source: sec.gov
2025Q1 · filed 2026-05-11 source: sec.gov

Show 47 older filings

2026Q1 · filed 2026-05-11 source: sec.gov
FY2022 · filed 2026-04-09 source: sec.gov
FY2023 · filed 2026-04-09 source: sec.gov
FY2024 · filed 2026-04-09 source: sec.gov
FY2025 · filed 2026-04-09 source: sec.gov
FY2023 · filed 2026-03-02 source: sec.gov
FY2024 · filed 2026-03-02 source: sec.gov
FY2025 · filed 2026-03-02 source: sec.gov
2024Q3 · filed 2025-11-10 source: sec.gov
2024Q4 · filed 2025-11-10 source: sec.gov
2025Q3 · filed 2025-11-10 source: sec.gov
2024Q2 · filed 2025-08-11 source: sec.gov
2025Q2 · filed 2025-08-11 source: sec.gov
2024Q1 · filed 2025-05-12 source: sec.gov
2025Q1 · filed 2025-05-12 source: sec.gov
FY2022 · filed 2025-03-11 source: sec.gov
FY2023 · filed 2025-03-11 source: sec.gov
2023Q3 · filed 2024-11-12 source: sec.gov
2023Q4 · filed 2024-11-12 source: sec.gov
2024Q3 · filed 2024-11-12 source: sec.gov
2023Q2 · filed 2024-08-09 source: sec.gov
2024Q2 · filed 2024-08-09 source: sec.gov
2023Q1 · filed 2024-05-10 source: sec.gov
2024Q1 · filed 2024-05-10 source: sec.gov
FY2021 · filed 2024-03-01 source: sec.gov
FY2022 · filed 2024-03-01 source: sec.gov
2022Q4 · filed 2023-11-15 source: sec.gov
2022Q3 · filed 2023-11-15 source: sec.gov
2023Q3 · filed 2023-11-15 source: sec.gov
2022Q2 · filed 2023-08-11 source: sec.gov
2023Q2 · filed 2023-08-11 source: sec.gov
2022Q1 · filed 2023-05-12 source: sec.gov
2023Q1 · filed 2023-05-12 source: sec.gov
FY2021 · filed 2023-03-28 source: sec.gov
2021Q3 · filed 2022-11-14 source: sec.gov
2021Q4 · filed 2022-11-14 source: sec.gov
2022Q3 · filed 2022-11-14 source: sec.gov
2021Q2 · filed 2022-08-15 source: sec.gov
2022Q2 · filed 2022-08-15 source: sec.gov
2021Q1 · filed 2022-05-17 source: sec.gov
2022Q1 · filed 2022-05-17 source: sec.gov
2022Q1 · filed 2022-05-17 source: sec.gov
FY2020 · filed 2022-03-09 source: sec.gov
2021Q3 · filed 2022-01-10 source: sec.gov
2020Q4 · filed 2022-01-10 source: sec.gov
2021Q2 · filed 2021-08-27 source: sec.gov
2021Q1 · filed 2021-07-27 source: sec.gov

Key ratios vs peers

Comparing against peers…

Institutional holders (13F, 2026Q2)

Institution	Shares	Value	Δ shares QoQ	% of 13F	% shares out
BlackRock Passive / index	33,984,088	$219.9M	1,575,327	0.0%	—
State Street Passive / index	10,048,444	$65.0M	963,203	0.0%	—
FMR (Fidelity) Passive / index	66,744	$431833	-26,693	0.0%	—

TradingPilot-computed columns (% of 13F, % shares out) are derived from as-filed 13F values; only name, shares and value are as-filed.

What a 13F can’t show you

Delayed 45+ days. Positions are as-of the quarter-end shown, filed up to 45 days later.
Long-only, US-listed. No shorts, cash, bonds, FX, foreign, or private holdings; listed options are reported at notional and excluded from per-position equity value.
$100M threshold. Smaller managers never file.
“% of 13F” is not “% of AUM”. A 13F is a partial book.
“Sold out” means “no longer reported” — not necessarily sold.
One CIK can aggregate multiple sub-entities. Confidential treatment can omit positions.
Passive index filers are not gurus. Their 13F changes are index flows, and they are excluded from TradingPilot’s smart-money signal.
Quant caveat. A quant filer’s 13F (e.g. Renaissance) is a long-only quarter-end snapshot and does NOT reflect its actual high-turnover, short-horizon strategy.
Guru pages and holdings never feed a signal (weightless).

Impact = company-prominence tier (High/Medium auto-assigned by graph centrality/size; Low only on curated minor events), not a magnitude forecast.

Upcoming

Date	Type	Event	Impact
2026-11-09	earnings	SoundHound AI earnings (estimated) — Date not yet announced — estimated from EDGAR filing cadence (last periodic filing + 91/365 days); confirm via IR. estimated	Medium

Past

Date	Event	Impact	Outcome
2026-08-05	SoundHound AI earnings confirmed	Medium	EPS beat · +10.1%
2026-05-07	SoundHound AI earnings confirmed	Medium	EPS miss · -7.8%

See the full calendar →