Sound Ventures is easiest to misunderstand from a distance. The celebrity names arrive first: Ashton Kutcher, the actor who became a product-minded investor, and Guy Oseary, the manager who helped turn musicians into global businesses. Up close, the Beverly Hills firm looks less like a velvet-rope curiosity and more like a compact operating shop. It invests in private technology companies, then tries to make its network useful when a founder needs to recruit an executive, sharpen a product, find a customer or explain an unfamiliar idea to the world.
That proposition is now entering a new phase. In May 2026, Sound said Oseary and general partner Effie Epstein would continue leading the firm with a tighter focus on companies capable of clear commercial scale. Kutcher would become a senior advisor while starting a separate firm for deep tech, infrastructure and energy. It was not framed as a dramatic divorce. It was a division of labor: different technologies require different capital, timelines and operating machinery.
A venture firm with a translation desk
Sound's practical job is familiar: collect money from limited partners, buy stakes in private companies and attempt to return more than it invested. Its direct users are founders, typically running early- or mid-stage technology businesses, and investors seeking exposure to that private-company growth. The portfolio crosses enterprise software, fintech, consumer products, health, education and AI. The firm says it commits early and stays hands-on. The less generic part is what “hands-on” can mean when one partner has spent decades building entertainment brands and another obsesses over how products are adopted.
Technical companies regularly suffer from a translation problem. Their builders know why the product matters, but customers, recruits and the wider market do not automatically follow. Sound's pitch is that it can help bridge that gap. Oseary's career is built around story, attention and long-term talent relationships. Epstein brings operating discipline gathered across investment banking, strategy, insurance, media and the early expansion of CLEAR. Kutcher, even in an advisory role, contributes product instincts and technical curiosity. Capital is the ticket to the table; interpretation is the service layered on top.
“Every single person in our portfolio has our personal phone numbers.”Ashton Kutcher, describing the firm's founder access in 2016
A phone number is not a business model, of course. The business model remains standard venture economics: management fees on committed capital and a share of investment profits. What the phone number signals is positioning. Sound is not trying to win by operating the largest analyst factory on Sand Hill Road. It wants founders to believe senior partners will engage directly, and that a Beverly Hills network can be an advantage when a product needs distribution, credibility or a more memorable story.
The small basket behind the big AI bet
The firm's most distinctive product is not a software tool. It is a portfolio design. In 2023, Sound announced a $240 million AI fund intended to hold roughly six or seven companies. About half had already been committed when the vehicle became public, with initial disclosed positions in OpenAI, Anthropic and Stability AI. At the time, Epstein explained that Sound expected only a small number of foundational-model companies to survive the talent requirements and punishing cost of computing infrastructure.
That cuts against the usual venture instinct to spread risk across a large batch of uncertain startups. Sound instead treated foundational AI as a market where a few platforms might capture disproportionate value. Backing rivals was acceptable because the thesis concerned the layer, not a single winner. The trade is blunt: concentration can make a correct thesis matter enormously, but it offers less shelter when the category or entry price is wrong.
By 2026, Sound described the broader AI strategy in still larger terms, saying it had deployed more than $800 million into early positions in Anthropic, OpenAI and World Labs. That figure comes from the firm and covers more than the original $240 million vehicle. It nonetheless clarifies where Sound believes its strongest result has come from: not a wide survey of every AI application, but early access to the expensive model layer beneath them.
Broad applications, narrow foundations
Sound also invests in the companies building on those models. Its 2026 activity included backing Warp, an AI-native employee-management platform automating payroll, HR, benefits and compliance, in a $60 million Series B. It led a reported $17 million round for Anomaly Insights, which helps health systems understand payer behavior. This side of the portfolio solves ordinary business pain with new machinery: administrative work, customer service, legal work, health-system payments and other expensive processes that still run on fragmented software and human follow-up.
The customer is not always the person writing code. It may be the HR manager buried in onboarding, the in-house legal team sorting routine questions or the healthcare operator trying to see why a claim was denied. For those buyers, “AI-native” matters only if the product removes steps, improves a decision or lowers a cost. Sound's brand and go-to-market experience becomes more relevant at precisely that point - when technical capability must become a credible purchase.
For a startup, that conversion involves dozens of small decisions: which customer to pursue first, what promise belongs on the home page, whether the product should hide its complexity or display it, and which senior hire can carry the company into its next market. Venture firms increasingly package this work as a “platform.” Sound's version is more partner-shaped. The advantage is senior attention and a network accumulated across technology, media and finance. The limitation is capacity. A compact team has to choose where its involvement can change an outcome, which makes portfolio selection and founder fit unusually important.
Where Sound fits in the market
Sound sits between two recognizable venture types. On one side are the large multi-stage firms - Andreessen Horowitz, General Catalyst, Lightspeed and Thrive - with broad teams, formal platform services and the ability to fund a company across many rounds. On the other are operator- and celebrity-led funds whose edge is personal access, specialized experience or audience. Sound borrows from both: institutional funds and follow-on capacity, delivered through a compact partnership with a conspicuous cultural network.
Its portfolio is sometimes credited with famous names such as Airbnb, Uber and Spotify, but precision matters. Those were defining investments of A-Grade, the predecessor vehicle started by Kutcher, Oseary and Ron Burkle in 2010. Sound launched five years later with institutional backing and a more formal structure. The A-Grade record created the permission to raise Sound; it should not be mistaken for the same fund.
The distinction also reveals what the founders learned. A-Grade showed that product curiosity, consumer intuition and an entertainment network could produce access to unusual deals. Sound attempted to institutionalize those personal advantages. It added staff, dedicated funds, operating support and a strategy that could outlast a handful of partners' angel checks. Epstein's rise to general partner and now co-leader is part of that transition from famous investors to an enduring firm.
This is also why the closest competitor depends on the founder's need. A company seeking a vast recruiting organization or a large in-house policy team may prefer a mega-fund. A founder who wants a sector specialist may choose a healthcare, climate or crypto investor with deeper domain benches. Sound is most legible when the difficult problem is crossing boundaries: turning a technical product into a mainstream category, introducing it to a buyer outside the founder's circle, or building a brand before the market has settled on a vocabulary. Its location, team and history all point toward that crossover work.
A cleaner second act
The 2026 reorganization is a useful act of strategic editing. Deep-tech infrastructure and energy projects can demand longer time horizons, heavier technical diligence and different reserves than enterprise applications or consumer platforms. Keeping every conviction under one brand might have looked expansive while making the underlying funds harder to explain. Sound chose two clearer machines.
Oseary and Epstein's version of Sound will emphasize companies with commercial scale and use storytelling, branding and go-to-market as levers. Kutcher's new effort can pursue infrastructure, energy and deep tech without forcing those bets into Sound's model. The firms can remain friendly and even collaborate, but their customers - founders and limited partners - receive a cleaner description of what each pool of capital is designed to do.
For founders, the practical test is simple. Sound can help when a strong technical product needs translation, high-level introductions and partners who will stay engaged after the check clears. It is less obviously suited to every company, geography or capital plan; a specialist network is useful when it overlaps the problem. For limited partners, the proposition combines conventional private-tech exposure with a team willing to make some unusually concentrated calls.
The Hollywood origin will remain part of the firm's magnetism because it is true and memorable. But after eleven years, it is no longer the whole explanation. Sound Ventures has become an experiment in converting taste and access into an institution - broad enough to fund everyday software, narrow enough to wager heavily on a few model companies, and disciplined enough to split when one strategy became two.