Signals Inbox·August 21, 2026·Voice AI

Is Wispr really worth $2B today?

Wispr can justify a $2 billion valuation today, but only as a hypergrowth bet: the revenue, retention and paid-business adoption are real, while Apple, Google and Wispr’s still-early expansion beyond dictation make the price demanding.

Q1What exactly changed when Wispr hit a $2B valuation?

Yes: Wispr’s jump to a $2 billion valuation is real, and the striking part is how quickly investors repriced the company.

On August 17, Wispr announced a $280 million Series B led by Menlo Ventures at a $2 billion valuation, with Notable Capital, NEA, Neo Ventures and 8VC returning, and Acrew, Forerunner, Goodwater and Peak XV joining. The round took total funding to $361 million. Nine months earlier, TechCrunch reported that a $25 million round led by Notable Capital valued Wispr at roughly $700 million post-money. That means the valuation rose about 2.9x in nine months.

The speed is easier to see when we look at the product timeline. Wispr was founded in 2021, pivoted away from hardware, and released Flow on Mac in 2024 before expanding to iOS, Windows and Android. Investors are assigning a $2 billion value to a software product that has only been broadly available for around two years. Granola shows that this kind of speed is no longer unique in AI: its valuation moved from $250 million to $1.5 billion in less than a year. Still, Wispr’s jump remains extreme by normal software standards.

Q2How much revenue is Wispr making now?

Wispr is making at least $50 million in annual recurring revenue (ARR), and the actual number today is almost certainly higher.

The cleanest number comes from Wispr itself. In a Head of Finance job posting published in May, the company said Flow had gone from $2 million to $50 million in ARR in under 12 months. That is much stronger evidence than an outside estimate because Wispr used the figure in its own recruiting material.

Since then, Menlo Ventures has said Wispr’s revenue is up more than 30x year over year. We should treat that May milestone as the last hard number we can verify, rather than as today’s exact ARR. The company has not publicly replaced it with a newer absolute figure.

Percentage growth still leaves a wide range of possible current revenue. If Wispr crossed the milestone shortly before the job post went live, today’s ARR could be much closer to $70 million than $100 million. If it crossed earlier, the number could be higher. We use the disclosed figure as a floor and show how the valuation changes at different revenue levels instead of guessing.

Q3Is Wispr’s revenue multiple too high?

At the last hard revenue figure, Wispr is priced at about 40x ARR; today, the real multiple is probably lower.

The last disclosed revenue base predates the financing, so the headline multiple is backward-looking. If ARR has already reached $75 million, the multiple falls to 26.7x. At $100 million, it is 20x.

The difference is huge. The top end demands near-perfect growth; 20x is still expensive, but recent private AI rounds show investors paying that kind of price for exceptional growth and retention. What matters now is how quickly Wispr can grow revenue into the price investors just paid.

Wispr valuation multiple at different ARR levels

Q4How expensive is Wispr compared with public software?

Wispr is currently valued at a multiple far above normal public software, so its valuation only works if its growth stays far above normal too.

A live benchmark covering 143 public SaaS companies puts the median revenue multiple at about 3.1x and the average at 5.4x. The same sample grows roughly 14% year over year on average. Zoom is currently around 6.3x annualized revenue while its latest reported quarter grew 5.5%. Duolingo is around 5.2x with recent annualized revenue growth near 18%.

The useful comparison sits at the other end of the market. Palantir recently reported 93% year-over-year quarterly revenue growth, and current public-market data values it at roughly 63x annualized revenue. Public investors can clearly tolerate huge sales multiples when growth, margins and market leadership are exceptional.

Wispr is much smaller and has far less proof on margins and large long-term contracts. Its growth rate deserves to be valued above ordinary SaaS, but the price investors just paid still needs more than growth alone.

Public software revenue benchmarks

Q5Is Wispr more expensive than ElevenLabs, Sierra and Granola?

Wispr looks expensive beside its closest private AI peers, but it is still inside the range investors are paying for breakout AI applications.

ElevenLabs is the cleanest benchmark. It raised $500 million at an $11 billion valuation after finishing 2025 above $330 million ARR, which works out to roughly 33x revenue. On disclosed numbers, Wispr is about 20% more expensive even though ElevenLabs already has a much larger revenue base.

Sierra shows how extreme the top of the private AI market has become. It passed $100 million ARR in seven quarters, entered this year above $150 million ARR, and later raised $950 million at a valuation above $15 billion. Even using that $150 million floor, the implied multiple is above 100x. Granola gives us a more product-adjacent comparison: its meeting-notetaker business jumped from $250 million to $1.5 billion in less than a year, without publishing enough revenue for a clean multiple.

The ranking is fairly clear. Wispr is more aggressively priced than ElevenLabs on disclosed revenue, cheaper than Sierra’s most extreme benchmark, and already valued above Granola despite moving into Granola’s meeting-notes territory only recently. High, yes. Unique, no.

Wispr versus selected private AI peers

Q6Is Wispr still growing fast enough to support $2B?

Yes, for now: Wispr’s recent growth is fast enough to make a $2 billion valuation plausible, even though that pace will inevitably slow.

Wispr’s disclosed ARR base expanded 25x in under a year. Earlier, TechCrunch reported revenue growing 60% month over month. By November, CEO Tanay Kothari said Flow had grown 40% month over month since June. Menlo’s latest update says revenue is now more than 30x higher year over year.

Those figures cover different periods and reflect a changing growth rate. The important part is that growth stayed abnormally high even as Wispr moved from a tiny base into tens of millions of ARR. That is enough to make a high multiple worth taking seriously.

The next test is harder because every new doubling now requires far more absolute revenue. If Wispr can still double while moving through $100 million ARR, today’s valuation starts to look much less aggressive.

Q7Do people keep using Wispr after the novelty wears off?

The retention data is unusually strong: roughly 70% of Wispr users are still around after 12 months.

Earlier, Kothari told TechCrunch that the subscription business converted about 19% of users from free to paid. A 19% free-to-paid conversion is unusually strong for a productivity app that users can try casually.

Usage also deepens over time. Wispr says that after six months, the average user writes about 72% of their characters with Flow across nearly 70 apps and sites. The company now says users have written more than 60 billion words with the product. That is more persuasive than download counts: the tool is replacing an existing behavior, not just sitting beside it.

There was also a useful failure case. In 2025, Wispr noticed weaker retention as more non-technical users arrived and failed to understand that Flow worked outside its own app. The company changed onboarding to push people into the apps they already used. Kothari later repeated the same 12-month retention level when discussing India, which gives the metric more weight than a single cohort snapshot.

Q8Are companies actually paying for Wispr at scale?

They are: Menlo now says tens of thousands of businesses pay for Flow, although the average contract size remains undisclosed.

Wispr also says people at almost all Fortune 500 companies use the product and that it is present in more than 10,000 enterprises. Those figures show broad workplace penetration, while company-wide contracts are a separate question.

The trajectory is also visible earlier. In November 2025, TechCrunch reported that Wispr had reached 270 Fortune 500 companies and was adding roughly 125 enterprise customers per week. Since then, the product has added centralized billing, shared dictionaries, admin controls, single sign-on, automated user management and audit features. Those are the kinds of features that move a tool from employees buying it themselves toward company-wide purchasing.

The missing piece is account depth. Wispr has not disclosed enterprise ARR, average contract value, net revenue retention or the percentage of revenue coming from centrally purchased seats. Businesses are clearly adopting Flow; we still do not know how deep those accounts go.

Q9Is voice AI demand big enough for Wispr to keep growing?

The market is growing fast enough to support a much larger Wispr, although dictation is only one slice of voice AI.

A G2 report published this year analyzed 1,419 verified reviews and surveyed 17 voice-AI vendors, including Wispr, ElevenLabs, Retell, Synthflow and others. It found that 88% of the platforms surveyed saw adoption rise rapidly over the previous 12 months, while 76% of users reported significant or transformational operational ROI. A separate consumer-and-business survey from Voices found that 55% of consumers had already used voice to interact with AI.

Company revenue is the cleaner proof. ElevenLabs ended 2025 above $330 million ARR, while Sierra crossed $100 million ARR in seven quarters and moved above $150 million shortly afterward. Granola has attracted a $1.5 billion valuation around a voice-heavy meeting workflow. Different jobs, different economics, but real money is already moving into voice-driven interfaces.

We would still avoid calling “voice AI” one giant market. Customer-service agents, speech generation, dictation and meeting notes have different buyers and economics. Wispr’s opportunity depends on whether it can expand from typing replacement into a broader work interface, which is exactly what its recent product moves are trying to do.

Q10Can Apple and Google make Wispr irrelevant?

Yes, they can damage Wispr badly, and this is currently the clearest threat to the $2 billion valuation.

Google introduced Rambler, a Gemini-powered dictation feature inside Gboard that removes filler words, understands mid-sentence corrections and supports multilingual code-switching. Gboard already ships to hundreds of millions of Android users. Apple then announced a new systemwide dictation experience built into the iOS keyboard, with automatic spelling, punctuation and capitalization corrections across apps.

That changes Wispr’s competitive problem. Earlier, the company mainly had to beat other startups such as Willow, Aqua, Superwhisper and Typeless. Now the default keyboards themselves are moving closer to Flow’s core experience. Apple and Google can distribute those features without asking users to install another keyboard or pay $12 to $15 per month.

Wispr needs to stay visibly better. Cross-platform continuity, personal dictionaries, context from previous usage, meeting memory and eventually voice-driven actions are the areas where it can create distance. If native dictation gets “good enough” and Wispr remains mostly a transcription layer, pricing power will weaken quickly.

Q11Can Wispr’s Canto model keep Apple and Google behind?

Canto improves Wispr’s technical position today, but we have not yet seen enough independent evidence to call it a durable moat.

Alongside the latest financing, Wispr introduced its first proprietary speech model, Canto. The company says it cuts word-error rates in difficult conditions such as wind, background noise and heavy accents from above 30% to around 5% to 10%, and should reduce the share of dictations needing edits by roughly 30% to 35%.

The metric Wispr cares about internally is “zero edit rate”: how often a user can dictate and send the result without touching the keyboard. That is a smart target because a small transcription error can destroy the convenience of voice input. If that improvement holds across millions of real dictations, the benefit should show up in retention and frequency rather than only in a benchmark.

Still, the current performance claims come from Wispr. Google, Apple, OpenAI and specialist speech companies have huge datasets and research teams of their own. The technical moat gets much stronger if user history, personal vocabulary, cross-app context and the proprietary model reinforce each other faster than competitors can copy the experience. That proof is still missing.

Q12Is Wispr becoming more than a dictation app?

Yes, lately Wispr has started moving beyond dictation in a way that makes the $2 billion valuation easier to understand.

The clearest move is Notetaker, launched recently as Wispr’s first product beyond dictation. It records meetings without sending a bot into the call, identifies speakers, creates transcripts and summaries, and lets users query meeting history. It also carries over context such as a user’s personal dictionary and can expose meeting notes to Claude, ChatGPT and other tools through MCP, a standard that lets AI tools connect to outside data.

The pieces fit together. Flow captures what a user says while working. Notetaker captures what happens in meetings. Canto improves the speech layer underneath both. Wispr’s Advanced Interfaces Lab is now working on systems that can use context and eventually turn spoken intent into actions.

For the valuation, we still have to focus on what exists today. Notetaker is new, and the action layer is still a roadmap. For now, Flow’s dictation business is doing most of the financial work. The upside comes from proving that these adjacent products increase revenue per user and make Wispr harder to replace.

Q13What revenue does Wispr need for $2B to look reasonable?

Wispr needs roughly $100 million ARR for $2 billion to look aggressive rather than extreme, and around $200 million for the valuation to look comfortable by high-growth software standards.

At $100 million ARR, the multiple falls to 20x. At about $133 million, it is 15x. At $200 million, it is 10x. Starting from the last revenue figure Wispr disclosed, Wispr needs roughly one more doubling to reach the first threshold and about 4x growth to reach the last.

Current pricing also makes the scale easier to picture. Flow Pro costs $15 per user monthly or about $12 on an annual plan. At $144 per year, $100 million ARR equals roughly 694,000 full-price-equivalent annual seats. $200 million equals about 1.39 million. Real customer economics will differ because enterprise deals and regional pricing change the average revenue per user, but the order of magnitude is useful.

Wispr already says it has millions of users. A relatively small share of that active base can support the valuation if those users keep paying and more companies move toward centrally purchased seats.

Revenue required to compress Wispr’s $2B multiple

Q14What would have to go right for Wispr to be worth much more than $2B?

Wispr can grow well beyond $2 billion if voice becomes a default way of interacting with software and Flow becomes the layer users choose across devices.

The bull case starts with habit. A product that captures most of a user’s daily writing can become unusually sticky because it sits inside messages, documents, prompts, code and search. Wispr already has evidence that usage deepens over time, which gives it a better starting point than an AI app people open for one narrow task.

There is also plenty of room to sell more inside companies. Menlo says tens of thousands of businesses already pay for Flow. If those accounts grow from a few enthusiastic employees into hundreds or thousands of company-paid seats, revenue can rise much faster than consumer subscriptions alone.

Then there is the product itself. Notetaker adds meeting context, while Canto gives Wispr more control over speech recognition. If spoken requests eventually trigger actions across software, customers could end up paying Wispr for far more than dictation. In that version of the story, today’s investors bought into the interface layer before it was fully built.

Q15What could break Wispr’s $2B valuation?

The valuation breaks if Wispr’s growth slows before the product becomes meaningfully harder to replace.

The most immediate risk is distribution. Google and Apple have already moved AI dictation into default keyboards. Microsoft, OpenAI and Anthropic also have obvious places to add richer voice input. Those companies can bundle the feature into much larger products, making a separate $12-to-$15 monthly subscription harder to defend.

India shows another problem. Kothari said it became Wispr’s second-largest market by users and revenue and briefly reached around 100% month-over-month growth after a local push. Sensor Tower found that India represented 14% of installs but only about 2% of in-app purchase revenue over one measured period. The local annual plan is also much cheaper than standard U.S. pricing. Wispr can add users very quickly without revenue growing at the same speed.

The newer products also need to earn their place. Meeting notes are already crowded, and owning a speech model does not guarantee a lasting lead. If Notetaker stays peripheral, Canto fails to hold an accuracy advantage and native keyboards close the quality gap, Wispr could still be a good business while looking overvalued at $2 billion.

Q16So, is Wispr really worth $2B today?

Yes, we think Wispr can justify a $2 billion valuation today, but only as a high-risk hypergrowth valuation rather than as a price already supported by today’s revenue and business scale.

What convinces us most is the combination of revenue, retention and speed. Menlo now describes growth as more than 30x year over year, 12-month retention is roughly 70%, millions of people use the product, and tens of thousands of businesses pay for it. There is enough business underneath the AI story to take the valuation seriously.

The price is still demanding. The last hard revenue figure implies roughly 40x ARR, richer than ElevenLabs was at its $11 billion round. Apple and Google are now attacking the same core behavior from the iOS and Android level, and Wispr has yet to prove that Notetaker, Canto or future voice actions create enough extra value to offset that distribution advantage.

Our verdict: aggressive but plausible. If current ARR is already moving toward $75 million to $100 million and Wispr keeps retention near present levels while expanding paid business deployments, the multiple will compress quickly and $2 billion will look reasonable. If growth falls sharply before that happens, the valuation will have run ahead of the business.

The next evidence we would watch is simple: a new absolute ARR figure, data showing whether business accounts are expanding, and whether users adopt Notetaker and Canto deeply enough to make Wispr more useful than the increasingly capable voice tools already built into their devices.