Wispr Raises $280 Million, Says Canto Cuts Noisy-Speech Errors to 5%–10%

Wispr Raises $280 Million, Says Canto Cuts Noisy-Speech Errors to 5%–10%

August 18, 2026

SAN FRANCISCO, August 17, 2026, 12:06 PDT

  • Wispr raised $280 million at a $2 billion valuation.
  • Canto is a preview, not a fully documented public benchmark.
  • Flow now spans dictation, meeting notes and speech-model research.

Wispr has raised $280 million in a Series B led by Menlo Ventures. The private financing values the voice-software maker at $2 billion. Wispr also previewed Canto, its first proprietary speech model.

The deal gives Wispr far more capital to improve dictation accuracy. It also raises the commercial bar. The valuation is almost three times the $700 million figure reported nine months ago.

For users, the useful test is simpler. Canto must reduce corrections enough to justify another paid tool. Built-in dictation is free, while Flow Pro costs $15 monthly.

Financing milestoneNew capitalTotal raisedValuationStatus
June 2025 Series A$30 million$56 millionNot disclosedCompany announced
November 2025 follow-on$25 million$81 million$700 millionReported post-money
August 2026 Series B$280 million$361 million$2 billionCompany announced

The Series B is 11.2 times larger than Wispr’s previous disclosed round. Its valuation has increased about 186% since November. The latest investor group includes existing backers and several new funds.

Chief Executive Tanay Kothari said accuracy determines whether voice becomes a daily interface. “The whole reason to talk instead of type is to stay inside your own train of thought,” he wrote.

Canto measureEarlier resultCanto claimWhat is known
Word error rate in hardest conditionsMore than 30%5%–10%Company test; background noise, wind, music or heavy accents
Dictations needing edits in everyday useBaseline not published30%–35% reductionCompany expectation
Model availabilityExisting Flow modelsPreviewGeneral rollout date not stated
Personal contextNot comparableUses dictionaries and nearby namesCompany description

The largest claimed gains appear under difficult acoustic conditions. That matters on phones, where cars, streets and offices create poor recordings. Wispr has not published an independent evaluation or a full test set.

Flow planPriceUsageSelected controls
BasicFree2,000 words weekly on desktop; 1,000 on iPhone; unlimited Android use is temporaryPrivacy Mode; custom dictionary; 100+ languages
Student$0 annually for eligible studentsUnlimited dictationClaimed on desktop or web
Pro$15 per user monthly; about $12 monthly on annual billingUnlimited across supported platformsCommand Mode; priority support; early features
EnterpriseContact salesUnlimitedSSO/SAML; enforced policies; advanced dashboards

The same speech engine serves every paid tier, according to Wispr. That makes limits, management and security the main price differences. The company supports Mac, Windows, iPhone and Android.

Product or initiativeDateStatusPrimary job
Flow DictationCurrentAvailableTurns speech into edited text across apps
Flow NotetakerAugust 5, 2026LaunchedCreates meeting notes, summaries and action items
CantoAugust 17, 2026PreviewedImproves speech recognition in real conditions
Advanced Interfaces LabJuly 24, 2026OperatingResearches voice-led human-computer interaction

Wispr is also moving beyond dictation. Its Notetaker product entered the meeting market this month. The Advanced Interfaces Lab is researching systems that convert speech into actions, not only text.

Scale gives that expansion a base. Wispr says Flow users have produced more than 60 billion words. It also claims use across more than 10,000 enterprises and almost all Fortune 500 companies. Those figures have not been independently audited.

Privacy remains central for business adoption. Wispr says Privacy Mode prevents dictation data from training its models. It lists SOC 2 Type II, ISO 27001 and HIPAA-related controls, with enforcement varying by plan.

Risks: Canto is still a preview, and its performance numbers come from Wispr. The company discloses no revenue in the announcement. A richer valuation therefore rests on product adoption, retention and model gains that outsiders cannot yet verify.

The financing buys Wispr time to prove those gains at scale. The next evidence should come from public benchmarks, broad Canto availability and measurable reductions in user corrections.

BEZ KABLI • EXTENDED COVERAGE

Further analysis

What did Wispr announce?
Wispr announced a $280 million Series B led by Menlo Ventures. The round values the private company at $2 billion and brings total capital raised to $361 million.
What is Canto, and can everyone use it now?
Canto is Wispr's first proprietary speech model. It is a preview. Wispr has not stated a general rollout date, so users should not assume the model is already active in every Flow account.
How much better does Wispr say Canto performs?
Wispr says word error rates in its hardest test conditions fall from more than 30% to 5%–10%. It also expects 30%–35% fewer edited dictations in everyday use. These are company claims, not independent benchmark results.
How much does Wispr Flow cost?
Flow Basic is free with weekly limits on desktop and iPhone. Flow Pro costs $15 per user monthly, or about $12 monthly with annual billing, and includes unlimited dictation. Eligible students can claim a separate free annual plan with unlimited use.
What is the biggest uncertainty for buyers?
Canto's real-world accuracy is not independently verified, and broad availability is unspecified. Buyers should test their own accents, languages, microphones and noisy settings before relying on the headline error-rate range.

Mateusz Ługowik

Mateusz Ługowik is a senior markets reporter at Bez-kabli.pl, specializing in technology stocks, artificial intelligence and global financial markets. A graduate of the University of Gdańsk, he previously worked in investment research and market analysis. His coverage helps readers understand the key trends, companies and innovations influencing investors worldwide.