Is Kalpa Labs profitable in 2026?

No. CB Insights lists Kalpa Labs' FY revenue as $0 as of June 22, 2026. The company is a 2-person, pre-revenue startup with its speech API still in gated early access as of July 19, 2026, and no disclosed paying customers.

How much funding has Kalpa Labs raised?

Reports conflict. CB Insights and Caplight report a $500,000 Seed VC round led by Y Combinator with Tekedia Capital as co-investor, closed around June 23, 2026, though StartupHub.ai separately cites $2M raised as of May 31, 2026, unconfirmed by primary sources.

Who is Kalpa Labs' biggest competitor?

ElevenLabs, which reportedly reached an $11 billion Series D valuation as of July 2, 2026, is the dominant named incumbent, alongside platform giants OpenAI (gpt-realtime) and Google (Gemini TTS) shipping comparable speech features as of mid-2026.

Could AI replace Kalpa Labs?

Kalpa Labs' entire product is itself an AI foundation model, so its niche is exposed to being absorbed: OpenAI and Google already bundle comparable speech capabilities (gpt-realtime, Gemini TTS) into general platforms as of mid-2026, per independent benchmark coverage.

How many employees does Kalpa Labs have?

Y Combinator's company page lists Kalpa Labs' team size as 2 as of July 20, 2026. The company was actively hiring founding ML research engineers at $140K-$220K in San Francisco as of December 2025 and July 2026 postings.

Does Kalpa Labs have any paying customers?

No named customers or revenue have been disclosed. CB Insights lists FY revenue as $0 (2026-06-22), and Kalpa's Speech API remains gated in early access as of July 19, 2026, with the company soliciting pilots rather than announcing signed clients.

YC Fall 2025 · United States

Kalpa Labs

Two-person startup building a generalist speech AI model, trained cheaply.

Little protection left

Assessed 27 July 2026 · Evidence as of 27 July 2026

Not much stands between this business and its pressures.

On the recordSince 27 July 2026 · Checkable from H2 2026 · Every call on the record Checkpoint: Kalpa's API moves from early access to a public launch with real pricing

The call

Kalpa is a tiny, fast-moving team entering a speech AI market where better-funded rivals and Google and OpenAI already ship comparable products.

Kalpa Labs is a 2-person company trying to build a general-purpose speech model that can handle many different voices and languages. The problem is that much bigger, better-funded companies like ElevenLabs, OpenAI, and Google have already entered this space and released competing products in the last 6 months. Kalpa has no special advantage that would make it hard to copy, such as exclusive data, customers locked into their system, or a distribution network that others cannot easily replicate. The biggest risk is that OpenAI, Google, or ElevenLabs could simply add Kalpa's features to their existing products faster than Kalpa can scale up, which would make Kalpa's offering unnecessary. The one thing that could shift this picture is if Kalpa attracts a major compute partner or real paying customers before the bigger players ship similar technology.

Why

AI pressure: OpenAI gpt-realtime and Google Gemini TTS. OpenAI and Google already ship speech models with similar tricks at far greater scale, funding, and distribution than Kalpa has. Could AI do the core job? Yes, likely.

Bigger players: ElevenLabs. ElevenLabs, valued near $11B, already has a generally available product (v3, Feb 2026) covering the same generalist speech ground. Could a giant copy it? Yes, easily.

What protects it. Kalpa trains models fast and cheap, claiming under $1,000 for an 800M model, but that speed is not something rivals can't copy.

What would change this call

Kalpa Labs' API moves from early access to public launch with published pricing within 6 months. Would show Kalpa can commercialize before larger rivals close the gap.
Google, OpenAI, or Cartesia ship RVQ tokenization or long-context features matching Kalpa's claims. Would show Kalpa's stated technical edge has been erased by frontier labs.
Kalpa Labs announces a named paying customer or a new funding round. Would show Kalpa is gaining real traction despite the incumbents.

Key numbers

Team size: 2 people (2026-07-20)
Claims sub-$1,000 cost to train an 800M model (2025-11-11)
ElevenLabs v3 reached general availability, 70+ languages (2026-02)

Named risks

No durable moat: no exclusive data, customer lock-in, or regulatory barrier identified. Severity: high
Compute constrained, publicly seeking 128x B200 GPU partner to scale beyond current models. Severity: high
Conflicting funding figures ($500K vs $1M vs $2M) suggest limited runway and unclear financial position. Severity: medium
Fast-moving competitive frontier (ElevenLabs, Cartesia, Hume, Google, OpenAI) shipping major releases every few months. Severity: high
Only 2 confirmed staff with no disclosed customers or revenue as of mid-2026. Severity: medium

Rivals, as read in July 2026

Rival	Direction	Evidence
ElevenLabs	losing to	$11B valuation vs Kalpa's $500K-$2M raised, unconfirmed, as of mid-2026
OpenAI	losing to	gpt-realtime and gpt-4o-mini-tts ship natively inside its platform as of May 13, 2026
Google	losing to	Gemini TTS reportedly ties top blind-listening rankings as of July 2, 2026
Cartesia	losing to	Shipped Sonic 4 with ~40ms latency by May 13, 2026, ahead of Kalpa's commercial launch
Hume AI	contesting	Shipped Octave 2 in the emotional-expressiveness niche as of May 13, 2026
Sesame	contesting	Full-duplex conversational speech model cited as adjacent competitor as of May 12, 2026

Questions people ask

Is Kalpa Labs profitable in 2026?: No. CB Insights lists Kalpa Labs' FY revenue as $0 as of June 22, 2026. The company is a 2-person, pre-revenue startup with its speech API still in gated early access as of July 19, 2026, and no disclosed paying customers.
How much funding has Kalpa Labs raised?: Reports conflict. CB Insights and Caplight report a $500,000 Seed VC round led by Y Combinator with Tekedia Capital as co-investor, closed around June 23, 2026, though StartupHub.ai separately cites $2M raised as of May 31, 2026, unconfirmed by primary sources.
Who is Kalpa Labs' biggest competitor?: ElevenLabs, which reportedly reached an $11 billion Series D valuation as of July 2, 2026, is the dominant named incumbent, alongside platform giants OpenAI (gpt-realtime) and Google (Gemini TTS) shipping comparable speech features as of mid-2026.
Could AI replace Kalpa Labs?: Kalpa Labs' entire product is itself an AI foundation model, so its niche is exposed to being absorbed: OpenAI and Google already bundle comparable speech capabilities (gpt-realtime, Gemini TTS) into general platforms as of mid-2026, per independent benchmark coverage.
How many employees does Kalpa Labs have?: Y Combinator's company page lists Kalpa Labs' team size as 2 as of July 20, 2026. The company was actively hiring founding ML research engineers at $140K-$220K in San Francisco as of December 2025 and July 2026 postings.
Does Kalpa Labs have any paying customers?: No named customers or revenue have been disclosed. CB Insights lists FY revenue as $0 (2026-06-22), and Kalpa's Speech API remains gated in early access as of July 19, 2026, with the company soliciting pilots rather than announcing signed clients.

Timeline

2026-07-20YC job listing confirms 2-person headcount, hiring ML engineers at $140K-$220K
2026-07-19Kalpa Speech API documented as still gated, early-access only
2026-06-22CB Insights reports $500K Seed VC round closed, led by YC with Tekedia Capital
2026-05-13Competitor Cartesia ships Sonic 4 with roughly 40ms latency
2026-03-05Third-party coverage (tryfondo) of Kalpa Labs' generalist speech model launch
2026-02Competitor ElevenLabs v3 reaches general availability with 70+ languages
2025-12-15Kalpa Labs posts founding ML infra role, seeks 128x B200 GPU compute partner
2025-11-11Kalpa Labs publicly launches on Y Combinator Launch platform
2025-09-10Kalpa Labs company blog covers early version of launch announcement
2025Kalpa Labs founded, joins Y Combinator Fall 2025 (F25) batch

Revision history

27 July 2026

First assessed. Little protection left. The text above is fixed as of this date and is not edited afterwards.

In the same band

Exonic Openroll Semiotic