Is Inviscid AI profitable in 2026?

No public revenue figures exist. Inviscid AI is a pre-revenue YC W26 startup (launched Feb 26, 2026) with two unconverted pilots announced March 2026; the company itself claims 36%+ energy savings but this is vendor-stated, not audited or revenue-linked.

Who is Inviscid AI's biggest competitor?

Phaidra, a Seattle-based, NVIDIA-backed startup that has raised roughly $120M including a $50M+ Series B in October 2025 and partnered with CoreWeave and Applied Digital by March 2026, making it 2-3 years ahead of Inviscid AI in the same data-center cooling niche.

Could AI replace Inviscid AI's own core technology?

A third-party technical review dated May 1, 2026 concluded Inviscid AI's physics-informed neural network architecture is not a defensible moat and would likely be replicated by competitors and academic labs within about 18 months.

Who owns Inviscid AI?

Inviscid AI is a privately held startup co-founded by Kabir Jain and Ziming Qiu, part of Y Combinator's Winter 2026 batch with Ankit Gupta as assigned YC partner as of February 26, 2026.

Has Inviscid AI raised outside funding?

No independently confirmed funding round has been reported beyond Y Combinator's standard batch terms, which typically provide around $500K in seed funding to accelerator companies; no lead investor or valuation is publicly documented as of mid-2026.

Does Inviscid AI have any real customers yet?

As of March 2026 it announced a smart-building pilot with Kajima/The GEAR in Singapore and a government MOU with MPSEDC in Madhya Pradesh, India, but no data confirms either has converted into a paid or renewed contract.

YC Winter 2026 · Singapore

Inviscid AI

Singapore startup using AI physics models to simulate industrial and data center cooling.

Protection under test

Assessed 27 July 2026 · Evidence as of 27 July 2026

Something worth protecting exists and is being tested right now.

On the recordSince 27 July 2026 · Checkable from Q1 2027 · Every call on the record Checkpoint: Kajima/The GEAR and MPSEDC pilots due to show paid conversion or renewal

The call

Inviscid AI is chasing a market where Phaidra already has customers, NVIDIA's backing, and years of head start, while its own core technology has no independent proof of holding up outside pilots.

Inviscid AI builds software that uses AI to simulate how things work in the physical world, helping construction and manufacturing companies predict outcomes and save energy. The core technology is not hard to copy because the underlying science will become standard within about 18 months, especially as NVIDIA (a chip giant that also funds the main competitor, Phaidra) releases similar tools for everyone. The one thing that could change this is if Inviscid's early customer projects prove that once a facility gets set up with their system, switching to a competitor becomes difficult and expensive; right now that proof doesn't exist.

Why

AI pressure: NVIDIA PhysicsNemo/Omniverse. NVIDIA's general-purpose physics AI toolkit could let others replicate Inviscid's core models within roughly 18 months. Could AI do the core job? Yes, likely.

Bigger players: Phaidra. Phaidra already runs live cooling optimization at CoreWeave and Applied Digital facilities with NVIDIA backing and $120M raised. Could a giant copy it? Yes, easily.

What protects it. Whatever integration depth and per-facility calibration data Inviscid builds during its Kajima and MPSEDC pilots, if it holds up.

What would change this call

MPSEDC India MOU becomes a paid multi-facility deployment by Q1 2027. Would show government interest translates into durable revenue.
NVIDIA ships a general-purpose Physical AI product for data center operations. Would show the core PINN technology is commoditizing as predicted.

Key numbers

Phaidra has raised about $120M total funding (2025-10-08)
Phaidra's Series B added $50M+, led by Collaborative Fund (2025-10-01)
Inviscid claims 36%+ energy cost cuts (vendor figure, unverified) (2026-03-04)

Named risks

Core PINN technology explicitly called not a moat, expected commoditized within ~18 months per third-party review. Severity: high
NVIDIA-backed Phaidra already has hyperscaler deals (CoreWeave, Applied Digital) in the same buyer segment. Severity: high
No confirmed paying customers; March 2026 pilots remain unconverted MOUs with no disclosed outcomes. Severity: high
No confirmed funding beyond YC's standard seed terms, raising runway uncertainty. Severity: medium
Conflicting public records on HQ location and founding year create identity/verification confusion. Severity: low

Rivals, as read in July 2026

Rival	Direction	Evidence
Phaidra	losing to	Phaidra raised ~$120M, NVIDIA-backed, partnered CoreWeave/Applied Digital by Mar 2026 vs Inviscid's unconverted pilots
NVIDIA PhysicsNemo/Omniverse	contesting	General-purpose PINN and digital-twin tooling threatens to commoditize Inviscid's core tech within ~18 months per May 2026 review
Ansys and Siemens Simcenter	contesting	Both incumbents already run PINN and AI-surrogate research programs in CFD/thermal simulation as of 2026
Submer Technologies	contesting	Signed its own data-center government MOU in Madhya Pradesh, India in July 2025, same target region as Inviscid's MPSEDC MOU

Questions people ask

Is Inviscid AI profitable in 2026?: No public revenue figures exist. Inviscid AI is a pre-revenue YC W26 startup (launched Feb 26, 2026) with two unconverted pilots announced March 2026; the company itself claims 36%+ energy savings but this is vendor-stated, not audited or revenue-linked.
Who is Inviscid AI's biggest competitor?: Phaidra, a Seattle-based, NVIDIA-backed startup that has raised roughly $120M including a $50M+ Series B in October 2025 and partnered with CoreWeave and Applied Digital by March 2026, making it 2-3 years ahead of Inviscid AI in the same data-center cooling niche.
Could AI replace Inviscid AI's own core technology?: A third-party technical review dated May 1, 2026 concluded Inviscid AI's physics-informed neural network architecture is not a defensible moat and would likely be replicated by competitors and academic labs within about 18 months.
Who owns Inviscid AI?: Inviscid AI is a privately held startup co-founded by Kabir Jain and Ziming Qiu, part of Y Combinator's Winter 2026 batch with Ankit Gupta as assigned YC partner as of February 26, 2026.
Has Inviscid AI raised outside funding?: No independently confirmed funding round has been reported beyond Y Combinator's standard batch terms, which typically provide around $500K in seed funding to accelerator companies; no lead investor or valuation is publicly documented as of mid-2026.
Does Inviscid AI have any real customers yet?: As of March 2026 it announced a smart-building pilot with Kajima/The GEAR in Singapore and a government MOU with MPSEDC in Madhya Pradesh, India, but no data confirms either has converted into a paid or renewed contract.

Timeline

2026-07-2623 GOP governors sign data-center energy pledge, a tailwind for cooling-efficiency vendors like Inviscid AI
2026-05-01Third-party review says Inviscid AI's PINN core is not a moat, replicable within about 18 months
2026-03-16Rival Phaidra partners with CoreWeave and Applied Digital on NVIDIA Max-Q AI factories
2026-03-04Inviscid AI publicly launches, publishing vendor-claimed performance benchmarks
2026-03-01Inviscid AI announces Kajima/The GEAR Singapore pilot and MPSEDC India MOU
2026-02-26Inviscid AI launches via YC W26 with Ankit Gupta as assigned YC partner
2025-10-01Rival Phaidra closes $50M+ Series B led by Collaborative Fund
2025-04-14NVIDIA-affiliated researchers publish Physical AI framework for data centers

Revision history

27 July 2026

First assessed. Protection under test. The text above is fixed as of this date and is not edited afterwards.

In the same band

Sequence Markets Piris Labs REV1