Feed/Companies/Tensor Machines

Tensor Machines

Every round on record, with the sources behind each and the investors on them.

BUILT ON

Visit site

A developer of AI hardware tools and physics-based models

tensormachines.orgSan Francisco, CAAI

LATEST ROUND

Pre-seed

$1.5M · 2026-10-07

RAISED ON RECORD · USD

$1.5M

across 1 round on record

INVESTORS

4

1 lead

SOURCES

2

from 2 publishers

Rounds

1

Feed/Companies

Company

Every round on record, with the sources behind each and the investors on them.

BUILT ON

Tensor Machines: $1.5M Pre-seed · Fireround

Pre-seed

$1.5M · 2026-10-07

led by Omni VC

Round	Amount	Announced	Lead	Buzz
Pre-seed	$1.5M	2026-10-07	Omni VC	2 outlets. #76 in Trending
2 sources · checked 1 time, last 5h ago [ VERIFIED · CORRECTED ] read against 2 sources Removed Kleiner Perkins · no source places them in this round Lead added Omni VC · finsmes.com Source 1 names Kleiner Perkins as lead; Source 2 names Omni VC as lead and lists three participants. I use Source 2’s more specific syndicate account.

Company research

Fetched today

Tensor Machines is building a physics-informed optimization layer that predicts what each GPU can sustain and guides how it should run.

Founders

Muneeb Rasool · Founder and CEO
At Meta Reality Labs, he led product work for Presence Platform; before Meta, he led product work at Metromile and was the founding Head of Product at ZenLedger, and he holds a master's degree in engineering and management from Carnegie Mellon University and served in the U.S. Army Reserve.
Mohamed Abdalla · Co-founder and CTO
Before Tensor, he led R&D electrical engineering at Ecolab and spent nine years at Atmosphere Recovery developing industrial instrumentation; he holds a master's degree in Machine Learning and a bachelor's degree in Electrical Engineering.
Hafiz Malik · Co-founder
He has spent fifteen years at the University of Michigan working at the intersection of hardware security, signal processing, and machine learning, with research using physical signals to authenticate, fingerprint, and monitor electronic systems.

WHY IT MATTERS

AI-infrastructure operators cannot tell from GPU specifications or hourly prices how much useful AI output their systems sustain, and Tensor Machines' tests found nearly 15% variation in token throughput across GPUs of the same model along with disproportionate power/performance tradeoffs. Its $1.5 million pre-seed funds an open-source benchmark and proprietary physics-based models that are in private beta with select bare-metal providers and neo-cloud design partners.

Checked against 9 sources today

SOURCES

Sources

[ 2 ]

Pre-seed$1.5M2026-10-072 sources

Investors

[ 4 ]

Pre-seed$1.5M2026-10-074 investors

Also named as backers in company research: NVIDIA Inception Program