DEPLOYDatabase

Company

Lambda

Lambda (legally Lambda, Inc., previously Lambda Labs) is a San Francisco GPU cloud founded in 2012 by brothers Stephen and Michael Balaban.

HQ
San Francisco, California, USA
Status
private

Overview

Lambda (legally Lambda, Inc., previously Lambda Labs) is a San Francisco GPU cloud founded in 2012 by brothers Stephen and Michael Balaban. The company began as a boutique deep-learning workstation and server builder for research labs and pivoted, roughly around 2018 to 2020, into a pure Nvidia-only public cloud that rents H100, H200, B200, B300, GB300 NVL72 and the newer Vera Rubin NVL72 rack systems by the hour, by reservation, or under multi-year committed contracts. Where CoreWeave courts hyperscalers and Nebius pitches a full Europe-hosted stack, Lambda has taken the middle of the neocloud market: on-demand access starting at a single GPU for developers alongside contracted supercomputer builds for named enterprise and lab customers, all on Nvidia silicon.

Its most-cited public number in 2026 is capital, not gigawatts. In August 2026 Lambda priced and closed a $926 million senior secured term loan B, rated Baa2 by Moody's and led by Morgan Stanley with MUFG, Citizens, Credit Agricole and Wells Fargo, describing it as the first investment-grade-rated TLB ever issued by a private neocloud. It is a GPU-asset-backed SPV structure secured by the servers themselves and by a multi-year contract with an unnamed investment-grade offtaker, and it follows a $1 billion senior secured credit facility Lambda closed in May 2026. That debt stack is the tell for how Lambda now grows: contracted customer, matching term-loan, deploy Nvidia gear, amortize over the useful life of the chips.

The executive bench built around that model is unusually telecom-flavored for a GPU cloud. The Balaban brothers moved to CTO (Stephen) and CPO (Michael), and Michel Combes, formerly CEO of Sprint and Altice, runs the company; Jerry Hunter (ex-Snap SVP Engineering) is Vice Chairman for Compute Delivery, Dave Ward is president of product and engineering, and Charles Fisher is CFO. Lambda has not published a company-wide GPU count, MW figure, or datacenter list, and does not disclose reserved-instance pricing publicly, so any specific inventory number for Lambda in 2026 is a trade-press estimate, not a company figure.

Verified record

Verified deployments
None on file
Active incidents
None on file

Key facts

Headquarters

Status

private

Type

brain

Founded

2012 in San Francisco by brothers Stephen and Michael Balaban

CEO

Michel Combes, formerly CEO of Sprint and Altice

Silicon partner

Nvidia-only fleet: A100, H100, H200/GH200, B200, B300 HGX, GB300 NVL72, and Vera Rubin NVL72 rack systems announced for the product line

Announced hourly on-demand pricing

H100 SXM 8x GPU $3.99/hr per GPU; B200 SXM6 8x GPU $6.69/hr per GPU; A100 SXM 80GB 8x $2.79/hr; GH200 $2.29/hr; A100 40GB $1.99/hr

August 2026 term loan B

$926M senior secured TLB, Moody's Baa2, SOFR + 3.00%, matures Dec 31 2030; billed as the first investment-grade-rated TLB by a private neocloud; Morgan Stanley lead, MUFG joint bookrunner, Citizens/Credit Agricole/Wells Fargo documentation agents

May 2026 credit facility

$1B senior secured credit facility closed May 2026, cited as the predecessor financing to the August 2026 TLB

Financing structure

GPU-asset-backed SPV secured by the GPU servers and the cash flows from a multi-year committed contract with an undisclosed investment-grade offtaker; fully amortizing over the useful life of the GPUs

Positioning

Nvidia-only public GPU cloud serving both single-GPU on-demand developers and multi-year enterprise supercomputer builds; not multi-vendor (no AMD/Cerebras/Gaudi), not hyperscaler-captive

Executive team

Stephen Balaban (CTO, co-founder), Michael Balaban (CPO, co-founder), Michel Combes (CEO), Charles Fisher (CFO), Leonard Speiser (COO), Steve Stewart (Chief Infrastructure Officer), Dave Ward (President, Product and Engineering), Jerry Hunter (Vice Chairman, Compute Delivery), Robert Brooks IV (CCO), Paul Zhao (GM Public Cloud)

Undisclosed

Lambda has not published a company-wide GPU count, total MW capacity, datacenter list, employee count, or reserved-instance pricing; any specific 2026 inventory number is a third-party estimate, not a company figure

Power capacity (planned)

3,000 MW target by 2030

GPU fleet (announced)

>1,000,000 target by 2030

Primary accelerator

NVIDIA H100 / B200 / GB300 NVL72

Data & sources

Web sources

8

8 sources backing this record.View all →

Partnerships

Every company Lambda has built with, alongside the ones only announced in a press release.

4 total3 operating1 just announced

Most-frequent partners

  1. 1Google1× joint117 total
  2. 2Microsoft1× joint
  3. 3NVIDIA1× joint1114 total
  4. 4NVIDIA1× joint

Most recent

Relationships

Current leadership (8)

Safety record

No incidents on record for Lambda.

Only active incidents are counted. Retracted incidents are excluded from this summary but remain reachable at their canonical URLs.

Full safety record: incidents, sourcing, and exposure data →

Lambda on the data-center map

Every verified data center Lambda operates that has a mapped coordinate (13 total). Filled circles are Tier-2 facilities and Tier-1 campuses under this operator; click any pin to open its record.

Lambda data centers (13)

AI data centers Lambda operates or is a party to, from the DEPLOY data center registry.

Recent coverage

Lambda in third-party press

Appears in DEPLOY reports