morphed

Luma AI / Dream Machine Statistics: Users, Funding & Scale

Last updated June 12, 2026

Luma AI: Dream Machine serves 30M+ users; raised a $900M Series C (Nov 2025) led by Saudi-backed Humain at a $4B+ valuation; building “Project Halo,” a 2-gigawatt GPU supercluster, with ~$1B+ cumulative funding.

Luma AI’s Dream Machine became one of the most-used AI video products in the world, reaching over 30 million users on the strength of fluid camera motion and photorealistic output. In November 2025 the company raised one of the largest rounds in AI video - $900 million led by Saudi Arabia’s Humain - and announced a 2-gigawatt GPU supercluster to train its Ray model series.

Key statistics

30M+2025–2026
Dream Machine users

Dream Machine reportedly serves over 30 million users globally - among the largest user bases of any AI video product.

$900MNov 2025
Series C raise

Luma raised $900 million in November 2025, led by Humain (Saudi PIF) with AMD Ventures and Andreessen Horowitz participating.

Source: CNBC
$4B+Nov 2025
Valuation

The Series C valued Luma AI upwards of $4 billion, confirmed by CNBC.

Source: CNBC
2GW2025
Project Halo supercluster

Luma and Humain are partnering on a 2-gigawatt AI supercluster in Saudi Arabia - one of the largest planned GPU deployments in the world.

Source: CNBC

How Luma’s war chest compares to AI video rivals

Luma’s $900M Series C is larger than the total lifetime funding of most AI video companies. Only Runway ($855M+ cumulative) is in the same capital league among Western pure-plays; Kling is funded internally by publicly listed Kuaishou.

CompanyCumulative fundingLatest valuationCapital strategy
Luma AI$1B+$4B+ (Nov 2025)Sovereign-backed compute (2GW Project Halo)
Runway$855M+$5.3B (Feb 2026)Enterprise revenue + venture rounds
Pika Labs~$135M$470–700M (2024)Lean creator-focused team
Kling (Kuaishou)Internal funding$20B reported spin-offParent-company balance sheet

Funding figures from CNBC, TechCrunch, CB Insights, and Sacra as of June 2026.

How does Luma compare to other AI video startups?

Luma raised more in one round than most AI video companies have raised in total, betting that frontier multimodal models require frontier compute.

$1B+2025
Cumulative funding

Total private capital exceeds $1 billion across rounds with a16z, NVIDIA, Matrix, Amplify, and Humain.

Source: Nextomoro
20212021
Founded

Founded September 2021 by Amit Jain, Alex Yu, and Alberto Taiuti - initially focused on 3D capture before pivoting to video generation.

Source: Nextomoro
$135M2024
Pika total funding (comparison)

Competitor Pika Labs has raised ~$135M at a $470–700M valuation - an order of magnitude below Luma’s war chest.

How we compiled this data

The Series C details come directly from CNBC’s November 2025 reporting, which named investors and confirmed the $4B+ valuation. The 30M user figure traces to industry profiles rather than a Luma disclosure, so we label it "reportedly." The funding-comparison table uses each company’s most recent confirmed round. Last full review: June 12, 2026.

Before you cite these numbers

  • The 30M+ user count is industry-reported, not a Luma press release, and the registered/active distinction is unspecified.
  • Project Halo’s 2GW supercluster is an announced plan; announced AI infrastructure projects routinely change scale or schedule.
  • Luma has published no revenue figures; do not equate its funding scale with sales traction.

Frequently asked questions

How many people use Luma Dream Machine?

Dream Machine reportedly serves over 30 million users globally, making it one of the largest AI video user bases alongside Kling AI’s 60M+ creators.

How much is Luma AI worth?

Luma AI was valued upwards of $4 billion in its November 2025 Series C, which raised $900 million led by Saudi-backed Humain.

What is Project Halo?

A planned 2-gigawatt AI supercluster in Saudi Arabia, built in partnership between Luma AI and Humain - one of the largest GPU deployments announced anywhere.

Sources

Figures on this page are compiled from the following publishers and reports. Where sources disagree, we present the range and note the methodology difference.