Innovation & Tech

Fast Company's 13 Applied AI Winners: Alembic Leads With $645M Valuation

Fast Company names 13 applied AI winners for 2026, led by Alembic's $145M Series B at a $645M valuation and Genspark's $250M annualized revenue.

By Amara Osei

5 min read

Updated

The 13 next big things in applied AI for 2026
The 13 next big things in applied AI for 2026Karen Roe / Openverse

What's News

  • Alembic closed a $145M Series B led by Accenture and Prysm Capital in November 2025, valuing it at about $645M.
  • Genspark says it reached $250M in annualized revenue by spring 2026, with more than 5,000 enterprise customers.
  • Replit says 85% of Fortune 500 companies build with its platform and it has more than 50 million users.
  • A Credo AI study found 60% of organizations deploy AI widely but only 4% govern it organization-wide.
  • Yutori's customers, including Meta, make more than 100 million API calls to its models every month.

Applied AI startup Alembic closed a $145 million Series B in November 2025 at a valuation of roughly $645 million — the strongest deal signal in Fast Company's new list of the 13 next big things in applied AI for 2026.

The round was led by Accenture and Prysm Capital. Nvidia was Alembic's first enterprise customer, and Nvidia CEO Jensen Huang has publicly praised the company's "causal engine," which analyzes thousands of variables — from broadcast and digital ad spend to macroeconomic shifts and executive appearances — to map enterprise actions to outcomes. Mars and Delta Airlines are also customers. Accenture now offers Alembic's causal-AI analysis to its consulting clients.

The list reflects a shift in how AI earns attention. As Fast Company puts it, 2026 may be the year "AI stopped getting brownie points simply for being amazing" — what impressives now is AI that performs useful tasks, quietly, in areas that let humans focus on the work they still do best.

Which companies made the cut?

The honorees span enterprise productivity, translation, robotics, and medical safety:

  • Alembic — causal AI that proves how decisions turn into outcomes
  • Canva — AI 2.0 design tools generating more than 400 designs per second
  • Credo AI — governance automation via its Govern AI Assistant (GAIA)
  • DeepL — real-time voice-to-voice translation, boosted by its Mixhalo acquisition
  • Genspark — an AI agent platform orchestrating roughly 70 AI models
  • Hanwha — autonomous data-center energy management
  • Odyssey — interactive world models for simulation and robotics
  • Replit — vibe-coding for full software projects via Agent 4
  • SAS — Unreal Engine digital twins for surgical sterilization
  • Tensor Auto — a self-driving car consumers can own
  • Vexcel — AI that reads the world from aerial imagery
  • Within — "Company Brain" operational context graphs
  • Yutori — web-using AI agents at Meta-scale

What do the revenue numbers show?

Genspark reports the steepest commercial ramp on the list. Once an AI search engine, the company now runs an AI Workspace with a full office suite, image generator, and specialized agents, orchestrated across about 70 AI models. Its persistent assistant, Genspark Claw, handles long-running multi-step workflows asynchronously across browsers, files, and Microsoft Office 365. More than 5,000 enterprises use the platform, and the company says business raced to $250 million in annualized revenue by spring 2026.

Within, formerly Klarity, says it saved its 93 customers — including OpenAI, Salesforce, ServiceNow, and Uber — $53.9 million in consulting fees during 2025. The company has raised more than $91 million from investors including NFDG, Elad Gil, and Y Combinator.

Yutori says customers such as Meta now make more than 100 million API calls per month to its models, with agents interacting with "several hundred thousand" merchant websites daily. Its new Navigator n2 is a 27-billion-parameter computer-use model built to operate full desktop environments. Cofounders Dhruv Batra and Devi Parakh led high-profile AI teams at Meta; other researchers came from leadership roles at Google DeepMind, Tesla, and Apple.

Where is enterprise adoption heading?

Replit says more than 50 million people use its platform and that 85% of Fortune 500 companies now build with it. Its Agent 4 assistant lets non-coders direct software design, build, and shipping in plain language, with hosting, databases, authentication, and monitoring handled by the company and more than 100 integrations with Databricks, Stripe, Slack, and Microsoft.

Canva's AI 2.0, powered by the Canva Design Model, lets users produce and refine campaigns from a few instructions — something Fast Company notes is notoriously hard with general-purpose AIs such as ChatGPT. More than 400 designs are created in Canva AI every second.

Credo AI's numbers point to a governance gap: a Credo study found 60% of organizations now deploy AI across multiple departments or company-wide, but only 4% govern those deployments organization-wide. GAIA, pre-loaded with Credo's governance libraries and frameworks such as the EU AI Act and NIST AI RMF, generates use case descriptions, recommends metadata, drafts questionnaire responses, and maps controls to risks.

What about physical AI and infrastructure?

Several winners target the physical world. Odyssey's Odyssey-2 Max world model earned the highest physics score among publicly evaluated world models on the VBench 2 benchmark — 58.52, up from 49.67 in the prior generation — and counts Jeff Dean, Elad Gil, Garry Tan, and NVIDIA's venture arm among its backers.

Tensor Auto, unlike Waymo's robotaxi model, is building a self-driving car consumers can own, with lidar sensing, an onboard supercomputer that accepts verbal commands, and a Tensor World Model that simulates physical conditions from prompts. The company has won a Level 4 DMV testing certification in California.

Hanwha, the South Korean conglomerate, built an Energy Management System that uses large language models and other machine learning to monitor data-center telemetry, alarms, and fault detection. In April 2026, Prime Group, one of the largest private real estate investors, adopted the technology for a nationwide edge data-center deployment.

In April 2026, DeepL unveiled real-time voice-to-voice translation within its DeepL Voice product. Pioneer, Aramark, Avendra International, and Inetum have all begun using it, DeepL says. The company's acquisition of Mixhalo, which specializes in ultra-low latency audio for live events, aims to enable real-time translation at conferences with thousands of attendees and hundreds of speakers.

SAS is building high-fidelity digital twins of environments such as operating rooms on Epic Games' Unreal Engine, simulating sterilization procedures for patient safety. The system is in use at Sterilcentral, a Danish facility cleaning thousands of surgical instruments. Vexcel's Intelligence system scans satellite imagery for complex descriptions — a set-back Victorian house, a specific crop — and works with Google Maps and disaster-response assessment, including after a tornado in Oklahoma and a fire in Georgia.

The honorees come from Fast Company's Next Big Things in Tech awards for 2026. The common thread across winners: proof of paid adoption — $250 million in annualized revenue at Genspark, 100 million monthly API calls at Yutori, a $145 million round at Alembic — rather than novelty alone.

Original: alembic.com

Share this article:

More from Amara Osei

Amara Osei

Show full bio

Senior reporter covering consumer brands and retail at Business Bearings.

619 articles

Related articles

« Previous articleNext article »