Inside the AI Startups Backed by AWS' US$1m Investment

AWS has chosen 40 AI startups for its Generative AI Accelerator programme, mentoring and granting each up to US$1m in cloud computing credits

The urgency to identify promising AI startups is rising as the challenges to understand, implement and capitalise on AI rise too.

Cloud providers including Amazon, Microsoft and Google are offering millions in credits and technical support to early-stage companies – betting that startups could become tomorrow’s major enterprise customers.

For smaller AI firms, access to computing infrastructure can make the difference between scaling rapidly or burning through venture capital on server costs.

Now, AWS has chosen its yearly AI startups for its third Generative AI Accelerator programme, offering each up to US$1m in cloud computing credits.

The eight-week programme brings together 40 companies working on everything from Arabic language models to AI systems that design new molecules for drug development.

What does AWS’s selection reflect about the AI market?

The selection shows how AI development has splintered into dozens of niches.

For instance, several companies are tackling languages that existing models handle poorly.

Trillion Labs is building models for Koreans, SCB 10X’s Typhoon project focuses on Thai – and Lisan AI is developing tools for Arabic speakers in government and business.

These efforts matter because most large language models (LLMs) are trained primarily on English text, which limits their usefulness in much of the world.

“Whether it’s in biotech labs, creative studios or industrial applications, the pace of Gen AI innovation is extraordinary – and it’s happening everywhere,” says Sherry Karamdashti, General Manager (GM) and Head of Startups in North America at AWS.

Some of the startups advancements in drug discovery and financial automation

Healthcare companies are now using AI for specific problems in drug development rather than broad research – and startups are assisting them.

For example, Chai Discovery trains models to engineer molecules, while Manifold Bio combines AI-driven protein engineering with testing in living organisms.

SyntheticGestalt has also built what it describes as a molecular-focused foundation model, though such claims about model capabilities are common in a sector where startups often promise more than they can immediately deliver.

Meanwhile in the financial sector, financial services selections include Hyperbots, which has developed an agentic AI platform for finance teams.

These systems can take actions autonomously rather than just answering questions.

The company’s HyperLM is a language model (LM) trained on financial data.

Meanwhile, Eloquent AI is working on similar automation for regulated operations, while Synthera AI builds tools for fixed income modelling.

How robotics startups are targeting manufacturing gaps

The robotics startups reveal how AI is being applied to physical tasks that have resisted automation.

RLWRLD is developing foundation models for industrial robots, training them on what it says is high-precision movement data.

Mimic Robotics is creating systems for retail and manufacturing, while Basetwo AI provides tools that analyse pharmaceutical plant data to suggest actions for engineers.

Simultaneously, infrastructure startups are addressing the costs of running AI systems.

All the AI startups AWS has chosen:

Hyperbots — Financial Services

Mary Technology — Legal

Pluralis Research — Software & Internet

RLWRLD — Computers & Electronics

SCB 10X — Software & Internet

SDio — Software & Internet

Smallest AI — Software & Internet

Stimuler — Education

SyntheticGestalt — Life Sciences

Trillion Labs — Software & Internet

Hemispheric (Cognitiv) — Life Sciences

Inephany — Software & Internet

Jentic — Software & Internet

Lettria — Software & Internet

Lisan — Software & Internet

Mimic Robotics — Robotics

Orakl Oncology — Life Sciences

VidLab7 — Software & Internet

AI Cube — Software & Internet

Dharma.ai — Software & Internet

Forlex — Legal

Qomplement — Software & Internet

Synthera AI — Finance

Basetwo AI — Manufacturing

Chai Discovery — Healthcare

Eloquent AI — Finance

Exaforce — Cybersecurity

Hedra Inc. — Media & Entertainment

Inception Labs — Computers & Electronics

Invisible Universe — Media & Entertainment

LlamaIndex — Software & Internet

Manifold Bio — Healthcare

NeuBird — Software & Internet

Nexxa.ai — Software & Internet

Pathway — Professional Services

Ravenna — Software & Internet

Reevo.ai — Software & Internet

Runloop AI — Computers & Electronics

Wondera AI — Media & Entertainment

Inception Labs claims its Mercury system operates 10 times faster and cheaper than current language models, using what it calls a diffusion approach.

Inephany builds optimisation tools to help companies train models more efficiently, which matters when training runs can cost hundreds of thousands of dollars.

Each company gets technical and business mentoring alongside the credits.

Across these startups, AWS’s programme covers ML performance, infrastructure setup and go-to-market strategies.

Participants will present their work at AWS re:Invent in Las Vegas this December, where they’ll meet potential investors and customers.

“This year’s cohort reinforces our mission to help that innovation move faster and deliver real-world impact for customers in every industry,” Sherry says.

“We’re removing the barriers and accelerating opportunities so these leaders can grow their world-changing solutions.”