A new venture aiming to provide “universal AI compute” has entered the flourishing artificial intelligence market in France.

With a €28.5 million war fund and an audacious goal to “unlock the future of machine intelligence,” Paris-based FlexAI came out of stealth today. FlexAI is re-architecting computational infrastructure in order to pick that lock.

The flagship product of the firm is an on-demand cloud service. The solution, which is scheduled for release later this year, will let programmers create and train AI applications using heterogeneous computational architectures.

The workload will be abstracted and orchestrated to the best architecture by an integrated software layer. Users can thus use various hardware configurations without requiring intricate code modifications.

This will maximize simplicity and efficiency. FlexAI seeks to increase access to AI by offering simple access to a broad variety of workloads and compute.

The ultimate objective of the firm, according to its founders Brijesh Tripathi and Dali Kilani, is to provide AI solutions for the most pressing issues facing humanity. They believe that more access will spur more inventions.

The CEO of FlexAI, Tripathi, stated, “We believe in the transformative power of AI to solve some of humanity’s biggest problems, but it will require a 1000x magnitude more compute to be able to realise this vision.”

“AI computing is currently only available to a limited number of people. Our goal is to provide everyone with access to computing power.

The Great Compute Problem in AI

Concern over AI’s insatiable thirst for processing power is growing as FlexAI emerges. The developments in the sector are getting more and more costly, unattainable, and unsustainable.

FlexAI promises to have answers for each of these issues. The service reduces compute expenses for companies, which can account for up to 80% of capital. FlexAI may guarantee that users only pay for the compute that is really used by automatically allocating the necessary processing power for a particular task.

Additionally, the strategy would lessen the growing carbon footprint of AI. AI is expected to use up more Indian electricity by 2030, according to the CEO of semiconductor manufacturer Arm. Model complexity and scale will only raise the requirement for higher energy efficiency.

The cloud service also seeks to address the escalating skills gap. The current AI infrastructure needs extremely specialized knowledge to function. FlexAI aims to empower more developers and businesses to create AI apps by developing a software layer that streamlines the process.

Seed funding for these three solutions has already totaled $30 million (€28.5 million). Heartcore Capital, Elaia Partners, and Alpha Intelligence Capital (AIC) led the financing.

The urgent necessity for FlexAI’s service was emphasized by AIC Managing Partner Antoine Blondeau.

“Compute is lagging, not leading, AI models’ increasingly potent capabilities,” he said. “To deliver on AI’s potential, the industry has to solve the issue of compute costs and that of stack complexity — both of which dramatically increase development costs and time to market.”

“Brijesh and Dali, uniquely, have the hardware and software skill sets to architect and build the ecosystem for the AI compute orchestration layer.”

Developing France’s AI Industry

The increasing strength of France’s AI ecosystem was also commended by the investors. Under the leadership of the unicorn Mistral AI, Paris has developed into a hub for artificial intelligence in Europe.

France’s AI industry has “global ambitions,” according to Blondeaue, with a solid scientific foundation, a burgeoning list of company founders, and more access to international financing.

The growth of this industry will also benefit Emmanuel Macron. The goal of the French president to establish “a startup nation” has so far elicited praise and derision. He now has proof that his idea can come true in the booming AI scene.