Three years in the past, Luminal co-founder Joe Fioti was engaged on chip design at Intel when he got here to a realization. Whereas he was engaged on making one of the best chips he may, the extra vital bottleneck was in software program.

“You can also make one of the best {hardware} on earth, but when it’s onerous for builders to make use of, they’re simply not going to make use of it,” he informed me.

Now, he’s began an organization that focuses totally on that downside. On Monday, Luminal introduced $5.3 million in seed funding, in a spherical led by Felicis Ventures with angel funding from Paul Graham, Guillermo Rauch, and Ben Porterfield. 

Fioti’s co-founders, Jake Stevens and Matthew Gunton, come from Apple and Amazon, respectively, and the corporate was a part of Y Combinator’s Summer time 2025 batch.

Luminal’s core enterprise is easy: the corporate sells compute, similar to neo-cloud firms like Coreweave or Lambda Labs. However the place these firms deal with GPUs, Luminal has targeted on optimization strategies that allow the corporate squeeze extra compute out of the infrastructure it has. Particularly, the corporate focuses on optimizing the compiler that sits between written code and the GPU {hardware} — the identical developer programs that precipitated Fioti so many complications in his earlier job.

In the mean time, the business’s main compiler is Nvidia’s CUDA system — an underrated ingredient within the firm’s runaway success. However many components of CUDA are open-source, and Luminal is betting that, with many within the business nonetheless scrambling for GPUs, there can be a number of worth to be gained in constructing out the remainder of the stack.

It’s a part of a rising cohort of inference-optimization startups, which have grown extra useful as firms search for quicker and cheaper methods to run their fashions. Inference suppliers like Baseten and Collectively AI have lengthy specialised in optimization, and smaller firms like Tensormesh and Clarifai at the moment are popping as much as deal with extra particular technical methods.

Luminal and different members of the cohort will face stiff competitors from optimization groups at main labs, which get pleasure from optimizing for a single household of fashions. Working for shoppers, Luminal has to adapt to no matter mannequin comes their manner. However even with the danger of being out-gunned by the hyperscalers, Fioti says the market is rising quick sufficient that he’s not anxious.

“It’s at all times going to be potential to spend six months hand tuning a mannequin structure on a given {hardware}, and also you’re in all probability going to beat any kinds of, any form of compiler efficiency,” Fioti says. “However our large wager is that something wanting that, the all-purpose use case continues to be very economically useful.”