The battle to find the next transformative AI chipmaker has begun. And tech giants and venture capitalists all want in.
What’s happening:
- With demand for generative AI hardware and compute still on the uptick, new startups are forming to try to capture some of the opportunity
- Because venture capital continues to pile into AI startups pursuing everything from consumer apps to drug discovery the wait times for GPUs are at an all time high
- As a result, hardware startups and AI chipmakers are raising capital at a rapid pace to try to fill the void
Who is making moves:
- d-Matrix recently raised $110M from Microsoft and other investors after unveiling their compute card they claim has nine times the throughput of Nvidia (NASDAQ: NVDA) GPUs for generative AI
- Further, d-Matrix claims their hardware could consume twenty times less energy for AI compute in comparison to other popular GPUs
- California based startup Cerebras Systems recently built its first ‘AI supercomputer’ which cost over $100M to build and aims to provide an a new alternative for AI generative workloads
- Notably, Cerbras Systems has a partnership with Abu Dhabi based G42 and has plans to bring their supercomputers to the UAE
- BrainChip Holdings (ASX: BRN) recently was awarded a first of its kind patent around their neuromorphic AI intellectual property, which enables their chip to process information in a similar way to the human brain
- BrainChip's flagship AI processor Akida is currently being used by NASA and Mercedes Benz
Why it matters:
- Everyone from Silicon Valley to Wall Street is wondering what the next big breakthrough in AI hardware will be, from GPUs to NPUs to new supercomputers
- With the backlog of Nvidia GPUs expected to continue for the foreseeable future there is a golden window for other hardware producers and chip makers to potentially thrive
- Currently, AI workloads require enormous amounts of computational power and energy. Reducing the amount of electricity and power required for generative AI workloads is an enormous opportunity
- The giants of tech and venture capital are positioning themselves to own some of the underlying infrastructure of AI compute so these mega funding rounds might just be getting started