Positron challenges Nvidia: $230 million to dethrone the king

Watching Nvidia conquer the world has become even a bit tedious. Jensen Huang has turned his company into a synonym for artificial intelligence, and his chips into the hardest currency of our time. However, monopolies rarely last forever, especially when truly big money and geopolitical ambitions enter the game. The Positron startup has just announced the closing of a Series B funding round for an impressive $230 million. This is not just another financing news story, but a direct challenge to the established order of things in the world of semiconductors.

The main investor is Qatar's sovereign wealth fund (Qatar Investment Authority). This is a critically important marker. While Silicon Valley debates algorithm safety and ethics, the Middle East is building physical infrastructure. Qatar clearly doesn't want to simply rent computing capacity from Microsoft or Google; it wants to own the technology that creates those capabilities. The investment in Positron is an attempt to jump aboard the departing train of semiconductor sovereignty before the ticket price climbs from millions to billions.

Why is everyone so obsessed with finding a Nvidia alternative? The answer is simple: scarcity, price, and architectural inertia. Companies queue for months to obtain coveted H100 accelerators, often overpaying by multiples. Moreover, the GPU architecture that Nvidia has refined over decades for graphics and gaming is not ideal for the specific tasks of training large language models. It contains much excess that consumes power and occupies precious space on the die. Positron promises something different—an architecture purpose-built for neural networks from the ground up, without the heavy baggage of last-generation graphics cards.

We have already seen many ambitious 'Nvidia killers' in the past five years. Remember British Graphcore or Israeli startup Habana Labs? The former encountered catastrophic financial difficulties and software problems, while the latter was swallowed by Intel and effectively dissolved within its unwieldy corporate structure. The main problem for newcomers has never been the hardware itself, but the software ecosystem. CUDA—Nvidia's software layer—has become a de facto industry standard. To force developers to switch to Positron, the startup needs to offer not just a faster chip, but a magic button that allows migrating all existing code without pain and years of rewriting.

The $230 million sum looks enormous for any software startup, but in the semiconductor world it is merely a modest entrance fee. Developing a single modern processor on a 3nm or 5nm process node requires colossal spending on design and booking manufacturing capacity at TSMC facilities. Positron will have to play very carefully to avoid burning through these funds in the furnaces of Taiwan's factories before presenting the market with even a first batch of working prototypes. In this industry, there is virtually no margin for error: either you demonstrate multiplicative efficiency gains over Nvidia, or you become another entry in the annals of expensive failures.

Nevertheless, the emergence of such players is necessary for market health. Nvidia's monopoly makes AI development prohibitively expensive for all but the top five technology giants. If Positron can deliver on even half of its promises, it will create the necessary pressure on prices and accelerate progress. Qatari capital here serves as a catalyst that can transform engineers' theoretical calculations into actual silicon. For us, it remains to watch whether the startup has the stamina for this marathon.

The key point: The market desperately seeks a 'Plan B', and Qatar is willing to bankroll it generously. Whether Positron can become a real threat to Jensen Huang's empire or will swell the list of ambitious but forgotten projects depends on their ability to overcome not just Nvidia's hardware, but its software.