NVIDIA AI chip rivals raise record funding as competition heats up

This report is from this week’s The Tech Download newsletter. Is it what you see? You can subscribe here.

Nvidia has established itself at the center of the AI ​​boom by dominating the most powerful chips for training and running models, but a growing number of startups are looking to challenge the company’s dominance.

And investors are committing increasingly large sums of money. According to Dealroom, in 2026, AI chip startups raised $8.3 billion in funding worldwide. Barring a near-total market collapse, record amounts are expected to be poured into the sector this year.

So what causes the spikes?

Nvidia’s graphics processing units (GPUs), originally designed for gaming, have been effectively repurposed for AI training, but the focus has now shifted to the most efficient way to actually deploy the technology in applications, known as AI inference.

Here’s what these chip startups claim: Because GPUs were not designed specifically for AI, the new system architecture provides significant energy and cost savings.

“Inference is now mainstream, and existing GPU architectures are not built for inference in the way that matters most at scale,” Patrick Schneider-Sikorsky, director of the Nato Innovation Fund (NIF), which has invested in British AI chip startup Fractile, told me.

Nvidia has a huge advantage as the world’s most valuable company with a nearly limitless supply of cash, but it’s still rushing to develop new chips to power AI.

In December, the company acquired assets from AI inference startup Groq for $20 billion, and in March it announced a $4 billion investment in two companies developing photonics technology.

The semiconductor giant also spent more than $18 billion on research and development in its most recent full year ending January 2026.

But investors aren’t shy about putting money into new AI chip technology that often hasn’t been tested at scale.

In the US, where many of the largest fundings have taken place, Cerebras Systems raised $1 billion in February, and MatX, Ayar Labs and Etched raised $500 million in 2026.

While European companies have raised relatively small amounts of capital, Axela and Orix have both raised more than $200 million this year. Other companies, including Euclyd and Optalysys, have reportedly told me they are planning at least $100 million in rounds in 2026, as well as Fractile and Arago.

“This is no longer a niche bet,” said Carlos Espinal, managing partner at European VC Seedcamp, which backs semiconductor startup Vere Computing. “This is becoming a core part of how people think about AI infrastructure.”

Anthropic and OpenAI have both announced major expansion plans in the UK. Anthropic announced new office space that can accommodate 800 people, and OpenAI announced the opening of its first permanent London office, which can accommodate more than 500 team members.

TSMC on Thursday reported a 58% increase in first-quarter profit, beating expectations and setting a new record as demand for artificial intelligence chips remained strong.

OpenAI has abandoned plans to rent capacity directly from its Norwegian data center, with Microsoft taking over the additional computing, days after confirming it had paused a similar project in the UK. ChatGPT makers will instead rent space from the following sites: microsoftthe company told CNBC.

Amazon announced Tuesday that it will acquire Globalstar in a deal worth about $11.57 billion, aiming to strengthen its fledgling LEO satellite internet business and compete with Elon Musk’s SpaceX.

Uber On Friday, it agreed to buy an additional 4.5% stake in German food delivery company Delivery Hero from the company’s largest shareholder. prosus.

ASML The stock has fallen since Wednesday’s earnings report, even though the company raised its 2026 revenue forecast and beat first-quarter revenue and profit estimates.

Astronomical expectations for the AI ​​boom likely caused a negative reaction, and tightened export controls reduced the percentage of net sales going to China.