AI chip challengers are attracting record investor money in 2026, and SambaNova’s fresh $1 billion raise at an $11 billion valuation is the clearest sign yet that the market for alternatives to Nvidia is maturing fast. For Pakistani IT firms and AI startups building on cloud infrastructure, this shift in the global hardware landscape deserves close attention.
What SambaNova Just Did and Why It Matters
SambaNova Systems raised $1 billion at an $11 billion valuation led by General Atlantic, in a first close of its Series F round, with more investors expected to join soon. The speed of the fundraising is striking. The latest round comes roughly five months after the Palo Alto-based startup unveiled its SN50 chip alongside a $350 million Series E in February.
A company valued at around $2 billion earlier in the year is now, on paper, worth more than five times that, a swing driven by investors racing to back any credible alternative to Nvidia.
The financing was led by General Atlantic along with participation from Seligman Ventures, T. Rowe Price, and Capital Group. Other new or existing investors include A&E Investment, Assam Ventures, Battery Ventures, BlackRock, Cambium Capital, Intel Capital, QFO Capital, Qatar Investment Authority, Vista Equity Partners, and Volantis.
Who Is SambaNova and What Does It Build?
Founded in 2017 by Rodrigo Liang, Stanford professor Kunle Olukotun, and Christopher Ré, SambaNova builds custom chips, hardware systems, and cloud services designed specifically for AI inference, which is the process of running trained AI models to generate responses.
SambaNova builds chips and systems designed to run large AI models efficiently, using an architecture it calls a Reconfigurable Dataflow Unit rather than the graphics-derived design Nvidia popularised. Its flagship products are the SN40L and SN50 processors. The SN50 is described as delivering more than three times as much throughput as Nvidia’s B200 graphics card, with a top speed described as being five times faster.
Importantly, SambaNova does not position itself as a pure Nvidia killer. SambaNova’s chips are designed to be used alongside Nvidia products, not replace them. The SN40 and SN50 chips can run the decode portion of inference five to ten times faster, which helps free up Nvidia chips for other tasks such as training.
JPMorgan, Saudi Aramco and the Enterprise Push
Alongside the new funding, SambaNova said it has been selected by JPMorgan Chase as an ‘inference-infrastructure partner,’ with its SN40L and SN50 systems set to power secure, on-premises AI inference at the bank. In addition to JPMorgan, it also names Saudi Aramco, Intel, and other Japanese firms as customers.
For banks and other industries where data is incredibly important, bringing AI infrastructure on premises, with models under the company’s own control and private data within its own firewalls, is a key aspect of running AI in a secure and private way. This on-premise model is also what makes SambaNova interesting for governments and large institutions in developing markets who are wary of sending sensitive data to foreign cloud servers.
AI Chip Challengers Are Now a Whole Wave, Not Just One Startup
SambaNova is not alone. SambaNova competes not only with Nvidia but with a growing cohort of well-funded startups including Groq, Cerebras, and d-Matrix, as well as in-house chip efforts from cloud providers like Google, Amazon, and Microsoft.
Nvidia has cemented itself at the heart of the AI boom with a near monopoly on the most powerful chips to train and run models, but a growing crop of startups are set on challenging its supremacy, and increasingly investors are throwing huge sums behind them. In 2026 alone, AI chip startups raised $8.3 billion in funding globally, according to Dealroom.
So far in 2026, investors have already funnelled more than $200 million into the Netherlands’ Axelera and the UK’s Olix. European, Asian, and American AI chip challengers are all raising at the same time, pointing to a structural shift rather than a one-off bet.
The core argument behind all these bets is simple. GPUs were not purpose-designed for AI, and therefore novel chip architectures will bring big savings in energy and cost. As AI moves from training big models to actually running them at scale for users, inference-focused chips gain a real performance edge.
The Nvidia Picture Is Still Huge, but Cracks Are Forming
Nvidia, which trades at a $4.8 trillion market cap, remains the dominant force in AI chips with roughly 80% market share in data centre accelerators. That is an enormous lead. Nvidia still dominates AI training and much of inference, and its customers, cloud providers and model labs alike, have every incentive to fund a second source.
That last point is the real driver. Big cloud companies do not want to depend on a single supplier. The more AI workloads grow, the more they need a backup, and that need creates the market that AI chip challengers are rushing to fill.
SambaNova’s CEO Rodrigo Liang has said the company is strongly considering an IPO in 2027, most likely in the US. SambaNova rival Cerebras Systems, which also designs inference-focused chips, climbed 68% following its initial public offering earlier this year.
Why Pakistani Tech Firms Should Pay Attention
Pakistan’s IT sector is growing fast. The government has frozen the IT export tax rate through 2029 to keep the sector competitive, as covered in our earlier report on Pakistan’s IT export tax freeze. As local software houses scale up and AI-powered products become part of their service offerings, the cost and availability of AI compute will matter more and more.
Right now, most Pakistani developers and startups using AI rely on US-based cloud providers that run on Nvidia hardware. That is fine today, but it comes with risks: high costs, US export policy changes, and no local data control. SambaNova’s on-premise model and its explicit focus on sovereign clouds, where governments fund local partners to build private clouds, is directly relevant to this challenge. If regional cloud providers or government-backed data centres in Pakistan one day adopt alternative AI hardware, costs for local developers could fall significantly.
There is also a broader signal here for Pakistani AI startup founders who track global venture trends. The AI inference chip market has become one of the most intensely funded corners of the semiconductor industry, as enterprises shift spending from model training to deploying AI at scale. Understanding where the hardware layer is heading helps founders make better decisions about which platforms to build on. The DeepSeek inference chip story we covered earlier, in our article about DeepSeek building its own inference chip, is another piece of the same puzzle: the AI hardware race is widening fast.
Frequently Asked Questions
What is SambaNova and what does it make?
SambaNova Systems is a US-based AI chip company founded in 2017. It designs custom processors called Reconfigurable Dataflow Units, built specifically to run AI inference workloads. Its SN40L and SN50 chips are its main products and are sold as full rack systems to enterprises and governments.
How did SambaNova reach an $11 billion valuation so quickly?
SambaNova’s most recent prior round, a $350 million raise, closed only in February, making the leap in valuation since then very sharp. Rounds that used to be spaced a year or more apart are now landing within months as demand for compute outpaces supply.
Are AI chip challengers actually a threat to Nvidia?
Not in the near term on overall market share, but they are carving out a real niche. The scale of SambaNova’s raise and the calibre of its investor base signals growing institutional conviction that the inference market is large enough to support meaningful alternatives. Inference, not training, is where most of the growth is now happening.
Does this affect how Pakistani developers access AI compute?
Not directly today, but the trend matters over the medium term. More AI chip challengers in the market means more competition, which can push down the cost of cloud AI services globally. If sovereign cloud projects gain traction in the region, local developers could eventually access cheaper, locally-hosted AI compute built on these alternative chips.
