On May 14, Cerebras Systems debuted on the Nasdaq under the symbol CBRS and its shares rose more than 100% on the first day of trading. With a valuation that reached USD 95 billion at close, it became the first big AI sector IPO of this cycle and an unmistakable sign that the financial market is betting heavily on companies that build AI infrastructure.

What does Cerebras do, why does it matter, and what does this valuation say about the state of the sector? Here we explain it.

What is Cerebras Systems?

Cerebras is a semiconductor startup founded in 2016 in Silicon Valley with a very specific goal: to build chips designed from the ground up to train and infer large-scale artificial intelligence models. Its flagship product is the WSE-3 (Wafer Scale Engine 3), the chip with the largest silicon surface ever manufactured in the industry.

Unlike Nvidia's GPUs, which are relatively small chips that connect in server racks, Cerebras's WSE-3 occupies a complete silicon wafer and has more than 4 trillion transistors. This makes it possible to train AI models much faster and with lower inference latency, especially for large-scale language models.

Why does it challenge Nvidia?

Nvidia dominates the AI chip market with more than 80% share in GPUs for data centers. But it has an Achilles' heel in a specific segment: inference latency. When you want a model to respond in real time to millions of simultaneous users, the network-connected GPU architecture generates bottlenecks that are hard to eliminate.

Cerebras attacks exactly that problem. Its chips enable inference up to 20 times faster than Nvidia's GPUs in certain benchmarks, according to data published by the company in its SEC filings. This makes them especially attractive for real-time AI applications: medical assistants, high-frequency trading systems, enterprise chatbots with high demand for simultaneous users.

According to TechCrunch, the success of the Nasdaq debut reflects investors' confidence that Nvidia's monopoly in AI infrastructure has real competition, and that Cerebras is one of the few players with differentiated technology to dispute it credibly.

The IPO: numbers and context

Cerebras raised USD 5.5 billion in its initial public offering, with shares initially valued at USD 37. At the close of the first day of trading, the price had risen more than 108%, bringing the market capitalization close to USD 95 billion.

To put it in perspective: that's the strongest stock market debut of a semiconductor company in more than a decade, and the first concrete sign that the IPOs of AI infrastructure companies that had been postponed since 2022 are ready to hit the market en masse.

As The Wall Street Journal reported, institutional investors oversubscribed the offering more than 10 times, which indicates an enormous market appetite for companies that build the hardware layer of AI, not just the models.

What does this mean for the global tech market?

The race for AI infrastructure consolidates

During 2023 and 2024, the market debate revolved around AI models (OpenAI, Anthropic, Google DeepMind). In 2025 and 2026, attention is moving toward who builds the hardware that runs them. Nvidia, AMD, Intel, Google with its TPUs, and now Cerebras are the central players in that dispute over control of the infrastructure layer.

AI IPOs return to the global agenda

Cerebras's debut may be the signal other AI companies were waiting for to go to market. Companies like Anthropic, Elon Musk's xAI, or Perplexity are on analysts' radar as possible IPOs of the 2026-2027 cycle. When the market absorbs an IPO of this magnitude well, it opens the window for the ones that follow.

The impact on LATAM

For tech companies and professionals in Argentina, Mexico, and Brazil, what happens with global AI infrastructure has a direct impact: it defines which AI services will be accessible, at what cost, and with what latency in the region. More competition in the chip market benefits all AI users, because it lowers prices and improves service quality.

If you want to understand how large corporations are integrating these technologies into their core operations, the article about how JPMorgan Chase turned AI into central infrastructure with USD 19.8 billion shows exactly that dynamic from the global financial sector.

Why does it matter for those studying AI in LATAM?

The boom in AI infrastructure investment has a direct consequence for the job market: more companies building AI systems means more demand for professionals who know how to work with these models, integrate them into real products, and design AI-based solutions for traditional industries.

If you want to understand the ecosystem from the inside and prepare for that growing market, these Coderhouse courses are the starting point:

• Introduction to Artificial Intelligence Course: to understand how language models work, what inference chips are, and why hardware matters in the AI ecosystem.

• AI Engineering Course: for those who want to go deeper and learn to implement AI solutions in production with real architectures.

• AI Agents Course: to learn to build autonomous agents that use these models to solve complex tasks independently in real environments.

Frequently asked questions

What is an IPO and why does Cerebras's matter?

An IPO (Initial Public Offering) is when a private company sells shares to the general public on a stock exchange for the first time. Cerebras's matters because it's the first large-scale IPO of the AI infrastructure sector, and the market's reception indicates how much confidence investors have in the future of these technologies in the long term.

Can Cerebras really compete with Nvidia?

In specific niches, it's already doing so. Cerebras doesn't seek to displace Nvidia in the entire GPU market, but to compete in the high-speed inference segment, where its wafer-scale architecture has real and documented technical advantages. The AI chip market is large enough to have several winners with differentiated propositions.

What is AI inference and why is it important?

The training of an AI model happens once and takes weeks or months. Inference is what happens every time the model responds to a question or generates content: it happens millions of times a day in production. Optimizing inference is key to reducing operating costs and improving the response speed of AI systems in the real world.

How does this type of investment affect the LATAM tech ecosystem?

When there's more investment and competition in AI hardware, the prices of access to powerful models drop and latency improves. For companies and developers in LATAM that use AI APIs, this translates into faster and more accessible services. It also implies more job opportunities in roles that integrate these technologies into local and regional products.

Is it worth investing in shares of AI companies like Cerebras?

That depends entirely on your risk profile and investment horizon. The shares of high-growth tech companies are very volatile: they can rise a lot and also fall significantly. Before making any investment decision in this sector, it's advisable to consult a qualified financial advisor and analyze the company's complete prospectus.