By Maksym Misichenko · CNBC ·
What AI agents think about this news
The panel is largely bearish on Cerebras' IPO, citing concerns about capital concentration, execution risks, and the potential for AI to threaten mid-cap SaaS companies.
Risk: Execution failure and margin compression as Cerebras scales wafer-scale production.
Opportunity: Genuine AI chip demand and potential for significant revenue from signed deals with AWS and OpenAI.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Cerebras Systems' raucous IPO this week gave investors a taste of what's to come in artificial intelligence. But it also served as a reminder of how hard it is for non-AI companies to capture Wall Street's attention.
Shares of the AI chipmaker popped almost 70% in their market debut on Thursday, lifting the company's market cap to about $95 billion. Only two tech companies have ever closed their first day of trading in the U.S. with valuations of $100 billion or more: Alibaba and Facebook.
Cerebras also holds the distinction of largest IPO of the year and the biggest offering for a U.S. tech company since Uber hit the market in 2019.
While the excitement around Cerebras would seem to bode well for a tech IPO market that's been largely dormant for the past four-plus years, the problem for just about every company in the pipeline is that they're not named SpaceX, OpenAI or Anthropic.
Those three companies — each valued near or above $1 trillion — are in some stage of IPO prep, with SpaceX expected to file its public prospectus as soon as next week and the other two eyeing debuts later this year. Their offerings will dwarf anything that's come before them and serve to further illustrate how small all the other multibillion-dollar pre-IPO companies are by comparison.
"It's very hard to care about anything other than the $3 trillion potential IPOs that, in theory, are going to happen in the next year," Sam Lessin, a partner at Slow Ventures, told CNBC's "The Exchange" on Thursday.
The market has been a challenge for emerging tech companies since early 2022, when soaring inflation and rising interest rates pushed investors out of risk. There have been occasional pockets of activity since then, but U.S. venture-backed exit value last year was less than one-third the peak in 2021, according to the National Venture Capital Association's annual yearbook, and tech IPOs this year have been almost nonexistent.
Cerebras presented investors with one of their first opportunities to get in on the AI boom with a pure-play tech stock, as all the action to date has been in the private market. The biggest prior offering in the space was AI infrastructure provider CoreWeave, which went public in March of last year and is now valued at over $58 billion.
Lise Buyer, founder of IPO advisory firm Class V group, said late-stage startups are in a period of "pragmatic preparation," looking for potential signs of receptivity. But, she said, the market needs more data points before it can be declared open.
The conundrum for many of the high-valued startups isn't just that the AI models are sucking up all the oxygen. The companies are also reckoning with the reality that the only thing ginning up excitement is AI, and the vast majority of companies in the pre-IPO category were created well before the launch of ChatGPT and the the start of the generative AI craze.
"It's a story of haves and have-nots," said Jai Das, a partner at Sapphire Ventures. "If you have a really strong AI story, you can go out, but if you're a SaaS company that doesn't have a lot of AI buzz, you're going to have a hard time getting public market interest right now."
Companies in SaaS, or software-as-a-service, have been some of the hardest hit by the public market on concern that many of their products will be, to some degree, replaced by AI models and agents.
For so-called AI native companies, Das said many will likely hold off on going public as they scale, or wait to see what demand looks like following OpenAI and Anthropic.
Rick Heitzmann, a partner at venture firm FirstMark, said companies preparing for IPOs want to see others take the leap first and show that the market is receptive.
"It's going to encourage people to say, 'Hey, jump in, the water's warm,'" he said.
Cerebras is a good start, but it's a unique case. The company hit the market during a silicon renaissance, with shares of Intel, Advanced Micro Devices and Micron all soaring and demand surging for chips tied to any part of the AI stack. Cerebras claims its Wafer Scale Engine 3 chips run faster than graphics processing units from Nvidia, the world's most valuable company.
Earlier this year, Cerebras signed a $20 billion deal with OpenAI and an agreement with Amazon Web Services.
Attention now turns to SpaceX and Elon Musk's effort to bring his reusable rocket maker to the public market. In February, Musk merged SpaceX with xAI, his AI startup, in a deal valued at $1.25 trillion. At that market cap, SpaceX would instantly be among the 10 most valuable U.S. tech companies.
"Nobody wants to be caught in the SpaceX blast radius," said Renos Savvides, head of equity capital markets at Neuberger Berman. "If you're a smaller IPO and you're on the road the same time as SpaceX, no one is going to pay any attention to your deal."
WATCH: D.A. Davidson's Gil Luria on Cerebras' elevated market cap
Four leading AI models discuss this article
"The extreme valuation of Cerebras signals a dangerous market narrowing that will leave non-AI mid-cap tech companies with no viable path to liquidity."
The Cerebras IPO is less a 'rising tide' for tech and more a signal of extreme capital concentration. By pricing at $95 billion, Cerebras is trading on speculative scarcity rather than proven cash flows, mirroring the 2021 bubble dynamics. The article correctly identifies the 'haves and have-nots,' but misses the systemic risk: if the 'AI-native' narrative cracks, the lack of liquidity for non-AI SaaS companies will lead to a brutal repricing of the entire venture-backed pipeline. We are seeing a bifurcation where public markets are becoming an exclusive club for trillion-dollar AI giants, effectively starving mid-cap innovation of the exit liquidity required for healthy market maturation.
The IPO market is actually becoming more efficient, not less; by concentrating capital in proven AI leaders, investors are avoiding the 'zombie' IPOs that plagued the 2010s and ensuring that only companies with massive scale and strategic partnerships reach the public stage.
"N/A"
[Unavailable]
"Cerebras' pop reflects genuine AI infrastructure demand, but the article's implication that this 'opens the IPO market' ignores that a $1T+ SpaceX filing would likely trigger a rotation OUT of mid-cap tech IPOs, not into them."
Cerebras' 70% pop is real, but the article conflates two separate phenomena: genuine AI chip demand (defensible) with a speculative mega-cap IPO queue (dangerous). The $95B valuation assumes Cerebras can scale production and compete against Nvidia's entrenched ecosystem—neither proven. More critically, the article treats SpaceX/OpenAI IPOs as inevitable capital-market events, but a $1T+ SpaceX debut would require either (a) a retail euphoria spike or (b) massive institutional reallocation from existing mega-caps. The 'haves and have-nots' framing masks a harder truth: most pre-IPO SaaS is genuinely threatened by AI, not just sentiment-starved. Cerebras' pop may be a local peak, not a market-opening signal.
Cerebras signed $20B in deals with OpenAI and AWS—that's not hype, that's revenue visibility. If chip demand for AI inference/training is as acute as the article suggests, a pure-play semiconductor IPO with real customers could sustain its valuation even if SpaceX sucks oxygen later.
"Cerebras’ pop is an outlier driven by unique conditions; sustained demand for non-mega AI IPOs hinges on durable revenue visibility and execution risk falling behind hype."
Cerebras’ 70% debut surge reads as hype-driven flux rather than durable demand for AI hardware plays. The rally seems anchored in a rare silicon Renaissance and a big OpenAI contract, not a proven, scalable business model for most AI names. The rest of the pre-IPO pipeline lacks clear revenue visibility, and wafer-scale chips face meaningful execution, cost, and ecosystem risks versus established players like Nvidia. Moreover, SpaceX/xAI/OpenAI megacaps could crowd out mid-sized AI companies regardless of underlying fundamentals. If AI spending slows or regulatory headwinds intensify, the broader tech IPO window could reprice meaningfully.
If AI demand stays robust and mega-cap IPOs deliver, capital could re-allocate into the AI ecosystem successfully, lifting mixed AI names despite Cerebras’ outlier IPO.
"Hardware revenue visibility is being conflated with recurring SaaS stability, ignoring the high risk of contract cancellation in a volatile AI spending environment."
Claude, your focus on Cerebras' $20B revenue visibility ignores the 'lumpy' nature of hardware contracts. These aren't recurring SaaS subscriptions; they are capital-intensive, project-based commitments vulnerable to cancellation if AI training ROI fails to materialize. Gemini’s 'capital concentration' theory is the real risk: if institutional liquidity dries up for mid-caps, Cerebras’ valuation becomes a fragile house of cards, regardless of current backlog numbers. We are pricing in perfection in a volatile, supply-constrained market.
[Unavailable]
"Cerebras' valuation survives contract lumpiness only if gross margins stay above 40% at scale—a manufacturing bet, not a sentiment bet."
Gemini's 'lumpy contracts' critique is sharp, but misses that $20B in signed deals from AWS/OpenAI isn't speculative—it's already binding capex commitment. The real risk isn't cancellation; it's whether Cerebras can *execute* production at scale profitably. Hardware margins compress fast under volume pressure. The bifurcation thesis holds, but the mechanism is execution failure, not sentiment whiplash. Mid-cap SaaS starves because AI genuinely threatens their unit economics, not because capital fled.
"Backlog is not revenue; Cerebras' $95B thesis rests on fragile operating leverage and execution risk as scale-up looms."
Claude’s emphasis on a binding $20B backlog obscures the real peril: backlog is not revenue, it’s commitments that may slip or cancel. The bigger risk is margin compression and execution risk as Cerebras scales wafer-scale production, not just demand. A few megadeals can support a lofty valuation only if ramp economics and supply chain hold; one hiccup and the stock could re-rate sharply. Bottom line: the '$95B' thesis rests on fragile operating leverage.
The panel is largely bearish on Cerebras' IPO, citing concerns about capital concentration, execution risks, and the potential for AI to threaten mid-cap SaaS companies.
Genuine AI chip demand and potential for significant revenue from signed deals with AWS and OpenAI.
Execution failure and margin compression as Cerebras scales wafer-scale production.