Steve Eisman, the investor known for his contrarian bet against the U.S. housing market ahead of the 2008 financial crisis, is expressing caution over SpaceX’s upcoming IPO, arguing that the company’s shift toward AI-linked infrastructure is making its business model significantly more capital-intensive and harder to justify at scale.

Speaking on CNBC, Eisman said he is not planning to short the stock when it begins trading, but raised concerns after reviewing SpaceX’s S-1 filing. His central argument focuses on a sharp rise in capital expenditure relative to revenue, which he sees as a structural change in the company’s financial profile.

He pointed to internal data showing that in fiscal 2023, SpaceX’s capital expenditure was about 42% of revenue. In contrast, in the most recent quarter, that figure reportedly surged to 215%, meaning the company is now spending more than twice its revenue on infrastructure investment.

That shift, according to Eisman, reflects a broader strategic transformation rather than a cyclical spike in spending. He argues that SpaceX is increasingly positioning itself not just as a space and satellite operator, but as an AI infrastructure and compute company tied to large-scale data center and processing ambitions.

Eisman’s critique centers on what he sees as a blurring of SpaceX’s core identity. Traditionally, the company’s valuation narrative has rested on two pillars: launch dominance and the Starlink satellite internet network. The newer AI-driven strategy, he argues, is now becoming the dominant driver of capital allocation and future expectations.

He said the company’s future is being increasingly framed around artificial intelligence rather than its established aerospace and communications businesses.

“The entire company is being bet on AI in terms of its future, not on space and not on Starlink,” Eisman said.

That shift places SpaceX in a more competitive and capital-heavy segment of the technology market, where returns depend on massive upfront investment in compute infrastructure, chips, networking systems, and energy capacity.

Eisman also broadened his critique beyond SpaceX, pointing to what he sees as an industry-wide escalation in capital spending driven by artificial intelligence.

He highlighted recent large-scale investment activity, including Alphabet’s massive funding plans, as evidence that major technology firms are moving from an asset-light model toward an increasingly infrastructure-heavy structure.

In his view, this shift is redefining the economics of the AI sector. While generative AI systems and large language models have demonstrated rapid adoption and technological progress, he argued that the underlying products are becoming increasingly interchangeable.

Despite the enormous spending, he said, differentiation between leading AI systems is narrowing, making long-term competitive moats harder to sustain.

“What’s being produced in terms of LLMs and agentic AI is not really differentiable,” he said. “It’s very commoditized.”

That dynamic, he warned, raises the risk that capital-intensive buildouts may not generate proportional returns if pricing power erodes across the industry.

A key theme in Eisman’s remarks is that capital intensity, rather than growth potential, is becoming the defining risk factor for AI-linked companies.

He noted that the scale of investment required for compute infrastructure, data centers, and advanced semiconductor supply chains has expanded rapidly, reshaping the economics of the sector.

This capital cycle is largely being driven by expectations of long-term AI demand rather than proven near-term profitability, creating what he sees as a widening gap between investment levels and monetization certainty. In that context, SpaceX’s IPO narrative is being interpreted not just as a space technology story but as part of a broader AI infrastructure expansion wave.

Eisman’s comments add a note of caution to what is shaping up to be one of the most closely watched public offerings in recent years.

The SpaceX IPO has already attracted significant attention from major banks and institutional investors, with revenue projections and long-term growth scenarios tied heavily to artificial intelligence integration, satellite connectivity expansion, and large-scale compute services.

However, concerns like Eisman’s highlight a growing divide in how investors are assessing the company: between those focused on long-term technological disruption and those scrutinizing the immediate capital requirements needed to sustain that growth trajectory.

While he stopped short of taking a bearish position, his comments imply that enthusiasm around the listing is being met with increasing scrutiny over whether the financial structure supporting the AI expansion cycle is sustainable at current spending levels.

Coinbase CEO Brian Armstrong has publicly outlined a pragmatic approach to managing AI costs, recommending, among other things, intelligently routing prompts to cheaper models for most tasks while reserving the most advanced systems for high-value work.

In a post on X on Sunday, Armstrong revealed that Coinbase has been actively optimizing its AI usage by directing the majority of workloads to more affordable models, allowing the company to keep overall costs relatively flat even as token consumption grows exponentially.

“We’re working hard on routing prompts to cheaper models where appropriate, and in some cases have been able to keep costs roughly flat, while token usage continues to grow exponentially,” he said.

Armstrong predicted that within 12 to 18 months, roughly 80% of workloads could run on models that are 99% cheaper than today’s frontier systems. He argued that the latest flagship models, such as Anthropic’s Opus 4.8 or OpenAI’s GPT-5.5, should be reserved for “IQ maxing” scenarios, including scientific breakthroughs, complex agent orchestration, and other high-stakes applications where maximum intelligence is essential.

“This leads me to think the limiting factor will be energy and compute, not better models,” he said.

For much of the past year, the prevailing culture, especially in startups, was one of “tokenmaxxing,” where companies proudly showcased massive token usage and rushed to adopt the newest, most powerful models. Y Combinator CEO Garry Tan famously advised founders to “let it rip” with tokens, and many embraced the approach without much regard for cost efficiency.

That era appears to be fading. Armstrong’s post struck a chord with several prominent figures. Venture capitalist Marc Andreessen called it “interesting,” while Hugging Face co-founder Julien Chaumond noted that “model routing is growing a lot these days.” Box CEO Aaron Levie described the 80/99 split as “a bit extreme” but agreed that AI usage would likely stratify, with “high-end” work handled by frontier models and “high-volume” tasks shifted to cheaper alternatives.

Harvey co-founder Winston Weinberg added: “Intelligence allocation is going to be extremely important.”

Glean co-founder Tony Gentilcore was more blunt, saying Armstrong’s view was “spot on” and that “everyone technical already knows this.” He suggested that financial markets were the only ones still extrapolating frontier model prices to infinite scale.

Why the Shift Matters

The move toward smarter routing addresses a growing tension in the AI ecosystem. While frontier models deliver impressive capabilities, they come at significantly higher computational and financial costs. When Anthropic released Opus 4.7, many users quickly hit rate limits and complained about rapidly escalating bills. The same pattern has played out across other providers.

By contrast, intelligent routing allows companies to match the right model to the right task — using lighter, faster, and far cheaper systems for routine queries while deploying top-tier models only when necessary. This approach not only controls costs but can also improve speed and user experience for everyday interactions.

For Coinbase, the strategy has clear business implications. As a major financial platform increasingly incorporating AI features, optimizing spend without sacrificing performance is critical for maintaining margins and scaling services. Armstrong implies the company is treating AI as a core infrastructure investment that must be managed with the same discipline as any other operational expense.

Some in the industry believe that this efficiency focus could reshape how the AI market develops. If more companies follow Coinbase’s lead, demand for mid-tier and specialized models may grow faster than expected, while pressure on frontier model providers could intensify to deliver better performance-per-dollar.

It also highlights a maturing understanding that raw intelligence isn’t the only metric that matters — latency, cost, privacy, and reliability are equally important for real-world deployment.

But the shift also has implications for startups. Some analyst point out that early-stage companies that previously burned capital on token-heavy experimentation may now need to adopt more disciplined approaches earlier in their lifecycle. At the same time, it creates opportunities for new players focused on orchestration, routing layers, and cost-optimization tools.

Armstrong’s post arrives at a moment when investor scrutiny of AI return on investment is increasing. With hyperscalers spending hundreds of billions on infrastructure and many enterprises still searching for clear ROI, the emphasis on efficiency and intelligent allocation feels timely.

In the broader AI ecosystem, this is seen as a healthy evolution. The initial gold-rush phase of throwing compute at every problem is giving way to a more nuanced, economically disciplined approach. As Armstrong and others suggest, the real constraint going forward may not be model intelligence but energy, compute availability, and thoughtful deployment.

Meta Platforms has escalated its long-running legal battle with Israeli spyware company NSO Group, filing a federal court motion for contempt on Monday after allegedly discovering new spear-phishing attempts linked to the controversial firm targeting WhatsApp users.

The social media giant said its messaging service recently disrupted fresh attacks that mirrored previous “1-click” phishing campaigns, in which users are tricked into clicking malicious links that can compromise their devices without requiring passwords or further interaction. Meta said it took down test accounts and groups created by NSO on its platform as part of the response.

“These attempts were similar to previous ‘1-click phishing campaigns’ aimed to trick users into clicking malicious links and direct them to external websites,” the company said in a blog post.

The move comes after a U.S. court issued a permanent injunction last year barring NSO from targeting WhatsApp or its users. While the court significantly reduced the punitive damages NSO owed Meta, from an initial $167 million down to $4 million, the injunction itself was viewed as a major blow to the spyware firm, which has faced widespread accusations of enabling human rights abuses through its Pegasus tool.

Meta’s latest filing accuses NSO of violating that permanent injunction. The company said the new attempts were part of a pattern of behavior it has been monitoring closely. By seeking contempt charges, Meta is signaling it will not tolerate continued efforts to undermine the security of its platforms, even from sophisticated state-linked actors.

NSO Group, which has been blacklisted by the U.S. government for activities “contrary to the national security or foreign policy interests of the United States,” has long maintained that its tools are intended only for legitimate law enforcement and intelligence purposes against terrorists and criminals. However, investigations by Amnesty International, Citizen Lab, and others have repeatedly linked Pegasus to the targeting of journalists, activists, opposition figures, and human rights defenders around the world.

Meta’s action also underscores the ongoing arms race between platform defenders and offensive cyber tools. “1-click” attacks are particularly dangerous because they require minimal user interaction, making them effective even against cautious targets. WhatsApp’s end-to-end encryption provides strong privacy protections, but sophisticated actors continue to probe for weaknesses in the broader ecosystem, including through social engineering and zero-click or one-click exploits.

Last month, Meta was joined by 12 prominent civil rights organizations, security researchers, privacy advocates, and digital rights experts who filed amicus briefs supporting the permanent injunction and opposing NSO’s appeal. This coalition indicates growing concern over the proliferation of commercial spyware and its potential for abuse by both authoritarian and democratic governments.

The legal fight serves multiple purposes for Meta, which includes protecting its users, bolstering the security of one of the world’s largest messaging platforms (with over 2 billion daily users), and sending a strong signal to other surveillance vendors. The company has invested heavily in security, encryption, and threat intelligence in recent years, partly in response to high-profile spyware incidents.

This latest chapter fits into a larger pattern. Tech companies like Meta, Apple, and Google have increasingly clashed with governments and spyware firms over encryption, backdoors, and user privacy. Apple has sued NSO in the past, and WhatsApp itself has taken legal action to protect its users.

These efforts underpin a growing recognition that commercial spyware poses systemic risks to democratic norms, journalist safety, and civil society.

At the same time, law enforcement agencies worldwide argue they need powerful tools to combat serious crime and terrorism in an encrypted world. This makes it difficult to strike the right balance, and cases like Meta vs. NSO often become proxies for larger debates about sovereignty, security, and human rights in the digital age.

As spyware capabilities grow more sophisticated, the cat-and-mouse game between platforms and attackers is likely to intensify, with significant implications for privacy, security, and trust in digital services. Meta’s contempt motion is expected to keep pressure on NSO while the appeal process continues.