It was a rough week across markets — public stocks, private sentiment, and bitcoin alike. This is your reset: a look at the private secondary market. We examine the viability of space data centers and the winners emerging beyond the SaaS downturn.
The SpaceX–xAI Combination and the High-Stakes Bet on Orbital Compute
Elon Musk pushed through the combination of his space company, SpaceX, with cash-hungry xAI on Monday. The deal values xAI at $250 billion and SpaceX at $1 trillion. SpaceX’s chief financial officer told investors the transaction is not expected to delay the company’s planned $50 billion initial public offering.
xAI told investors it burned roughly $9.5 billion of cash during the first nine months of last year, while generating about $210 million in revenue over the same period. SpaceX, by contrast, has recently begun to generate meaningful cash. The company told investors it produced between $1 billion and $2 billion in free cash flow last year, driven largely by the rapid expansion of its Starlink satellite internet business. Executives said revenue climbed to roughly $16 billion in 2025, with about $8 billion in earnings before interest, taxes, depreciation, and amortization.
Musk justified the SpaceX–xAI combination by pointing to what he sees as an impending constraint on Earth-based AI infrastructure. In a blog post announcing the deal, he argued that AI data centers would ultimately need to move off-planet, where power is less constrained. “My estimate is that within two to three years, the lowest-cost way to generate AI compute will be in space,” Musk wrote.
Musk is not alone in exploring the idea. According to a Wall Street Journal report published in December 2025, Sam Altman contacted Stoke Space to explore a similar concept, while Blue Origin, backed by Jeff Bezos, has maintained a team for more than a year working on technologies required for orbital AI data centers.
Google has also examined the concept from a feasibility standpoint. In a research paper published in November 2025, the company outlined “Project Suncatcher,” which explores whether future AI workloads could run on constellations of satellites equipped with specialized accelerators and powered primarily by solar energy. The report concluded that “the core concepts of space-based ML compute are not precluded by fundamental physics or insurmountable economic barriers,” while emphasizing that major engineering challenges remain.
Google said its next milestone will be a learning mission in partnership with Planet, slated to launch two prototype satellites by early 2027. The experiment is designed to test how machine-learning models and TPU hardware operate in space and to validate the use of optical inter-satellite links for distributed workloads.
Separate research has reached similarly cautious conclusions. A study by Ascend, backed by French defense contractor Thales, found that space-based data centers could be economically viable and environmentally feasible, provided carbon emissions from launches fall by roughly an order of magnitude. The project envisions deploying up to a gigawatt of orbital capacity over the next 25 years.
SpaceX, however, does not appear to be thinking on a 25-year timeline.
The company has already sought permission to launch as many as one million satellites intended to function as a massive orbital data-center network, according to filings with the Federal Communications Commission. The proposal would dwarf any satellite deployment previously disclosed by SpaceX or its competitors. According to the European Space Agency, only about 25,000 satellites have ever been launched into orbit in total.
In its filing, SpaceX argued that “by directly harnessing near-constant solar power with little operating or maintenance cost, these satellites will achieve transformative cost and energy efficiency while significantly reducing the environmental impact associated with terrestrial data centers.”
On the Other Side of the “SaaSpocalypse”
This week felt genuinely dramatic for public SaaS markets. Software stocks sold off sharply, with the S&P Software & Services Select Industry Index down 5.79% in a single session and the BVP Nasdaq Emerging Cloud Index falling 6.50% day over day.
Bloomberg labeled the moment a “SaaSpocalypse” — an apocalypse for software-as-a-service stocks — as investors rushed to exit positions amid growing fears that artificial intelligence could structurally undermine large parts of the traditional SaaS model.
The panic, however, did not emerge in a vacuum. One catalyst repeatedly cited by analysts was Claude Code, which many now describe as an inflection point for software development itself. According to research from SemiAnalysis, roughly 4% of public GitHub commits are already being authored by Claude Code. At the current pace, SemiAnalysis projects that AI-generated code could account for more than 20% of daily commits by the end of 2026. As the report puts it bluntly: while the market blinked, AI began consuming software development.
That same analysis went further, publishing an economic model of Anthropic. The model projects Anthropic’s future revenue trajectory and estimates the total addressable markets for AI agents across financial services, legal, consulting, and other professional industries — precisely the domains that have historically supported premium SaaS multiples.
Anthropic, however, is not the only company reshaping the landscape. Alongside the pressure on traditional SaaS, a new cohort of AI enablers is emerging — the “shovels and shields” of the AI transition. These companies are not competing with AI models directly; instead, they enable adoption through orchestration layers, developer tooling, cybersecurity, and identity infrastructure.
Some of the strongest performers of 2025 came from this category. Replit posted growth of 365% during the year, while Vercel grew 181%, underscoring where value is accruing as AI reshapes software creation and deployment.
More hyper-growth companies are likely to surface in 2026 as this transition accelerates. We explore this shift — and its implications for markets, business models, and valuations — in our 2026 outlook in the full report.
Regulation Is Coming to Prediction Markets — and Prediction Markets Are Coming to the LBXpro Index
Regulation is beginning to catch up with prediction markets in the U.S. The U.S. Commodity Futures Trading Commission, which oversees prediction markets, plans to introduce new rules aimed at clarifying what types of contracts these platforms are permitted to offer.
The CFTC intends to withdraw a Biden-era proposal that would have banned political and sports-related contracts — by far the two most popular categories on prediction market platforms. Sports betting remains regulated at the state level, and several states have sued to block prediction markets from offering sports-related contracts, arguing that they constitute illegal gambling under state law.
Following the February reconstruction of the LBXpro 25 Index, Polymarket entered the index as a new holding, driven by elevated secondary trading activity in the fourth quarter of 2025. Polymarket ranked 20th in our activity ranking, with Kalshi close behind at 22nd.
OpenAI Launches Frontier Platform for Enterprise AI Agents
OpenAI unveiled a new platform for business customers to develop, run and manage AI agents that can take actions on behalf of employees. OpenAI lists HP, Intuit and Uber, among others, as customers of its Frontier platform.
The company says Frontier should help businesses overcome some of the key hurdles of implementing AI, such as helping agents get access to the right tools, managing identity and permissioning and incorporating feedback to improve them.
OpenAI is also hiring hundreds of AI consultants, also known as forward-deployed engineers, to support businesses’ adoption of AI.
Fresh Funding Rounds and Liquidity Events
Self-driving ridesharing company Waymo said it raised $16 billion at a $126 billion valuation, with Alphabet remaining the majority shareholder. The round underscores sustained investor appetite for autonomous driving, even as commercialization timelines remain long.
Cerebras Systems raised approximately $1 billion in a new round that values the company at $23 billion post-money. The financing comes as Cerebras is expected to pursue a public listing in the coming months. The company recently announced a major commercial agreement with OpenAI, which committed to purchase 750 megawatts of compute capacity through 2028, a deal estimated at roughly $10 billion.
Anthropic is planning an employee tender offer at a reported $350 billion valuation.
ElevenLabs also raised $500 million in a round led by Sequoia Capital, valuing the company at $11 billion — more than three times higher than its valuation twelve months ago.
Smart-ring maker Oura Health is preparing a tender offer that would allow existing shareholders to sell shares at a roughly 25% discount to its last private valuation. Oura raised about $900 million at an $11 billion valuation in its most recent Series E.
Cyberstarts led a new employee tender offer at Cyera through its $300 million Employee Liquidity Fund.
Enjoyed this article?
Subscribe for private market insights delivered to your inbox.