A Grand AI Narrative Meets Stark Usage Numbers
SpaceX’s planned IPO leans heavily on a sweeping artificial intelligence narrative. After merging with xAI, the company pitches its Grok chatbot and broader AI stack as a core growth engine, part of an “innovation engine” meant to power everything from scientific discovery to future enterprise applications. The prospectus frames AI for large institutions as a market worth USD 26.5 trillion (approx. RM122.0 trillion), helping justify a targeted USD 1.75 trillion (approx. RM8.1 trillion) SpaceX IPO valuation and an offering size of USD 75 billion (approx. RM345.0 billion). Yet those ambitions collide with sobering data from the federal AI inventory. Across more than 400 documented deployments naming a vendor, Grok appears just three times. While SpaceX promotes an AI lab that it says will transform multiple industries, the government’s actual usage suggests Grok is, so far, a marginal player in real-world enterprise AI adoption.
Federal AI Adoption Data Shows Clear Preference for Rivals
The latest consolidated AI inventory paints a lopsided picture of federal AI adoption. OpenAI models, including ChatGPT, Codex and Microsoft Copilot, are named in 234 deployments. Alphabet’s Gemini family accounts for 33 listings, while Anthropic’s Claude appears 26 times, despite later political headwinds. In sharp contrast, xAI’s Grok is present in only three cases, a signal that agencies are overwhelmingly choosing alternatives. These figures emerge from records assembled by the federal budget office and corroborated through interviews with civil servants and contracting specialists. Crucially, agencies can access most major AI vendors through existing procurement frameworks, so Grok is not locked out by process. Instead, the data suggests that when government buyers weigh performance, trust, security and integration, they gravitate toward established enterprise AI competitors rather than newer entrants like Grok, underscoring the importance of proven track records in sensitive institutional environments.
Price Isn’t the Problem, So What Is?
Grok’s struggle inside government is not easily blamed on cost. Federal agencies have been able to access the chatbot for eight months at just 42 cents per agency, a symbolic fee aligned with industry practice of seeding low-cost access to encourage habitual use before larger contracts follow. Competitors similarly discount early usage, yet they have achieved vastly higher adoption. That disparity points toward deeper concerns about Grok’s suitability for high-stakes institutional work. One enterprise AI CEO argues that the weak federal traction hints at gaps in security rigor, a critical criterion for public-sector deployments. Others note that xAI has lost its non-Musk founding team and that the broader SpaceX conglomerate is juggling rockets, satellite communications, a social platform and AI—all while burning cash in its AI arm. For risk-averse government buyers, these operational signals may reinforce caution, driving them toward vendors perceived as more stable and battle-tested.
Valuation Hype vs. Enterprise AI Reality
The disconnect between Grok’s limited federal use and SpaceX’s AI promises feeds a broader debate about AI market skepticism. SpaceX is effectively asking public investors to underwrite an IPO at a valuation approaching USD 1.75 trillion (approx. RM8.1 trillion) on the expectation that AI for large enterprises will eventually dwarf its existing businesses. Yet its AI unit currently “burns USD 2 (approx. RM9) for every dollar it brings in,” and the company reports only one recently profitable year, at USD 791 million (approx. RM3.6 billion). Without tangible validation from major institutional customers, those projections look more aspirational than inevitable. Critics warn that, absent stronger evidence of adoption beyond early enthusiasts, SpaceX’s AI narrative risks echoing earlier tech hype cycles. The federal numbers may thus serve as an early stress test: if Grok cannot gain traction in a massive, price-insensitive buyer, investors may question how quickly it can conquer the broader enterprise AI market.