In the last two decades, social media platforms have become the main medium for informational exchange on a global scale. This has yielded an unprecedented opportunity for corporate actors to make use of this data in various sectors. Among others, finance has been able to exploit such data most efficiently in order to understand retail sentiment, capture market movements, and analyze the spread of financial narratives across the social media landscape.

This report examines the recent SpaceX IPO as a great case study in the finance sector, focusing observational attention on 'X' (formerly Twitter), the social media platform with the largest volume of data on the topic. It has the additional benefit of being text intensive as a medium of informational exchange, which facilitates representative data collection without multi-modal pipelines.

Being the most followed account on X, Elon Musk, with 241.5M followers as of September 2026^[2], has a unique advantage in the case of finding enough demand to carry out an IPO. Whilst other companies have to rely on investment banks and foraging for large buyers before listing, SpaceX was able to comfortably hand off 30% of the initial share allocation to retail buyers^[3]. Where a usual offering would sell only 5-10% of the float to individual traders, Elon's company had the largest IPO in history outdo that figure threefold and have it be oversubscribed by fourfold reaching $250B in orders^[4]; non-professional traders were said to be pulling a book of $70B when the total offering itself was 555,555,555 Class-A shares @ $135 totalling $75B in proceeds ($74.4B net after underwriting discounts)^[3].

Before the social media era, such avid demand for equity in a company would've been either for strong fundamentals or simply social exaltation uncharted by the networks we have today. In this case, we saw a peak valuation of $2.95T on June 16th^[5], a 67% increase from the IPO value, carrying a massive 158x price to sales ratio with a net earnings for FY2025 of -$4.97B^[3].

Although the reported TAM is $28T^[6] and firm believers of the company's potential would argue that these metrics don't do justice to the undervaluation, the significant impact of Elon Musk's persona and social power cannot be understated.

Having identified that a substantial portion of this IPO's value was pulled from Musk's social capital and influence, there can be an instinctive curiosity to find the correlation between price movement and volume when compared to social media metrics like keyword count data or even a sentiment index on the stock. The principal issue with such an approach is the inherent circularity of these data series. On one hand, the current price is a composition of fundamentals, news, shocks and sentiment yet on the other hand, the current sentiment is derived from those same fundamentals, news events, social shocks and of course the movement of the market price. As we have so many unknowns in these formulas, we can't determine the factors of interdependence in the structure of either of these metrics.

This means that the research for alpha in forecasting financial market movements through social network transformations can't be cut short at first-order quantitative data.

Fetching the informational landscape of narratives and sentiment on social media can be quite challenging to the average participants of both the IPO and the network of investors. The topology of this web of interactions has to be inferred from the quantitative data and further processed and analyzed by specialized infrastructure, making it difficult for retail traders to properly model its anatomy.

Indeed, two-thirds of new brokerage accounts opened in 2025 belonged to investors under 45, a cohort entering markets with little independent investment experience and no established track record^[7]. This unsophistication is characteristic of the modern landscape of retail investing; as a result, the potential for large-scale, irrational investing decisions—epitomized in herding behavior—has never been so high. Exposure to environments of potential collective group coordination, namely social media platforms, has significantly increased the probability of extreme herding events. Herding events are particular in their self-reinforcing nature, drawing direct parallels to engagement effects of viral social media content, where the herding itself generates its own subsequent attention. Tools capturing such retail events at their initial phases have become a great source of profitability for sophisticated traders and a main focus of the financial alternative-data market.

Additionally, ordinary motives for retail investment decisions, including, among others, retirement provision, diversification, and risk reduction contrast sharply with a growing population of financially illiterate and social-media-driven retail investors. The wide availability of brokerage fintech, and the increasingly frictionless interfaces built to onboard new users, have greatly expanded the pool of first-time investors: mobile apps now account for around 75% of all retail stock trades globally, and Gen Z and Millennial participants together represent more than 60% of retail trading activity^[7], a cohort contributing roughly $302 billion in 2025, up 53% from $197 billion in 2024^[10]. This population is greatly influenced by attention-induced and return-chasing biases: inexperienced investors without independently developed selection criteria default more heavily to whatever draws attention (Seasholes & Wu, 2007)^[8], and are more prone to chasing recent performance than investors with an established track record of their own (Greenwood & Nagel, 2009)^[9].

SpaceX further illustrates herding behavior, largely social-media-mediated, given the reach of the individual at its center. Holding the most followed account on X, researchers at the University of Pennsylvania's Center on Media, Technology and Democracy have found that Elon Musk's posts carry a baseline view count of millions of impressions, independent of content^[11]. This would merely be a continuation of similar interaction between finance and the same social media environment, with Tesla's ownership base having been characterized in existing market research as primarily young, affluent, and technically engaged, and a documented tendency toward brand loyalty that extends beyond the product into identification with Musk personally; a 2025 NBER paper attributes a measurable share of Tesla's recent sales volatility directly to shifts in public sentiment toward Musk as an individual, separate from the vehicles themselves^[12]. That same audience, already influenced by years of engagement with Musk's public persona across Tesla, SpaceX, and X, represents a population whose collective behavior around SPCX is plausibly less a reaction to the company's fundamentals and more of an extension of a community identity already well established in the data, the topic of the last section.

Sentiment has become one of the main data points to extract on social media platforms as a proxy for investment intent, a shift reflected in institutional adoption: 78% of hedge funds now incorporate alternative data into their process^[13], with social and sentiment data tracked as a standalone category^[14][15], and large funds now spending tens of millions of dollars annually to access it^[16]. Its value increases considerably once it is tied to other social data. Distinguishing communities by their stance on particular narratives, events, or opinion leaders ahead of key market events enriches the profile and deepens the resolution at which a target group may be analyzed and predicted^[18]. Additionally, network topology of target communities, algorithmic clusters, dynamic collective attentional vectors^[17], and many more factors are among the fundamental invisible variables surrounding sentiment, which alone can only capture a limited set of features.

Therefore, forecasting social media movements and isolating their impact on price action proves complex. Most importantly, communities have become entities of their own on social media platforms. Measures of group adhesion, recruitment rate, centrality measures^[19], competition of narratives, segmentation of social groups, and community lifespan^[20] must all be tracked in order to understand how information and opinions propagate through networks. With an appropriate monitoring of the social media environment, prediction of propagating sentiment for a given stock or platform-wide virality, Pythia's purpose, become possible.