The company will continue losing money for another five years
Em um momento em que a inteligência artificial redefine o valor das empresas antes mesmo de gerar lucro, a Goldman Sachs apresenta ao mercado uma visão de futuro para a SpaceX que transforma o foguete em coadjuvante: é a divisão de IA, a xAI, que carregaria o peso de uma avaliação de US$ 1,78 trilhão. A projeção — de US$ 3,2 bilhões em receita em 2025 para US$ 322 bilhões em 2030 — não descreve uma empresa que existe hoje, mas uma aposta coletiva em uma supremacia tecnológica ainda por ser conquistada. É o otimismo de Wall Street transformado em prospecto.
- A SpaceX se prepara para um dos maiores IPOs da história, e os números que Goldman Sachs apresenta a investidores exigem um salto de fé quase sem precedentes.
- A divisão xAI acumulou US$ 6,4 bilhões em prejuízo em 2025, mas o modelo do banco projeta uma virada imediata: crescimento de 388% já em 2026, chegando a US$ 15,6 bilhões em receita.
- O negócio que tornou a SpaceX famosa — os foguetes — ficaria com apenas US$ 8,3 bilhões em 2030, enquanto a xAI sozinha superaria em mais do dobro a receita da Starlink.
- Para que as projeções se concretizem, o Grok precisaria superar OpenAI, Google e Anthropic nas frentes mais disputadas da IA: agentes autônomos, programação, cibersegurança e linguagem.
- O fluxo de caixa livre só deve se tornar positivo em 2031, o que significa que os investidores estão sendo convidados a financiar cinco anos de perdas em troca de uma promessa de transformação radical.
A SpaceX caminha para um IPO histórico, e os números que Goldman Sachs circula entre potenciais investidores revelam uma aposta quase inteiramente depositada na inteligência artificial. O banco, coordenador da oferta, trabalha com uma avaliação de US$ 1,78 trilhão — cifra que só se sustenta se a divisão xAI crescer de US$ 3,2 bilhões em receita este ano para US$ 322 bilhões até 2030. As projeções foram apresentadas durante o roadshow do IPO e confirmadas ao Financial Times, como parte do esforço para atrair os gestores de capital que poderiam levantar até US$ 86 bilhões.
O que torna esse cenário revelador é a hierarquia que ele impõe ao negócio. A Starlink, serviço de internet via satélite que já gera receita real, chegaria a US$ 144 bilhões em 2030. Os foguetes — a origem de tudo — contribuiriam com modestos US$ 8,3 bilhões. A xAI, sozinha, superaria ambos combinados. Esse reordenamento reflete o entusiasmo de Wall Street com a IA, onde qualquer ativo com conexão plausível à tecnologia atrai capital em abundância.
O ponto de partida, porém, é de perdas: a xAI registrou prejuízo de US$ 6,4 bilhões em 2025. Ainda assim, o modelo do Goldman projeta uma virada quase imediata, com crescimento de 388% já em 2026. Para que isso aconteça, os modelos Grok precisariam não apenas alcançar, mas superar OpenAI, Google e Anthropic nas áreas mais competitivas da IA — agentes autônomos, programação e cibersegurança.
A realidade financeira presente contrasta com esse horizonte. A SpaceX queimou US$ 13,8 bilhões em fluxo de caixa livre no ano passado, e Goldman não espera que esse número se torne positivo antes de 2031. Os investidores estão sendo convidados a acreditar em um futuro tão distante e tão dependente de avanços tecnológicos ainda incertos que o presente financeiro se torna, por ora, quase irrelevante.
SpaceX is preparing for one of the largest initial public offerings in history, and the numbers Goldman Sachs is circulating to potential investors tell a story of almost unimaginable growth—one that hinges almost entirely on artificial intelligence.
The investment bank coordinating the IPO has modeled a valuation of $1.78 trillion for the company. To justify that figure, Goldman has sketched out a path where SpaceX's AI division—the xAI unit—explodes from $3.2 billion in revenue this year to $322 billion by 2030. That's a hundredfold increase in five years. The projections were presented verbally during the IPO roadshow and later confirmed to the Financial Times, part of Goldman's effort to attract the capital managers whose commitments could raise as much as $86 billion.
The scale of this bet becomes clearer when you look at what xAI is supposed to become relative to SpaceX's other businesses. Starlink, the satellite internet service that has captured public imagination and generated real revenue, is projected to bring in $144 billion by 2030. The rockets themselves—the original business, the thing that made SpaceX famous—are expected to contribute just $8.3 billion. xAI alone would dwarf both of them combined. This reordering reflects the fever pitch of AI enthusiasm on Wall Street right now, where investors are chasing any asset with a plausible connection to the technology.
What makes this projection particularly striking is the starting point. xAI lost $6.4 billion in 2025. The division is bleeding money. Yet Goldman's model shows it pivoting to profitability and explosive growth almost immediately. In 2026 alone, revenue is projected to surge 388 percent to $15.6 billion. By 2027, it would reach $34.5 billion. The bank also forecasts that SpaceX's overall adjusted earnings before interest, taxes, depreciation, and amortization will climb from $6.6 billion this year to $352 billion by 2030—a fifty-fold jump.
For any of this to happen, Elon Musk's Grok family of AI models would need to do something that hasn't been done yet: not just match but decisively beat OpenAI, Google, and Anthropic in the areas that matter most—programming, autonomous agents, cybersecurity, and conversational AI. These are the fronts where the competition is fiercest and where superiority is hardest to establish and maintain.
The current financial reality of SpaceX sits in sharp contrast to these projections. Last year, the company burned through $13.8 billion in free cash flow. Goldman doesn't expect that number to turn positive until 2031, when it projects free cash flow will reach $72 billion. In other words, the company will continue losing money for another five years even as its valuation soars and its AI division supposedly transforms into a revenue machine. Investors are being asked to believe in a future so distant and so dependent on technological breakthroughs that the present financial picture becomes almost irrelevant. That gap between where SpaceX is now and where Goldman says it will be is the entire story of this IPO.
Notable Quotes
xAI would need to not just match but decisively beat OpenAI, Google, and Anthropic in programming, autonomous agents, cybersecurity, and conversational AI— Goldman Sachs projections
The Hearth Conversation Another angle on the story
Why does an AI division matter so much to SpaceX's valuation? Isn't the company supposed to be about rockets and satellites?
Because Wall Street has decided that AI is where all the value is going to be. Starlink is real revenue, real customers, real infrastructure. But it's mature. Rockets are the same. AI is the frontier—it's where the trillion-dollar markets are supposed to be. Goldman is telling investors that xAI is the reason to buy in.
But xAI is losing billions of dollars right now. How does that become a $322 billion revenue business in five years?
That's the bet. The losses are supposed to be investment—building the models, training them, competing with OpenAI. Once Grok becomes the best, the theory goes, the revenue follows. It's a familiar Silicon Valley story, just at a much larger scale.
What would it actually take for Grok to beat OpenAI and Google?
It would need to be demonstrably better at things like writing code, running autonomous tasks, and securing systems. Not just comparable—better. And it would need to do that while those companies are also improving. It's not impossible, but it's not inevitable either.
And if it doesn't happen?
Then the entire valuation collapses. The rockets and satellites can't support a $1.78 trillion company. The AI projections are doing almost all the work.
So investors are really just betting on Musk's ability to build the best AI?
They're betting on that, yes. But they're also betting on the market being so large and so hungry for AI that even if Grok isn't the absolute best, it can still capture enough value to justify the price. That's a different kind of bet—one about market size rather than technological superiority.