Google has a massive $300 billion question hanging over its lofty AI ambitions, and Bank of America thinks investors need to pay attention. That’s roughly the amount of money the bank expects Alphabet (GOOGL) to spend on capital expenditures in 2027, building data centers and AI chips. Take something as familiar as searching for a restaurant. Over the past several years, Google has steered you toward reviews, menus, and reservation pages. Now, its AI tools can pull together answers and recommendations before you ever click through. Though that feels like a small convenience. Behind the scenes, though, providing those answers entails costly chips and data centers. And as people’s search habits change, Google needs to find ways to keep its lucrative advertising business growing. That’s quite a balancing act, which involves paying for the future without weakening the business financing it. That said, in a note shared with me, Bank of America remains bullish, maintaining a Buy rating and $430 price target. However, the company’s CapEx forecast raises a big question for shareholders: whether Google’s growing AI and cloud businesses can generate enough cash to justify the bill. Bank of America expects Alphabet’s 2027 capital expenditures to approach $300 billion. Michael M. Santiago / Getty Images Bank of America sees Google’s cloud growth masking a costly problem On the surface, it’s easy to see why Google will continue on an upward trajectory. More Google: Google is making a change in November that might catch Gemini users off guard Google set to challenge Meta Muse in the Agentic AI race Google just disclosed something troubling about its AI Alphabet brought in $119.8 billion in Q2 sales, up 24% from a year earlier. Google Cloud, once a smaller piece of the business, grew 82% to $24.8 billion, an extraordinary pace for a company of Google’s size. Post expects the momentum to continue when Alphabet reports on October 28. His Q3 forecast calls for $111.4 billion in sales after traffic-acquisition costs and $3.07 in GAAP earnings per share, ahead of Wall Street’s comparable estimates of $111 billion and $2.98. That said, the growth isn’t evenly spread across Google. Bank of America expects Search revenue growth to cool from 17% in the previous quarter to 14%, while Cloud accelerates from 82% to 93%. Google’s traditional cash-generating business is growing much more slowly, even as its AI infrastructure operations expand rapidly. The more surprising finding is BofA’s cash-flow forecast. The bank expects negative free cash flow of $12.4 billion in Q3, compared with about $3.9 billion expected by other analysts. And it forecasts around $300 billion in capital spending in 2027 alone. That’s a huge financial balancing act. Google can continue to deliver strong sales, even blow past earnings estimates, while pouring more cash into the data centers powering its AI ambitions. So why is a business this profitable consuming so much cash? Google’s AI ambitions could consume $500 billion in just two years BofA expects Google’s AI ambitions to come with a massive bill. It estimates Alphabet will spend about $199 billion in CapEx in 2026 and another $301 billion in 2027. That’s roughly half a trillion dollars in just two years, much of it tied to data centers, computing chips, and the infrastructure needed to handle rising AI demand. But Google isn’t simply buying more servers and hoping customers show up. It’s building its own Tensor Processing Units (TPUs), expanding cloud capacity, and selling computing power and AI services to other businesses. Google Cloud CEO Thomas Kurian offered a sense of that demand at a September technology conference. He said over 300 customers had each committed over $100 million, while customers were spending, on average, more than 50% above their contractual commitments. Kurian also said Google’s AI servers pay for themselves in under two years, with its own TPUs expected to pay back faster than conventional graphics processors. Nevertheless, there are a few caveats to consider. Bank of America notes that an index tracking AI-model usage costs fell 33% in Q3. Cheaper AI might attract more customers, but Google needs enough additional usage to offset the lower per-unit earnings. And its original moneymaker faces a different challenge. As Google adds AI answers and rivals develop agents that complete tasks directly, users might have fewer reasons to visit websites. Pew Research found traditional search-result clicks occurred in 8% of visits with AI summaries, versus 15% without them. So even though Google has the customers and technology to make this work. Whether they can generate enough cash to justify the bill is another matter. Buffett’s big Google bet faces an October 28 reality check Interestingly, that question also matters to Warren Buffett’s Berkshire Hathaway (BRK.A), too. Years ago, Buffett admitted he’d missed an opportunity in Google’s advertising business, despite the potential it saw through GEICO’s search-ad spending as reported on CNBC. He eventually initiated Berkshire Hathaway’s Alphabet investment, just as Google was becoming a much more expensive business to build, which I covered a few months ago. Now, investors have an important date circled which is when Alphabet reports earnings on October 28 for Q3. Bank of America expects Google Cloud revenue to grow 93%. But another impressive growth number won’t settle the debate. Investors will want to know whether Search advertising is holding up as AI changes how people find information, and whether Google can turn soaring Cloud demand into stronger margins. Then there’s the spending. Any increase in Alphabet’s CapEx outlook, or a deeper free cash flow shortfall, would raise fresh questions about how quickly its AI investments can pay off. Even headline earnings may need a closer look. BofA estimates Alphabet’s SpaceX stake could produce a significant mark-to-market headwind, complicating comparisons with the unusually large investment gains recorded earlier this year. The big new challenge is whether those businesses can generate enough cash over the coming years to support the extraordinary infrastructure being built around them. Related: Morgan Stanley sees new hurdle for Microsoft stock
CryptoAlerta — análise de criptomoedas e mercado em tempo real