Marc Benioff is openly questioning whether Big Tech’s AI spending spree is delivering meaningful progress. The Salesforce CEO, known for his blunt industry assessments, has raised concerns about the billions of dollars being poured into artificial intelligence, particularly by Microsoft, Google, and Meta.
During Salesforce’s earnings call on Wednesday, Marc Benioff emphasized that the company is avoiding investments that could drain its financial resources without ensuring substantial returns.
“We aren’t building huge $10 million, $20 million, $30 million, $100 billion data centers,” Benioff stated. “We’re not doing some of these kind of engineering efforts that may or may not have some kind of huge payoff, but is going to take down all of our cash and all of our margin for the next several years.”
His comments come as some of the biggest players in AI start to scale back their AI infrastructure plans, while others double down on massive investments.
Microsoft had committed $80 billion to AI infrastructure, aiming to dominate cloud-based artificial intelligence. However, the company has halted construction on a $3.3 billion AI data center in Wisconsin and canceled multiple cloud expansion leases.
Google, despite its aggressive AI push, faced a stock decline after announcing $75 billion in AI investments, leaving investors uneasy about its long-term profitability. Meanwhile, OpenAI, a key Microsoft partner, is increasingly shifting away from Azure, after securing a $40 billion investment from SoftBank to expand its AI compute infrastructure independently.
Microsoft Adjusts AI Infrastructure Plans Amid Demand Recalibration
Microsoft’s AI ambitions have always revolved around its cloud division, particularly its partnership with OpenAI. The company’s investment in OpenAI gave it a competitive edge, integrating the AI research firm’s models into products like Microsoft 365 Copilot and Azure AI. However, recent moves suggest that the company may have overestimated short-term AI infrastructure demand.
Despite its initial expansion plans, Microsoft has begun slowing its AI infrastructure growth. The company has cut back on data center leases, affecting hundreds of megawatts of planned cloud capacity. Analysts suggest that Microsoft’s optimistic AI demand forecasts did not fully materialize, particularly as OpenAI seeks alternative compute providers.
Satya Nadella has publicly reaffirmed Microsoft’s close ties with OpenAI, stating that “OpenAI APIs are exclusive to Azure going forward even, so nothing changes there… OpenAI has committed to Azure in a very significant way.”
However, OpenAI’s latest moves indicate otherwise. The company is reportedly shifting some of its workloads to SoftBank in the Stargate Project, a large-scale AI data center initiative backed by the Trump administration, diversifying its infrastructure providers.
Google’s $75 Billion AI Investment Draws Investor Scrutiny
While Microsoft is adjusting its AI infrastructure plans, Google is taking an aggressive expansion approach. The company has pledged $75 billion to AI development, including expanding its cloud AI services and training advanced machine learning models.
However, rather than boosting confidence, the investment led to a decline in Alphabet’s stock since its announcement, with investors worried about whether these costs will translate into immediate revenue.
Google’s cloud division plays a central role in its AI strategy, yet recent earnings reports indicate that revenue growth did not meet expectations, prompting further concerns over the sustainability of AI infrastructure investments.
While Google remains a leader in AI research with its Gemini models, financial analysts remain skeptical about whether its aggressive cloud expansion will deliver strong near-term returns.
OpenAI Moves Away From Microsoft, Strengthens SoftBank Partnership
One of the most striking developments in AI infrastructure is OpenAI’s shift toward a multi-cloud strategy. The AI company has relied heavily on Microsoft Azure, but recent investments suggest that it is looking to reduce dependency on a single cloud provider. OpenAI has secured a $40 billion investment from SoftBank, which will help it expand beyond Microsoft’s cloud services.
A SoftBank executive explained the move, stating that “Software alone won’t be enough”, emphasizing the importance of AI-specific hardware and infrastructure. OpenAI is expected to allocate a portion of its workloads to SoftBank’s Stargate Project, which aims to provide independent AI computing resources. This marks a shift in OpenAI’s long-term strategy, ensuring that it has access to multiple AI compute providers.
Meta’s $200 Billion AI Bet: A Power Move or a Risky Gamble?
While Microsoft and Google adjust their AI infrastructure strategies, Meta is pushing forward with one of the largest AI investments ever recorded. The company is reportedly planning a $200 billion AI data center expansion, positioning itself as the dominant player in AI compute power.
Meta has already built a 100,000 Nvidia H100 GPU cluster to train its AI models, including the upcoming Llama 4. However, this unprecedented expansion raises questions. If AI demand slows or infrastructure costs outpace adoption, Meta could be left with excess computing capacity and significant financial strain.
Some analysts argue that Meta’s strategy is an all-in bet on the future of AI, while others warn it could create long-term operational inefficiencies if the AI market does not grow as expected.
Apple’s AI Investment Strategy Focuses on Hardware
Unlike its competitors, Apple is taking a different approach to AI investment. The company has committed $500 billion to AI and semiconductor manufacturing, but its focus is not on cloud-based AI models. Instead, Apple is prioritizing on-device AI capabilities and chip production, ensuring that its AI processing remains tightly integrated with its hardware ecosystem.
Apple’s strategy aligns with its historical preference for controlling its supply chain and reducing reliance on third-party vendors. This investment also plays into the broader U.S. semiconductor push, which has been incentivized by government initiatives such as the CHIPS Act. While other companies are betting on large-scale cloud AI, Apple is ensuring that its AI functions are optimized for its own devices, reinforcing its long-term product roadmap.
Benioff’s History of Criticism Toward Microsoft
Benioff’s skepticism about AI spending is not new. His criticism of Microsoft dates back to 2016, when he stated, “I just kinda came to the conclusion at that point that the new Microsoft was actually the old Microsoft.”
His frustrations stemmed from Microsoft’s aggressive business tactics, particularly following its acquisition of LinkedIn, which Salesforce attempted to block on antitrust grounds.
Now, Benioff’s concerns center on whether AI investments are being made strategically or simply as a reaction to industry pressure. He has warned against chasing AI trends without clear business outcomes, contrasting Salesforce’s targeted AI approach—such as its expansion of Agentforce 2.0—with the infrastructure-heavy investments being made by Microsoft and its competitors.
The Future of AI Spending: Sustainable Growth or Overbuilt Infrastructure?
The shifting AI infrastructure strategies of Big Tech raise fundamental questions about whether AI investments are sustainable or speculative. Microsoft’s retreat from some data center projects suggests that early growth projections may have been overly optimistic. OpenAI’s move to diversify its infrastructure with SoftBank also reflects an industry shift toward more flexible and cost-efficient AI computing models.
With billions already invested and further spending planned, the coming years will show whether AI infrastructure expansion will fuel transformative breakthroughs—or if companies will be left with overbuilt and underutilized data centers.
