Analysts See Undervaluation in Meta Platforms Despite Aggressive AI Capex
Market reaction to increased capital expenditure contrasts with analyst views on long-term returns from AI-driven ad revenue growth and reduced reliance on Nvidia hardware.
Analysts are arguing that the market is currently undervaluing Meta Platforms (META) because of its aggressive, full-stack artificial intelligence strategy. Despite recent negative market reactions to hikes in capital expenditure, CEO Mark Zuckerberg is maintaining high spending on AI initiatives, betting on long-term returns rather than short-term stock fluctuations. This approach involves leveraging AI to boost advertising revenue through smarter targeting and automation, a move that is already showing signs of rapid scaling.
The company's AI-powered ad tools have seen adoption double, rising from 4 million advertisers to approximately 8 million advertisers in the current period. This expansion underscores the firm's confidence that its technology can effectively monetise AI across its platforms. Polen Focus Growth Strategy has re-initiated a position in META in its Q1 2026 investor letter, citing strong execution in this area despite concerns over the elevated data centre capital expenditure.
To reduce reliance on Nvidia GPUs, Meta is developing custom Training and Inference Accelerator (MTIA) chips designed to operate at roughly 25W. This is significantly lower than the power consumption of competing Nvidia A100s, which draw 250–500W, and H100s, which use around 350W. The newer MTIA generations are expected to handle recommendation systems and generative AI workloads, potentially unifying AI training, inference, and advertising workloads under its own hardware stack.
Funding this full-stack AI approach requires significant financial discipline, leading the firm to implement cost-cutting measures including layoffs and an efficiency push. While the market reacts negatively to these CapEX hikes, the leadership remains focused on the long-term potential of the technology. The strategy aims to create a unified ecosystem that reduces dependency on external hardware suppliers while driving revenue growth through enhanced ad capabilities.
In the broader context of the sector, institutions continue to show heavy buying of Nvidia shares amid strong earnings, providing a backdrop to Meta's strategy of reducing reliance on Nvidia hardware. Meta Platforms ranks ninth in Cathie Wood's list of 11 biggest AI and data centre stock picks, with her stake valued at approximately $197 million. This institutional interest highlights the divergence between immediate cost concerns and the perceived long-term value of deep AI integration in major tech platforms.
While Polen Focus Growth Strategy notes the uncertainty around the ultimate return on investment without a comparable cloud business, the firm remains encouraged by Meta's execution. The re-initiation of the position suggests that investors are beginning to look past the immediate costs and focus on the structural advantages Meta is building through its custom hardware and advertising ecosystem.
