Why Meta stock below $600 is attracting attention despite concerns about massive AI spending

Meta stock below $600 is attracting attention because its advertising business is growing rapidly while its valuation remains below several major technology peers. The lower price also reflects serious concerns about cash flow, rising debt, data center commitments, and uncertain returns from artificial intelligence.

The debate is not simply about Meta spending too much or its shares looking cheap. Investors must decide if AI-driven advertising gains and possible new businesses can grow faster than the long-term costs of Meta’s infrastructure expansion.

Why does Meta look inexpensive below $600?

Meta looks inexpensive because its stock trades near 22 times trailing earnings while revenue continues growing at a strong rate. That combination has encouraged investors to reconsider the shares after a sizable decline from their 52-week high.

The stock recently traded around $590, roughly 25% below its reported high near $796. A lower share price has placed Meta below the valuation assigned to several large technology businesses and the broader S&P 500.

Still, the price of one share does not determine if a company is cheap. Investors must compare the company’s market value with its earnings, cash generation, debt, expected growth, and future financial obligations.

A trailing price-to-earnings ratio is also based on results from the previous 12 months. It cannot fully account for future depreciation, energy costs, cloud revenue, AI adoption, or another increase in infrastructure spending.

Meta’s latest earnings contained unusual costs that also affected the valuation calculation. The quarter included a $2.4 billion legal charge and $1.18 billion in severance costs connected to the company’s May workforce reduction.

Those charges contributed to a 14% decline in quarterly net income, but they were not all recurring AI expenses. Investors should separate temporary items from continuing infrastructure costs before concluding Meta’s valuation.

Is Meta’s advertising engine still growing?

Yes, Meta’s core advertising operation remains strong despite concerns surrounding its AI strategy. Meta’s official results show second-quarter revenue increased 28% from the previous year to $60.8 billion.

Ad impressions delivered across Meta’s family of apps rose 14% from a year earlier. The average price per ad increased 12%, showing that growth came from both higher advertising volume and stronger pricing.

The company’s family of daily active people reached an average of 3.60 billion in June, up 3% from the previous year. This massive audience gives Meta valuable data and broad distribution across Facebook, Instagram, Messenger, and WhatsApp.

AI is already supporting that advertising system through recommendations, targeting, ranking, and creative tools. These improvements can increase engagement and advertiser results without requiring users to purchase a separate AI subscription.

This distinction is important because chatbot market share does not measure Meta’s entire return from AI. The company can monetize AI indirectly when stronger recommendations keep people engaged or when improved targeting makes businesses willing to spend.

Meta’s SEC filing says AI-powered recommendations have already improved engagement and monetization. The central question is now whether those gains can expand enough to cover Meta’s rapidly rising infrastructure costs.

Fun fact: The analyst targets an entry price of $550 with a liquidation target above $750, emphasizing disciplined entry and hedging due to ongoing risks.

How large is Meta’s AI spending problem?

Meta’s spending problem is large because the company is committing cash to infrastructure faster than its advertising operation can currently replenish free cash flow. Capital expenditures and finance lease payments reached $31.08 billion during the second quarter.

Operating cash flow remained strong at $31.86 billion, but infrastructure spending absorbed nearly all of it. Free cash flow consequently fell 91% from $8.55 billion a year earlier to only $784 million.

Meta expects 2026 capital expenditures of $130 billion to $145 billion, compared with a previous range of $125 billion to $145 billion. The higher minimum signals that management remains committed to an aggressive buildout.

The commitment extends well beyond the current year. Meta reported approximately $349.31 billion in noncancelable contractual obligations, mainly involving cloud capacity, servers, networking equipment, data centers, and Reality Labs hardware.

Meta also disclosed about $278.99 billion in leases that had not started as of June 30. It entered additional data center leases worth approximately $68 billion in July, with those agreements expected to begin during 2027 and 2028.

These figures cover different categories and should not simply be added together as one total. They still show that Meta’s AI expansion is a long-term financial obligation rather than a short construction cycle that can easily be reversed.

Could an AI cloud business change the story?

An AI cloud business could create another revenue stream by allowing outside customers to rent Meta’s computing capacity or access its models. It could also reduce Meta’s dependence on advertising if enterprise demand proves strong.

Reuters reported that Meta was developing a cloud operation that might sell computing power and hosted access to AI models. However, the strategy was still under development, and Meta declined to confirm the report.

Zuckerberg later said outside businesses had offered to access Meta’s computing capacity at a meaningful premium over its construction cost. That suggests the infrastructure may have market value even before Meta develops a full cloud platform.

Selling capacity does not necessarily mean Meta lacks internal uses for its data centers. Management has described computing power as scarce, giving the company the option to shift resources between internal products and outside customers.

However, renting raw computing power is different from running a complete cloud platform. Amazon Web Services, Microsoft Azure, and Google Cloud offer databases, security products, development tools, support systems, and established relationships with corporate customers.

Meta could enter the market faster by selling basic capacity, but matching the largest cloud providers would require years of additional investment. Investors need evidence about customers, contracts, pricing, utilization, and margins before treating cloud services as a proven business.

Fun fact: Zuckerberg is considering selling Meta’s compute capacity to other AI companies, and while external utilization is currently low, a successful pivot could lead to rapid growth; however, it faces fierce competition from established players like Amazon.

Can custom chips make the AI investment work?

Custom chips could improve Meta’s AI economics by lowering costs and reducing its dependence on expensive processors from outside suppliers. They may also help Meta design hardware for specific advertising, recommendation, and model workloads.

Reuters reported that Meta planned to begin manufacturing its Iris AI chip in September 2026. Broadcom is helping with the design, while Taiwan Semiconductor Manufacturing Company is expected to produce it.

The chip is intended to supplement, not completely replace, graphics processors from Nvidia and AMD. Meta reportedly plans to introduce additional custom chips through 2027 as its computing requirements continue expanding.

Meta plans to reach 7 gigawatts of computing capacity during 2026 and 14 gigawatts in 2027, according to an internal memo reviewed by Reuters. Greater scale could support more advanced models, advertising tools, recommendations, and new consumer products.

Yet raw capacity does not guarantee an attractive financial return. Investors need to watch chip performance, energy efficiency, utilization rates, reliability, customer demand, and the revenue generated for each dollar invested.

One supplied article incorrectly used the earlier figure of 5.5 gigawatts for the second half of 2026. Reuters corrected that number to 2.5 gigawatts, making the corrected figure important for any accurate discussion of Meta’s expansion.

What risks could erase Meta’s apparent discount?

The largest financial risk is that depreciation will continue rising after new servers and data centers enter service. Depreciation and amortization reached $6.36 billion during the second quarter, up about 46% from $4.34 billion one year earlier.

Meta’s net property and equipment increased to $225.72 billion, compared with $176.40 billion at the end of 2025. Much of today’s capital spending will therefore affect reported profits gradually over several future years.

Debt creates another expense that the stock’s trailing valuation may understate. Long-term debt rose from $58.74 billion at the end of 2025 to $83.66 billion in June, while quarterly interest expense climbed 225% to $783 million.

Reality Labs remains a separate but substantial drain on earnings. The division lost $4.62 billion during the second quarter and approximately $8.65 billion during the first half, supporting virtual reality, augmented reality, neural interfaces, and wearable devices.

Meta also faces legal, regulatory, privacy, competition, model-safety, and energy-supply risks. The company warned that ongoing youth-related legal cases could create material losses, while AI development introduces risks involving misinformation, copyrights, cybersecurity, bias, and harmful content.

Meta below $600 could reward investors if advertising growth, custom chips, cloud sales, and new products produce durable returns. It could remain under pressure if cash flow, debt, depreciation, and long-term commitments rise faster than those businesses can generate profits.

TL;DR

• Meta generated $60.8 billion in second-quarter revenue, up 28%, as ad impressions and average ad prices both recorded double-digit growth.
• Meta stock trades near 22 times trailing earnings, but a price below $600 does not prove that the shares are undervalued.
• Capital spending reached $31.08 billion during the quarter, leaving Meta with only $784 million in free cash flow.
• Meta may sell AI computing capacity to outside customers, but the proposed cloud business remains under development and could change.
• Custom chips could lower infrastructure costs, while rising depreciation, debt, leases, and Reality Labs losses could weaken future returns.

This article was made with AI assistance and human editing.

If you liked this, you might also like:

• Could Cloudflare’s stablecoin wallet rollout reshape how AI agents access digital services?
• Could the shift to Deribit help Coinbase capture a larger share of the crypto derivatives market?
• Galaxy posts an $85 million Q2 loss, so why are its AI data center investments still attracting attention