Meta Platforms shares closed at $646.01 on July 17, 2026, giving the owner of Facebook, Instagram, WhatsApp and Threads a market value of about $1.66 trillion. That price rewards one of the world's most profitable advertising businesses, but it also asks investors to look through an unusually expensive artificial-intelligence construction cycle.
Our 10-year discounted cash-flow model produces a base value of about $610 a share. A more cautious case falls near $281, while an optimistic case reaches roughly $1,106. The current price is only about 6% above the base case, yet the width of the range shows how much depends on the return Meta earns from data centers, chips and AI talent.
This is a valuation framework, not personalized investment advice. It uses Meta's latest filed results and a July 17 market snapshot, then makes the assumptions visible so readers can test a different view.
The short answer
Meta looks close to fairly valued in the base case, but it is not obviously cheap. The market is paying for continued double-digit revenue growth, durable advertising margins and a sharp recovery in cash conversion after 2026. If AI improves recommendations and ad performance enough to offset the infrastructure bill, today's quote can work. If spending becomes a permanent cost of defending the platforms, the downside is substantial.
The key valuation question is not whether Meta can generate operating profit. It produced $22.87 billion in the first quarter alone. The question is how quickly those accounting profits turn back into distributable cash after management's planned $125 billion to $145 billion of 2026 capital expenditures.
The numbers
Meta generated $200.97 billion of revenue in 2025, up 22%, and $83.28 billion of operating income at a 41% margin. Family of Apps earned $102.47 billion of operating profit, while Reality Labs lost $19.19 billion. Advertising supplied $196.18 billion of revenue, making the core business powerful but highly concentrated.
Cash conversion was already tightening before the 2026 surge. Operating cash flow reached $115.80 billion in 2025, but capital expenditures including finance-lease principal totaled $72.22 billion, leaving $43.59 billion of reported free cash flow. Meta also paid $5.32 billion in dividends and repurchased $26.26 billion of shares during the year.
The first quarter accelerated both sides of the story. Revenue rose 33% year over year to $56.31 billion. Ad impressions increased 19%, the average price per ad rose 12%, and operating margin held at 41%. But quarterly capital expenditures climbed to $19.84 billion, and free cash flow was $12.39 billion. Management then raised the full-year capital-spending range by $10 billion.
Meta ended March with $81.18 billion of cash, cash equivalents and marketable securities against $58.75 billion of long-term debt. We therefore add about $22.43 billion of net cash to the operating value and divide by 2.564 billion diluted shares. The balance sheet offers flexibility, though the filing also lists $237.67 billion of noncancelable contractual commitments, much of it tied to cloud capacity, servers, networks, data centers and Reality Labs hardware.
How the model works
The model starts with 2025 revenue of $200.97 billion and projects 10 years of unlevered free cash flow. Instead of pretending that 2026's capital intensity is normal, each scenario explicitly changes the free-cash-flow margin over time. This is the share of revenue left after operating costs, normalized taxes, capital investment and working-capital needs, before interest payments.
In the base case, revenue growth begins at 18% in 2026 and slows to 3.5% by year 10. The free-cash-flow margin starts near zero in 2026 because the midpoint of management's capital-spending range absorbs most of the business's after-tax operating cash generation. It recovers to 8% in 2027, 15% in 2028, 22% in 2029 and 30% by the later years as the infrastructure build moderates and new capacity supports revenue.
Those cash flows are discounted at 9%, with a 3% perpetual growth rate after year 10. Adding net cash produces an equity value of about $1.56 trillion, or $610 per share. This is not a forecast precise to the dollar. It is a structured statement of what must happen for the present valuation to make sense.
Bear, base and bull cases
- Bear case: about $281 a share. Revenue growth starts at 14% and fades faster, the free-cash-flow margin recovers only to 24%, the discount rate is 10.5%, and terminal growth is 2.5%. This fits a world where AI raises delivery and depreciation costs, regulation constrains ad targeting, Reality Labs losses persist, and Meta must keep investing heavily just to defend engagement.
- Base case: about $610 a share. Revenue growth starts at 18%, cash-flow margin rises toward 30%, the discount rate is 9%, and terminal growth is 3%. AI improves recommendations and advertising returns, while the current infrastructure wave creates capacity that does not need to be repeated at the same intensity forever.
- Bull case: about $1,106 a share. Revenue growth starts at 22%, cash-flow margin reaches 34%, the discount rate is 8%, and terminal growth is 3.5%. This requires Meta to turn AI into stronger ad pricing, messaging revenue and new products while controlling long-run capital needs and Reality Labs losses.
Why the stock can still outperform
The operating evidence is stronger than a simple capex warning suggests. In the first quarter, both ad volume and price increased at double-digit rates. Meta says AI systems are improving recommendations, ad ranking and creative tools, while more than 3.5 billion people use at least one of its apps daily. Even modest gains in engagement or advertiser return can be meaningful at that scale.
There is also optionality outside the mature feed businesses. Paid messaging, Meta Verified, business agents, AI assistants and wearables could add revenue streams that are not central to this model. The bull case does not assign a separate venture-style value to each project; it lets successful products appear through faster consolidated growth and a higher cash-flow margin.
The caveat
The base case makes an aggressive normalization assumption: free-cash-flow margin climbs from near zero in 2026 to 30% later in the decade. That can happen if much of the current spending is front-loaded growth investment. It will not happen if compute requirements rise as quickly as AI usage, if equipment becomes obsolete faster than expected, or if depreciation and energy costs permanently reset the economics of the ad business.
Reported earnings also need normalization. Meta's first-quarter net income included an $8.03 billion tax benefit, which added $3.13 to diluted earnings per share. The valuation model therefore focuses on operating cash generation and does not capitalize that one-time accounting benefit as recurring profit.
Governance and regulation matter too. Meta's dual-class structure gives Mark Zuckerberg control over major shareholder votes. The company faces privacy, competition, content and AI regulation in multiple jurisdictions. Those risks are difficult to model directly, so they appear mainly through the discount rate and the lower growth and margin assumptions in the bear case.
What to watch next
The next earnings report should be judged on more than revenue and EPS. Watch the full-year capital-spending range, operating cash flow, depreciation, contractual commitments, ad impressions, price per ad and Family of Apps margin. A credible path toward lower capex as a percentage of revenue would support the base case even if near-term free cash flow remains weak.
At $646, Meta is priced like a high-quality company whose AI investment earns a solid return. The shares do not require every new product to succeed, but they do require the core advertising engine to stay strong and the capital cycle to normalize. Investors who disagree most sharply about those two assumptions will naturally arrive at very different values.
Sources and methodology
Financial figures come from Meta's first-quarter 2026 Form 10-Q, its 2025 Form 10-K, and the company's first-quarter results. Valuation outputs are PS News calculations from the assumptions shown above. Market prices move quickly, and readers should update the model before relying on it.
