Summary: Advertising revenue continues to grow, and AI monetization is beginning to show results, but massive capital expenditures remain the market’s biggest concern.

Editor’s note: Meta’s latest earnings report highlights a central tension in Big Tech’s AI investments: artificial intelligence is beginning to generate revenue for existing businesses, but the investment needed to build the next generation of AI infrastructure still far outpaces the realization of profits from new businesses.

On October 22, Goldman Sachs published a review of Meta’s third-quarter results (Q3'26 Review: Core Advertising Momentum Builds), maintained its “Buy” rating on Meta, and raised its 12-month price target from $860 to $885.

Goldman Sachs is bullish on Meta not simply because it is betting on AI products, but because it believes the company’s core advertising business is continually proving the commercial value of AI investment. At the same time, the newly launched AI agent platform Muse opens up new growth opportunities for Meta in areas such as subscriptions, shopping transactions, and business services.

On the other hand, Goldman Sachs expects Meta’s capital expenditure to reach around $230.3 billion in 2027, with capital expenditure approaching 73% of revenue. This means that even if advertising revenue continues to grow, the company could still face pressure on margins and free cash flow.

Therefore, the key question for Meta going forward is no longer just whether AI can drive growth, but whether growth can come quickly enough to cover the ever-increasing costs of computing capacity and infrastructure.

Here is the original text:

The advertising business remains Meta’s biggest source of growth confidence

Meta generated total revenue of $64.601 billion in the third quarter of 2026, up 26.1% year over year. This exceeded Goldman Sachs’ previous estimate of $63.702 billion and the market consensus of $63.517 billion.

Family of Apps revenue reached $64.158 billion, up 26.4% year over year, while advertising revenue reached $63.061 billion, up 25.9%. Advertising revenue in the United States and Canada grew 33.3% year over year, continuing to be an important source of growth.

However, revenue growth did not translate into a corresponding increase in reported profit. Meta’s GAAP operating profit in the third quarter was $12.988 billion, with an operating margin of around 20.1%, below Goldman Sachs’ previous estimate of 30.4%.

It is worth noting that the quarter’s results included around $9.89 billion in expenses related to the settlement of lawsuits concerning teenagers. According to Meta’s earnings report, excluding this expense, operating profit would have been around $22.88 billion and diluted earnings per share would have risen from $3.83 to $7.40. Therefore, the decline in the reported operating margin for the quarter cannot be attributed entirely to AI investment.

In Goldman Sachs’ view, more important than quarterly profit is that the AI technology Meta invested in earlier has begun to generate measurable returns in its advertising business. In the third quarter, more than 10 million advertisers used at least one of Meta’s generative AI advertising features, and adoption of its video generation feature grew 25% quarter over quarter.

In addition, advertisers that adopted recommendations from Meta AI Business Assistant saw around 5% more conversations while maintaining the same cost per conversion. They subsequently spent 9% more on advertising than similar advertisers who did not adopt the tool. This means AI can not only improve the content users see, but also help advertisers increase conversion efficiency and further strengthen the platform’s appeal to advertising budgets.

This shift is especially important because Meta has a vast existing commercial ecosystem. Company survey data shows that at least around 500 million merchants who identify as small businesses use its Family of Apps to reach customers.

Goldman Sachs believes that as AI advertising tools become more widely adopted among existing advertisers, Meta still has room to raise its advertising revenue outlook. And because the core advertising business is highly profitable, further upward revisions to revenue forecasts could also provide a significant boost to the company’s overall operating profit.

From this perspective, Meta’s AI strategy is not simply a matter of waiting for a new product to succeed in the future. The company already has a vast, mature commercialization system that can directly turn algorithm improvements into greater advertising efficiency and revenue growth.

Muse shows commercialization potential, but scaling it up still comes with cost hurdles

If the advertising business represents proven returns on Meta’s AI investments, Muse represents a potential second growth curve for the future.

Goldman Sachs notes that Muse, a consumer AI agent platform, quickly attracted user attention after its launch. By the time Meta reported its third-quarter results, management disclosed that cumulative downloads of Muse in the United States and Canada were approaching 10 million.

Meanwhile, daily user engagement with Muse is around 25%, and shopping has become one of the platform’s most representative use cases.

According to figures disclosed by management, more than 25% of Muse’s weekly active users use its shopping features, with an average purchase amount exceeding $200. Goldman Sachs took note of these figures because, compared with traditional chatbots, shopping, consumer decision-making, and transaction execution indicate clearer commercial intent and could also generate more direct revenue.

Meta’s management sees several potential monetization paths for Muse, including subscriptions, commercial transactions, advertising, and paid services for businesses. Subscriptions could be among the first models to launch, while shopping and business services could offer longer-term growth opportunities.

Goldman Sachs also highlighted the potential synergies between Muse and Meta’s existing app ecosystem.

According to figures disclosed by management, Muse users in the United States currently generate average annual revenue of several hundred dollars across Meta’s Family of Apps, with the annual Family of Apps ARPU exceeding $1,000 for around 20% of users.

This suggests that Muse’s early users have high commercial value. However, it is important to distinguish that this revenue comes from the existing app family and does not mean that Muse itself has already achieved monetization at a similar scale.

For Meta, if AI agents can eventually participate directly in shopping recommendations, ad placements, and commercial transactions, they could create new revenue streams beyond the company’s existing advertising revenue.

Over the longer term, Muse could also enter the enterprise and small-business markets, helping merchants acquire customers and manage business processes, while enabling consumer agents and merchant agents to complete transactions directly.

However, this growth opportunity still faces practical constraints.

Muse relies on substantial computing power for inference, and a growing user base means higher computing and service costs. Management expects Muse’s compute cost efficiency to improve by more than 10 times over the coming year, mainly through model optimization, improved inference efficiency, and more effective use of idle computing capacity.

Goldman Sachs believes these efficiency gains are crucial to scaling Muse. However, the company will still need to continue investing in infrastructure before it can generate stable revenue.

Therefore, the key test for Muse is not whether it can continue to attract downloads, but whether user engagement, transactions, and commercial revenue can ultimately support a sustainable business model with durable margins.

Capital expenditure could reach $230 billion in 2027 as AI investment enters a high-intensity phase

Compared with Muse’s early growth, Goldman Sachs believes that investment in AI infrastructure and its impact on Meta’s financial position remain the central concerns for investors. Meta’s management narrowed its 2026 capital expenditure guidance to $135 billion–$145 billion, from the previous range of $130 billion–$145 billion.

However, in Goldman Sachs’ forecast, 2026 may only mark the start of a new round of capital expenditure growth.

According to the research report, Goldman Sachs expects Meta’s capital expenditure to reach around $141.2 billion in 2026, then rise to around $230.3 billion in 2027, an increase of 63.1% year over year. Capital expenditure is projected to reach around $250.4 billion in 2028.

This means that Meta’s capital expenditure is expected to equal 72.9% of revenue in 2027 and remain elevated at 66.9% in 2028.

It is important to emphasize that the capital expenditure figures for 2027 and beyond are Goldman Sachs analysts’ forecasts, not formal annual guidance from Meta’s management. In its latest earnings report, Meta said that its 2027 budget and computing capacity plans are still being developed, but that it expects to continue investing aggressively in AI.

Higher capital expenditure will further affect the company’s depreciation expenses, margins, and free cash flow.

Goldman Sachs expects Meta’s depreciation and amortization expenses to rise from around $18.6 billion in 2025 to around $34.7 billion in 2026, then reach around $63 billion and $94.9 billion in 2027 and 2028, respectively.

Meanwhile, Goldman Sachs expects Meta’s free cash flow to fall from around $46.1 billion in 2025 to around $6.4 billion in 2026, turn negative at around $33.5 billion in 2027, and recover slightly to around $1.4 billion in 2028.

This indicates that even if Meta’s revenue and operating cash flow continue to grow, large-scale infrastructure construction could significantly reduce the cash available for discretionary use in the coming years.

In fact, Meta generated $32.79 billion in operating cash flow in the third quarter of 2026, but free cash flow for the quarter was still negative $2.12 billion, already reflecting the impact of capital expenditure on cash flow.

Meta is also adjusting how it allocates investment across its different businesses.

Management expects Reality Labs’ annual operating loss to reach around $18 billion–$19 billion in 2026, then narrow to around $15 billion in 2027. Goldman Sachs believes that a narrower loss at Reality Labs would help ease some operating cost pressure. However, given that investment in AI computing capacity is still expected to increase, this improvement would not be enough to fully offset the impact of the new wave of technology investment on margins.

In addition, Meta’s announcement that it would establish non-voting Class C common stock has raised investor questions about its future capital structure and financing plans.

Goldman Sachs notes that this arrangement gives Meta a more flexible capital structure, but management did not indicate on the earnings call that the company would immediately raise funds as a result.

Therefore, over the coming quarters, investors may pay closer attention to Meta’s financing needs, the efficiency with which it uses computing capacity, and the extent to which revenue from new businesses can cover infrastructure costs.

Goldman Sachs raises price target to $885, remains bullish on long-term growth

Despite the possibility that capital expenditure will continue to weigh on financial performance, Goldman Sachs remains positive on Meta.

Goldman Sachs raised its 12-month price target for Meta from $860 to $885 and maintained its Buy rating. Based on the reference share price of $697.52 cited in the report, the new target implies potential upside of around 26.9%.

For its earnings forecasts, Goldman Sachs raised its 2026 revenue estimate for Meta to around $257.2 billion and its 2027 estimate to around $315.9 billion.

However, as its cost forecasts were also adjusted, Goldman Sachs lowered its 2026 earnings-per-share estimate from $31.67 to $28.84 and its 2027 estimate from $32.26 to $31.02.

This adjustment reflects Goldman Sachs’ two-sided view of Meta: the revenue growth outlook has improved, but continued investment expansion is still weighing on earnings in the short to medium term.

For its valuation methodology, Goldman Sachs uses a multiple of enterprise value to GAAP operating profit and an adjusted discounted cash flow model, assigning a 50% weighting to each.

According to its scenario analysis, Meta’s price target is $885 in the base case, could fall to $520 in the bearish case, and could reach $1,170 in the bullish case.

Goldman Sachs believes that the key factors affecting Meta’s share price going forward fall into three main areas.

First, can the core advertising business sustain strong growth? As AI recommendation algorithms, ad creation tools, and business support features become more widely adopted, Meta has the opportunity to keep improving ad conversion efficiency and strengthen its competitive position in the digital advertising market.

Second, Muse’s commercialization progress. Although Muse has already built an initial user base and shown strong demand for shopping features, it remains to be seen whether it can establish sustainable revenue models for subscriptions, transactions, and business services.

Finally, and most importantly, can the economic returns on AI investment gradually materialize? If computing utilization, inference efficiency, and model operating costs continue to improve, Meta will have the opportunity to expand its AI products while increasing investment efficiency.

Conversely, if capital expenditure continues to rise while growth in new businesses or improvements in computing efficiency fall short of expectations, the company could face margin pressure for a longer period.

Goldman Sachs also cautions that Meta faces competition from internet, media, and e-commerce companies, as well as uncertainties including regulatory scrutiny, antitrust risks, and the possibility that AI commercialization will fall short of expectations.

Over the long term, Goldman Sachs still believes Meta is at the intersection of structural growth trends in AI, social platforms, and spatial computing. The cash flow and commercialization capabilities of its core advertising business provide a foundation for the company to explore the next generation of AI products.

However, with capital expenditure expanding rapidly, the criteria the market uses to evaluate technology companies are also changing.

In the past, investors focused more on how much technology companies were investing in AI. Going forward, the market may care more about how much additional revenue, profit, and free cash flow those investments actually generate.

For Meta, the advertising business has already provided initial proof of AI’s value, while Muse points to the possibility of a next-generation business model. The key challenge still to be addressed is how to turn its ever-growing investment in AI infrastructure into substantial and sustainable returns for shareholders.