Meta is investing at a scale few companies can match, but the artificial intelligence race remains highly competitive. The company has enormous financial resources, a massive user base, and some of the world’s largest consumer platforms. Meta has recently narrowed the performance gap, although OpenAI and Anthropic remain among the strongest frontier-AI competitors.
AI is already moving beyond traditional chatbots. Assistants and generative AI features now appear across smartphones, wearable devices, search products, and workplace software, while companies continue to develop more capable agents and connected experiences.
Meta’s challenge is no longer proving its commitment to AI. The company expects 2026 capital expenditures of $130 billion to $145 billion, much of which will support AI infrastructure and the company’s broader business. The question is whether those investments can produce enduring technical advantages, new revenue streams, and products that businesses and consumers continue to use.
Meta’s AI strategy faces scrutiny
Meta has reorganized its artificial intelligence efforts to focus on developing advanced AI systems capable of competing with the strongest models in the industry. The company has increased spending on research teams, infrastructure, specialized hardware, and computing capacity.
The company’s rising expenses highlight how expensive the AI race has become. Meta has increased its capital spending plans to support new data centers and additional computing resources. The company has also expanded agreements with external computing providers.
These investments demonstrate Meta’s commitment to becoming a major AI player, but they also raise questions about efficiency. Building infrastructure is necessary for advanced AI development, yet competitors are already showing how those investments can translate into popular products and business services.
Is spending rising faster than results?
Meta’s massive AI spending continues to attract scrutiny, especially after the second-quarter free cash flow fell sharply. However, its competitive position has improved significantly in recent months. Meta’s newest models have narrowed the gap with the frontier, although OpenAI and Anthropic remain widely regarded as leading competitors.
OpenAI continues to have enormous consumer and workplace reach through ChatGPT, while Anthropic has built a strong enterprise business around Claude. Google remains a major AI competitor with Gemini integrated into products such as Search, Android, and Workspace, although recent leadership changes and model setbacks have created fresh uncertainty about its position.
Meta also has an advantage that many AI developers lack: its Family of Apps averaged 3.60 billion daily active users in June 2026. The central question is whether Meta can turn that distribution, along with its improving models and infrastructure, into durable AI businesses beyond the benefits AI already provides to its existing products.
AI glasses become Meta’s major bet
Meta’s most visible consumer AI strategy focuses on wearable technology. The company believes smart glasses could become one of the most natural ways people interact with artificial intelligence because they allow users to access assistants without constantly reaching for a phone or opening an application.
Through its partnership with EssilorLuxottica, Meta has continued developing AI-powered smart glasses designed for everyday use. These devices combine cameras, speakers, and AI assistants to provide features such as answering questions, capturing images, translating languages, and offering hands-free assistance.
This strategy shows that Meta is competing through product integration as much as AI model performance. While other companies are focusing on software subscriptions and enterprise services, Meta believes the future of AI could depend on making assistants available wherever users go.
Can open models close the gap?
Meta has returned to an open-weight AI strategy, but its current push has moved beyond the Llama family.
On August 10, 2026, Meta released Muse Glimmer, an open-weight model designed to run agentic tasks locally on consumer hardware. The company has also said it plans to release the weights for Muse Spark 1.2, its most advanced foundation model.
Llama remains an important part of Meta’s open-model history and reached more than 1 billion downloads by March 2025. However, Meta’s newest strategy centers increasingly on the Muse family developed by Meta Superintelligence Labs.
Open weights can make models easier to customize and deploy, potentially expanding Meta’s developer ecosystem. But openness alone will not determine the AI race. Meta still needs competitive performance, useful products, attractive economics, and sustained adoption by developers, businesses, and consumers.
Investors want clearer AI returns
Meta’s growing AI expenses have attracted increasing investor attention, with investors seeking evidence that the spending will yield meaningful financial returns. The company remains highly profitable because of its advertising business, but AI has not yet become a major standalone revenue source.
Higher infrastructure costs, additional hiring, and increased investment in research have created pressure on Meta’s financial outlook. Although the company has enough resources to continue spending, investors want proof that these investments will eventually create a new growth engine.
Interesting fact: Meta signed a 20-year agreement with Constellation Energy to support the continued operation of Illinois’ Clinton nuclear plant, which generates about 1,121 MW of power.
Meta still has room to compete
Despite concerns, Meta remains one of the strongest competitors in artificial intelligence. The company has financial strength, engineering talent, global platforms, and consumer products that provide opportunities few companies can match.
The company continues to improve its AI models, expand its Superintelligence Labs, and develop smart glasses while attempting to connect AI with its existing ecosystem. Its ability to test AI features across billions of users gives it a unique advantage.
However, the AI industry is moving quickly, and early investment does not guarantee future success. Meta must show that its combination of social platforms, hardware, and open models can create advantages that competitors cannot easily copy.
The consumer AI battle expands
The artificial intelligence race increasingly depends on where and how people experience AI in daily life. Meta is using its social platforms and AI glasses to make assistants available during activities users already perform, from messaging and content creation to taking photos and asking questions hands-free.
Competitors are pursuing different but overlapping strategies. Anthropic has built a strong workplace and enterprise offering around Claude, while Google integrates Gemini across both consumer and productivity products, such as Android and Workspace. Meta itself is also moving beyond consumer services by offering developers access to its models.
That means the competition is no longer neatly divided between consumer AI and enterprise AI. Meta’s advantage lies in its distribution across social apps and wearables, but it still has to prove that those experiences are useful enough to become enduring habits rather than occasional features.
TL;DR
- Meta is investing billions into artificial intelligence infrastructure, but critics question whether spending alone can help the company surpass established AI leaders.
- The company’s AI glasses strategy aims to bring assistants into everyday life through cameras, audio, and hands-free interactions.
- OpenAI, Anthropic, and Google currently have stronger positions through enterprise adoption, developer services, and platform integration.
- Meta’s Llama open-model strategy could increase its influence, but the company still needs stronger proof of performance and profitability.
- Investors are watching to see whether Meta’s AI investments can create a new business opportunity rather than merely improving its existing advertising operations.
This article was made with AI assistance and human editing.
If you liked this, you might also like:
