There are two ways to read an earnings call. The first is to focus on the numbers: revenue, profit, margins, and guidance. The second is to read between the lines, paying attention to the words executives use over and over again—and the words they suddenly stop saying.
On Meta’s Q3 2025 earnings call, the numbers mattered, but the vocabulary told an even sharper story. According to coverage of the call, Meta execs used the word “compute” 22 times and “metaverse” zero times. “Capacity” came up 11 times. “Novel,” in the context of AI models, appeared seven times. The metaverse, once the centerpiece of Meta’s narrative, was completely absent from the hour-long conversation.
Three years earlier, Meta CEO Mark Zuckerberg had said “metaverse” eight times on a single third-quarter earnings call, framing it as the company’s future and the reason for massive Reality Labs spending. This time, he did not say the word once.
Instead, the spotlight fell on compute, AI infrastructure, and capital expenditure. Meta raised its capex guidance to an astonishing seventy to seventy-two billion dollars for 2025, with even higher spending expected in 2026, mostly to build out AI data centers and supercomputing capacity. Investors, already uneasy after years of expensive metaverse bets, reacted by sending the stock into a double-digit slide.
The title “ ‘Compute’ 22, ‘metaverse’ 0: The words Meta execs did and didn’t say on its earnings call” captures more than a clever count. It marks a profound rhetorical pivot from a futuristic virtual world story to a present-tense AI and compute story. In this article, we break down what this language shift suggests about Meta’s strategy, how it is trying to reassure investors and what it could mean for the fate of the metaverse inside Meta.
Counting words: What “compute 22, metaverse 0” really signals
The lexicon of Meta’s Q3 2025 call
Earnings calls are carefully scripted performances. Executives do not repeat certain words by accident, and they certainly do not drop a core brand concept by mistake. On this Q3 2025 call, “compute” became the anchoring term. Meta’s leadership used it to describe everything from AI training clusters and inference capacity to long-term infrastructure bets and rising capital expenditures. Analysts who tallied the transcript noted that:
Zuckerberg and CFO Susan Li used “compute” 22 times combined, framing it as the scarce resource Meta must aggressively secure. “Capacity” appeared 11 times, emphasizing both physical data center space and the number of advanced chips the company can deploy. “Novel” was used seven times to describe new AI models, “novel products,” and “novel capabilities” that Meta hopes to unlock with its compute build-out.
In stark contrast, “metaverse” registered exactly zero mentions. The word “Horizon,” Meta’s brand for its virtual world platform, was also absent. Reality Labs, the division that houses VR headsets and AR devices, was discussed primarily in financial terms—losses, spend, and some commentary about hardware—rather than as the gateway to a grand metaverse vision. This is not just a matter of fashion. It is a deliberate reset of Meta’s story to Wall Street, designed to shift attention from speculative virtual worlds toward tangible, if still expensive, AI infrastructure and compute capacity.
A sharp break from the metaverse-heavy calls of the past
To understand the significance of “metaverse” disappearing, it helps to recall just how dominant the term was a few years ago. In 2021 and 2022, Meta rebranded from Facebook, renamed its stock ticker, and poured tens of billions into Reality Labs, all while Zuckerberg talked relentlessly about building a fully immersive metaverse. On a prior Q3 call, he used the word “metaverse” eight times in prepared remarks alone.
That metaverse narrative came with severe market consequences. As Reality Labs losses mounted and VR adoption lagged, investors began to see the metaverse as an expensive distraction. Criticism intensified that Meta was overbuilding infrastructure for a future that might not arrive quickly enough, if at all.
Now, with AI in the spotlight, Meta is trying hard not to repeat that mistake. By saying “compute” instead of “metaverse”, management is attempting to reframe its spending as a necessary response to the AI arms race—an investment in foundational infrastructure, not an indulgence in a speculative side project. The word counts on this earnings call are the clearest proof of that rhetorical pivot.
Why “compute” became the star of the call
The AI arms race and the scarcity of compute
In 2025, the most valuable resource in tech is arguably not talent or data, but high-end compute: the GPU clusters and data centers needed to train and run massive AI models. Meta’s executives made this point repeatedly. They described compute as the “bottleneck,” the limiting factor that determines how quickly they can ship new AI features across Facebook, Instagram, WhatsApp and Threads.
Zuckerberg emphasized that Meta must front-load its AI infrastructure build-out, buying chips, building data centers, and wiring up networks now in order to support future breakthroughs. The company has created a “Superintelligence Labs” team, expanded its AI talent pool and is purchasing huge volumes of cutting-edge accelerators. On the call, compute is presented not just as an asset, but as a moat. Meta argues that owning unprecedented compute at scale will:
Allow it to train “novel” frontier models that differentiate its AI assistants, recommendation engines and ad systems. Enable it to run those models efficiently in production, improving engagement and monetization across its family of apps. Position it alongside or ahead of other AI leaders in both consumer experiences and AI infrastructure. In that frame, “compute” becomes a reassuring word for investors. It suggests hard assets, data centers, and chips—things that can be counted and depreciated—rather than an nebulous, hype-laden idea like the metaverse.
Capital expenditure and the AI supercomputer narrative
Of course, compute is not cheap. Meta raised its capital expenditure forecast to seventy to seventy-two billion dollars for 2025, nearly double its 2024 spend, and indicated that capex will climb even higher in 2026 as it builds out AI supercomputing infrastructure.
On the call, executives tied this spending directly to compute. They talked about constructing next-generation data centers, optimizing for AI workloads, and ensuring enough capacity to support both training and inference. The phrase “compute capacity” becomes a way to justify the sheer scale of these investments, implying that the company’s future depends on owning as much high-quality compute as possible.
Investors, however, are nervous. Meta’s share price fell more than ten percent after the call, erasing over two hundred billion dollars in market value as analysts worried about “runaway AI spending” and drew uncomfortable parallels to the metaverse overbuild of earlier years. In that context, the word “compute” is doing double duty. It is meant to sound disciplined and infrastructure-focused, even as the numbers attached to it keep climbing.
The metaverse goes quiet: Strategy shift or just storytelling?
Zero mentions on the call, but not zero importance inside Meta
So what happened to the metaverse? Did Meta simply abandon it? The answer is more nuanced than the word count suggests.
Internally, Meta leaders insist that Reality Labs and the metaverse strategy remain important. CTO Andrew Bosworth has reportedly told employees that this year is critical to prove whether the metaverse is a visionary success or a “legendary misadventure,” and internal memos continue to frame metaverse work as a company-wide priority.
Externally, however, the metaverse brand has become toxic with investors. The term is now associated with years of heavy losses, slower-than-expected adoption of VR headsets and skepticism about whether consumers truly want to live in virtual worlds.
By saying “metaverse” exactly zero times on the earnings call, Meta is effectively telling Wall Street: we are not here to sell you a dream about avatars; we are here to talk about AI, compute and cash flow. The company is still spending billions per quarter on Reality Labs, but it is choosing to foreground more tangible, near-term narratives.
From “metaverse company” to “AI and compute company”
This rhetorical pivot mirrors a broader strategic repositioning. Meta once described itself in almost grandiose terms as a metaverse company, implying that social networking was just a stepping stone. Now, it talks about being an AI-first company with a huge user base, running AI recommendation systems, generative tools, and assistants across Facebook, Instagram, WhatsApp and Threads.
On the call, the platforms mentioned most frequently were Facebook, Instagram and Threads, with WhatsApp oddly in the background. Horizon Worlds, the flagship metaverse product, did not come up. Instead, Meta highlighted AI-boosted engagement metrics, like increased time spent on video, better ad performance and more relevant recommendations. In short, the company is now selling itself as an AI platform with massive distribution, rather than as the architect of a new reality. Metaverse efforts continue, but they are being folded into a broader compute and AI story, not treated as the headline act.
Reading the silence: What Meta’s omissions tell us
The power of not saying “metaverse”
Silence can be louder than repetition. By refusing to utter “metaverse,” Meta signals a desire to distance its current narrative from the most controversial chapter of its recent history.
For investors, this is both reassuring and unsettling. On one hand, it suggests that Meta heard feedback about metaverse overexposure and is trying to show more financial discipline. On the other hand, the continued Reality Labs losses—which still run into billions per quarter—have not gone away. They have simply been pushed offstage, tucked behind a curtain of compute-centric language.
The question becomes whether this is a temporary storytelling adjustment—an attempt to avoid spooking Wall Street while AI investments are ramping up—or a deeper admission that the original metaverse vision will never be the company’s central narrative again.
The missing competitors, politics, and Horizon
Another notable aspect of the call is what else went unmentioned. Meta executives did not dwell on competitors like Apple’s Vision Pro or emerging mixed-reality ecosystems. They said little about regulatory threats or political storms beyond briefly acknowledging a huge one-time tax charge triggered by new U.S. legislation.
The absence of Horizon Worlds and the broader “metaverse” vocabulary fits into this pattern. Meta wanted a tight, focused script that hammered three core ideas: AI is working, compute is scarce, and we are building the infrastructure to win. Anything that did not reinforce that message—and the metaverse fits squarely into that category—was left out.
Investor takeaways: Between AI opportunity and capex anxiety
AI is delivering engagement gains, but at a rising cost
Meta’s executives did not only talk about compute in the abstract. They tied AI to concrete improvements. On the call, they highlighted that AI recommendation systems are driving more time spent on Facebook and Threads, with video as a particular bright spot. AI is powering smarter ad targeting, better Reels suggestions and new creative tools, all of which support revenue growth.
This is the optimistic side of the compute-heavy narrative. The company is not just building giant AI clusters for fun; it claims those clusters are already translating into better engagement and monetization.
The pessimistic side is the cost. Capex is surging. Operating expenses are rising. Reality Labs is still a drag. The stock’s sharp post-earnings drop reflects a fear that Meta is replaying its metaverse strategy with a new label: this time the big bets are on AI supercomputing rather than virtual worlds, but the cash burn looks familiar.
Can “compute” avoid becoming the new “metaverse”?
One risk of leaning so heavily on compute as the buzzword is that it may eventually inherit the baggage of “metaverse” if returns do not materialize quickly enough. The market tolerated massive Reality Labs losses for a while, but eventually rebelled. The same could happen to AI infra spending if Meta cannot convincingly show that each new billion in capex is generating durable revenue.
For now, though, the word count story is clear. Meta wants investors to associate it with compute, capacity and AI, not with a capital-M Metaverse. Whether that rebranding sticks will depend less on the language of future calls and more on how the numbers evolve in the next few years.
What it means for the future of the metaverse inside Meta
Metaverse work becomes infrastructure and OS, not slogan
Even if the word “metaverse” has disappeared from Meta’s earnings vocabulary, its components are still very much alive. Reality Labs continues to develop Quest headsets, mixed-reality experiences and Horizon OS, the underlying platform for XR devices. Recent restructurings suggest Meta is separating the OS and platform work from the broader “metaverse” marketing story, making it look more like a long-term infrastructure bet than a consumer mega-brand.
In this new framing, the metaverse is not a product line that must be defended on every earnings call, but one of several long-horizon projects sitting on top of the same compute and AI foundation. That makes it easier for Meta to scale back or redirect metaverse efforts without admitting defeat, while still claiming progress in spatial computing, AR glasses and mixed-reality interfaces.
AI and metaverse quietly converging
There is also a technical convergence underway. The same AI models Meta is building for assistants, search and content creation will eventually be crucial in immersive environments. Generative AI can power lifelike NPCs, adaptive worlds and personalized experiences inside virtual and augmented reality.
From that perspective, the shift from “metaverse” to “compute” is not necessarily a retreat; it is a recognition that the metaverse, if it ever becomes mainstream, will be built on top of enormous AI compute stacks. By focusing on compute now, Meta is arguably laying the groundwork for both an AI-first future and a more subtle, less hyped version of the metaverse down the line.
Why earnings-call language matters more than ever
Words as a window into strategic priorities
In the AI and platform wars, capital markets reward clarity. When a company like Meta shifts its language from metaverse to compute, it is sending a signal about what it wants to be judged on.
Saying “compute” 22 times tells investors: judge us on our ability to build infrastructure, scale AI, and monetize it across our apps. Saying “metaverse” zero times tells them: do not anchor your expectations on a grand VR vision that may take a decade to mature. For analysts and observers, tracking these linguistic shifts can be just as revealing as reading the balance sheet. They expose where management feels confident, where it feels defensive and how it wants to frame risk.
From buzzwords to balance
Ultimately, Meta’s challenge is to avoid letting “compute” become just another buzzword. The company has already lived through one cycle where a visionary concept—the metaverse—drove huge investments and bold promises, only to be followed by skepticism and backlash.
If “compute” and AI infrastructure are to remain credible pillars of Meta’s story, they will have to be backed by sustained improvements in user value, revenue and margins. The Q3 2025 call suggests Meta understands this, even if the market remains unconvinced in the short term. The phrase “ ‘Compute’ 22, ‘metaverse’ 0” may be clever shorthand for one earnings call, but it also captures a deeper truth: the company once known for social networking and metaverse dreams now wants to be known, above all, as a builder and owner of world-class compute.
Conclusion
The linguistic scorecard from Meta’s Q3 2025 earnings call—“compute” 22, “metaverse” 0—is more than a curiosity. It is a concise summary of how the company wants to reposition itself in the eyes of investors and the broader tech world.
By foregrounding compute, capacity and novel AI models, Meta is arguing that its future rests on building enormous AI infrastructure and using it to improve products that already reach billions of people. By scrubbing “metaverse” from its spoken script, it is quietly distancing its present narrative from a costly and controversial chapter, even as Reality Labs and XR work continue behind the scenes.
For observers, the key is to look beyond the buzzwords and watch how strategy, spending and results align. If Meta can turn its AI and compute-heavy capex into durable advantages across its apps, the new rhetoric will look prescient. If not, “compute” risks becoming the next “metaverse”—a word that dominated calls for a few years before disappearing in a cloud of disappointment. For now, the words Meta execs did and did not say on this earnings call offer a clear message: the metaverse may still exist inside the company, but in the world of investor communication, compute sits firmly at the center of the stage.
FAQs
Q: Why did Meta executives say “compute” so often and avoid “metaverse” on the call?
Meta executives leaned heavily on the word “compute” to emphasize their focus on building AI infrastructure, data centers and GPU capacity, which they see as critical to future products and revenue. At the same time, they avoided saying “metaverse” to distance their current narrative from earlier years when that term was associated with heavy losses and skeptical investors. The new language frames spending as disciplined investment in AI and compute, rather than as a speculative bet on virtual worlds.
Q: Does the absence of the word “metaverse” mean Meta is abandoning the metaverse entirely?
Not necessarily. Internally, Meta leaders still describe Reality Labs and metaverse-related work as important. Hardware like Quest headsets, AR glasses and Horizon OS continues to receive investment. However, the company no longer wants the metaverse to be the headline message to Wall Street, especially while it is committing tens of billions of dollars to AI and compute. The silence on “metaverse” is more about investor storytelling than about shutting down the underlying projects.
Q: How are Meta’s AI and compute investments affecting its financials?
Meta’s AI and compute infrastructure investments are driving a significant increase in capital expenditures. The company has guided capex to around seventy to seventy-two billion dollars in 2025, with expectations of even higher spending in 2026. While AI is already improving user engagement and ad performance across Facebook, Instagram and Threads, investors worry that the scale of spending echoes the earlier metaverse overbuild. The stock dropped sharply after the earnings call as analysts questioned whether these investments will generate sufficient long-term returns.
Q: What role does Reality Labs play now that Meta is focusing on compute?
Reality Labs still plays a major role inside Meta as the home of VR, AR and spatial computing initiatives. The division continues to report large operating losses, but it is now framed as one of several long-term bets rather than the company’s central identity. In the new narrative, Reality Labs work is built on the same AI and compute foundation as Meta’s other products. This makes it easier for the company to discuss Reality Labs in financial terms while reserving the main spotlight for AI, compute and core app performance.
Q: Should investors and observers keep tracking earnings call word counts like “compute 22, metaverse 0”?
Yes, with some caution. Word counts on earnings calls are not a substitute for financial analysis, but they are a useful lens on strategic priorities and messaging. When a company dramatically increases use of certain terms and drops others, it reveals what leaders want the market to focus on and what they would rather downplay. In Meta’s case, the shift from “metaverse” to “compute” reflects a deliberate attempt to reframe its big-spending narrative around AI infrastructure. Tracking how this language evolves over time can help investors understand how the company’s story—and possibly its strategy—is changing.
