The race to build artificial general intelligence is no longer an abstract research contest. It is reshaping how the largest tech companies hire, spend, and talk about the future of the internet, and nowhere is that tension sharper than between Google and Meta. Their clash over talent, infrastructure, and ideology is turning the AGI push into a defining corporate rivalry that will influence how billions of people search, socialize, and work with AI.
At stake is not just who gets to claim the first working AGI system, but who sets the rules for how powerful models are built and deployed. As I see it, the fight now dividing Google and Meta is really a struggle over control: of engineers, of data centers, of safety narratives, and of the very definition of “open” AI.
AGI as the new holy grail
Artificial general intelligence, or AGI, has become the industry’s shorthand for a system that can match or outdo humans across a wide range of tasks, and both companies increasingly frame their long‑term strategies around that goal. In one widely shared description, AGI is described as the “holy grail of technology” that could outperform people and is already pushing leaders to think “far, far ahead” about what comes after today’s chatbots and copilots, a framing that captures how central AGI has become to Silicon Valley’s self‑image and ambition as they race toward AGI.
In that context, the rivalry between Google and Meta is less about incremental product features and more about who gets to define what AGI is for. Strategic scenarios already being discussed range from “Ethical Leadership,” where global norms and guardrails shape deployment, to more disruptive futures in which immersive platforms and AI agents transform how people socialize and experience reality, a set of possibilities that analysts have laid out when comparing OpenAI, Google, and Meta and asking what happens if Here the industry chooses different paths.
Meta’s talent offensive and Google’s alarm
One of the clearest flashpoints in the AGI contest is hiring. Meta has embarked on an aggressive campaign to recruit senior researchers and engineers from rivals, offering multimillion‑dollar packages to lure people away from established labs. The chief executive of Google DeepMind, Jul, has publicly complained that Meta is offering “millions” to attract AI talent and argued that the competition is “not just money,” a pointed remark that underscored how seriously Google’s leadership views Meta’s push and how directly it feels the impact of Meta on its own teams.
That frustration is not theoretical. Meta Poaches Top Engineers for AGI Team has become a recurring storyline, with one high‑profile example being Jack Rae, a principal researcher at Google DeepMind who confirmed he is moving to Meta to work on its AGI efforts. The fact that One of the engineers leaving is Jack Rae, a figure closely associated with cutting‑edge research, signals that Meta Poaches Top Engineers for AGI Team is not just about headcount but about transplanting institutional knowledge from Google into Meta’s new superintelligence push, a shift that has been documented as Meta Poaches Top Engineers for AGI Team gathers pace.
Lucrative paydays and what they reveal
Behind the scenes, the financial terms of this hiring war are reshaping expectations across the AI labor market. Reports that Meta Continues To Poach AI Talent From Competitors With Lucrative Paydays Because It Is Behind In The Race, Says Googl highlight that the company is willing to pay far above traditional compensation bands to close what it sees as a gap with Google and others. When a rival’s chief executive publicly characterizes those offers as evidence that Meta is “behind in the race,” it reinforces the perception that Meta is trying to buy time and expertise rather than relying solely on its existing research pipeline, a dynamic that has been captured in detail as Meta Continues To Poach AI Talent From Competitors With Lucrative Paydays Because It Is Behind In The Race, Says Googl.For Google, this raises uncomfortable questions about retention and culture. If Meta can peel away senior figures with cash, Google must decide whether to match those packages or lean on its own narrative about mission, safety, and long‑term impact. The tension is heightened by the fact that both companies are under pressure to ship new AI products quickly, and industry observers have warned that Tech companies that are leading the way in artificial intelligence are prioritizing products over research, a pattern that includes Meta, Google, and OpenAI and suggests that the same urgency driving these “lucrative paydays” may also be pulling attention away from deeper safety work, as seen in analyses of how Tech firms balance risk and speed.
Open source ideals, closed‑door pivots
Ideology is another front in the AGI fight. Meta has long positioned itself as a champion of open models, arguing that sharing code and weights can democratize AI and accelerate innovation. In a blog post that accompanied the release of Llama 2, the company aligned itself with Open‑source promoters and criticized what it saw as exaggerated fears about releasing powerful systems, with some voices around Meta suggesting that claims about catastrophic risks are “preposterously stupid,” a phrase that captured the sharp divide between those who favor open access and those who want tighter control over Open models.
Yet Meta’s stance has not been static. Aug brought a notable reversal when Meta just flipped on open‑source AI, with Zuckerberg shifting from declaring that “Open Source AI is the Path Forward” to signaling that Today the company will not open source its most advanced systems, citing safety concerns. That pivot, described as a move that “sounds like caution, but feels like consolidation,” shows how Meta, Zuckerberg, and the phrase Open Source AI is the Path Forward have become part of a more complicated story in which public commitments to openness collide with the commercial and regulatory realities of AGI, a tension laid bare when Meta recalibrated its message.
Search, social, and the battle for user attention
Beyond labs and licensing, Google and Meta are colliding in the products that sit in front of users. Google’s core business is search, and it is racing to integrate generative answers into that experience without cannibalizing the ads that fund it. Meta, for its part, is building AI systems that could challenge the very foundations of how people find information, embedding assistants into messaging apps and exploring AI‑driven discovery inside Facebook, Instagram, and WhatsApp. Analysts have argued that Meta is building something that could redefine the internet’s information layer and that this is Why Google Shoul be worried, because an AI‑driven search engine battle could erode Google’s dominance if Meta succeeds in turning its social graph into a new kind of Meta‑powered search.
In practice, that means AGI research is not just about abstract benchmarks but about who controls the interface where people ask questions and get answers. If Meta can use its AI to keep users inside its own apps for search‑like tasks, it chips away at the habit of “Googling” everything, while Google is trying to ensure that its own generative systems remain the default gateway to the web. The rivalry is especially intense on mobile, where a single tap to an AI assistant inside Instagram or a Gemini‑powered Chrome could determine whether Meta or Google owns the next generation of discovery, a contest that makes the AGI race feel very immediate for both companies’ revenue streams.
Superintelligence labs and the infrastructure land grab
To support their AGI ambitions, both companies are pouring staggering sums into infrastructure. Meta and Google tout aggressive AI infrastructure investments focused on data center builds and power, with each committing tens of billions of dollars in new investments for massive infrastructure buildouts that include specialized chips, new campuses, and long‑term energy contracts. These Meta and Google plans are not just about scaling today’s models but about preparing for far larger systems that could require entire new classes of data centers, a reality that has been underscored as Meta and Google unveil their buildout strategies.
Meta has gone a step further by reorganizing its AI efforts around a new internal structure. Meta Launches Superintelligence Labs in Major AI Restructuring, a move that concentrates its most advanced research into a dedicated group focused on superhuman systems. The announcement comes as Meta continues to invest billions in AI infrastructure, including specialized hardware and data centers needed to train increasingly sophisticated models, a combination that shows how seriously Meta is treating the path from today’s large language models to potential AGI and how it is trying to match or surpass Google’s long‑standing investments in its own global compute Meta footprint.
Cloud giants, power demands, and who can afford AGI
The infrastructure race is not happening in a vacuum. Google, Microsoft, AWS and others have continued investing billions in AI infrastructure and capabilities this year, a reminder that even as Google and Meta battle each other, they are also competing with other hyperscalers for GPUs, power, and enterprise customers. For Google, its cloud platform is a natural distribution channel for its models, while Meta, which does not run a public cloud at the same scale, must rely more heavily on partnerships and its own consumer apps, a structural difference that shapes how each can monetize AGI and how they position themselves among Google, Microsoft, AWS and the rest of the cloud field.
These capital requirements also act as a barrier to entry. Training frontier‑scale models requires not just money but access to reliable energy and advanced chips, which is why Meta and Google are locking in long‑term supply and building out new sites. The more they invest, the harder it becomes for smaller players to keep up, and the more the AGI race looks like a contest among a handful of firms that can afford to spend tens of billions of dollars per year on compute. That concentration of power is one reason regulators and researchers are increasingly focused on how these companies govern their systems and whether their internal safety processes are robust enough for the stakes involved.
Safety warnings and the “pull the plug” question
Inside and around these companies, there is a growing debate about how far and how fast to push. Former leaders have started to speak more bluntly about the risks of unchecked AGI development. One of the most striking interventions came from Google CEO Eric Schmidt in his post‑Google role, when he warned that in a worst‑case scenario society might need to “Pull the plug” on a system that behaves dangerously, a phrase that captured both the urgency and the uncertainty around how to govern models that could act in unpredictable ways and that has been widely shared as Pull the warning.
At the same time, critics argue that corporate incentives still tilt toward speed. Analyses of leading AI firms have found that Tech companies that are leading the way in artificial intelligence are prioritizing products over research, with Meta, Google, and OpenAI all cited as examples of organizations that ship new features faster than they publish safety work. That tension is visible in the way Meta reversed course on open sourcing its most powerful models and in the way Google balances its Gemini roadmap with calls for more rigorous evaluation, leaving both companies trying to convince regulators and the public that they can manage the risks of AGI even as they race each other to build it.
More from MorningOverview