Meta’s Infrastructure Team Needs A Culture Reset
SubscribeSign in
Meta’s Infrastructure Team Needs A Culture Reset<br>Meta Infrastructure has become bloated, with middle managers expending resources on over-engineered technology solutions that lose sight of broader organizational needs.<br>Wayne Ma, Myron Xie, Julien Martin-Prin, and 2 others<br>Jul 22, 2026<br>∙ Paid
112
Share
In our recent newsletter piece about Meta Superintelligence, we expressed reasons to be optimistic on Meta AI. MSL now has many of the right ingredients to catch up with Anthropic and OpenAI to return to the frontier. However, we also briefly alluded to cultural issues plaguing Meta’s infrastructure teams. This article will dive into how these cultural issues have manifested into expensive missteps, whether it be with acquisitions like Rivos or strange choices on hardware architecture.<br>We believe that Meta Infrastructure needs a cultural reset to better serve the Meta AI organization, especially the world class researchers who are at MSL. This is even more important as Meta embarks on the path of selling its compute to outside customers, not just serving captive internal users.<br>Meta Infrastructure has become bloated, with middle managers expending resources on over-engineered technology solutions that lose sight of broader organizational needs. The company appears burdened by far too many disparate groups that are over-optimizing for certain metrics as opposed to delivering usable technology for the company as a whole. Middle managers will do everything to justify their proposals to protect their positions within Meta, which has become an extremely political organization.<br>One big issue is Meta’s six-month performance review cycle, in which the bottom 10% to 15% are cut every review round. The result is an organization of employees that optimize for short-term wins rather than long-term strategies. Some managers push for highly visible projects that can be delivered quickly, a practice known as “window washing” and then promptly pivot or abandon them. Few openly challenge leadership, which leads to bad decisions going uncorrected. The whole system discourages long-term thinking and leads to risk-adverse behavior.<br>Within Meta Infrastructure, supply chain teams also have little say over engineering teams. The result is technology decisions driven by political motivations rather than thoughtful software/hardware co-design for the broader company.<br>Frequent pivots are also common. And because Meta has a reputation for throwing money at problems and executing at high speed, these U-turns end up becoming more costly versus other companies that take a more disciplined or conservative approach. Suppliers also lose faith when given design wins are later cancelled. This has lead to less supply chain prioritization on new designs. Some suppliers favor focusing on Amazon or Google designs due to Meta’s frequent reshuffling.<br>A lot of Meta’s issues come from a lack of financial discipline, with managers creating new projects and headcount to fill out and justify the massive budgets given for initiatives like AI. It’s a repeat of what happened to Meta’s Reality Labs division, where billions of dollars were spent on engineers and R&D until layoffs, partly brought on by the company’s pivot to AI, slashed the team and their related projects starting in 2022 and continuing through this year.<br>Rivos Acquisition
First, one of Meta’s most recent missteps has been the more than $2.5 billion it sunk last year into the acquisition of Rivos, a chip startup.<br>Few inside Meta’s chip division have a full understanding of why the company bought Rivos in the first place, and those who championed the deal internally have since gone quiet. The prevailing theory is that Meta had the money, the custom silicon space was heating up, and it was already licensing Rivos’ IP for a future chip, so leadership figured it might as well own Rivos’ technology outright than let anyone else have it.<br>With the Rivos acquisition, Meta also gained the ability to bypass partners like Broadcom to manage the manufacturing and testing of their own custom chips, a practice known as customer-owned tooling. However, paying more than $2.5 billion for that privilege doesn’t make much sense either given that a COT team could be built from scratch for maybe $100 million a year or more.<br>Although Meta wanted only Rivos’ accelerator and GPU team, the startup’s founders insisted on an all-or-nothing deal. Meta bought the entire company and then heavily cut employees in the parts it didn’t want.
Source: SemiAnalysis, Rivos<br>Some former Meta chip employees say the acquisition was led by Meta’s silicon chief, Yee Jiun Song, who pushed for the deal against the wishes of some of those under him but has since lost interest. The result is an organization that has rejected the transplant, leaving Rivos staff stuck to fend for themselves. Existing Meta chip managers treated the acquisition as a pool of free...