Andrew 'Boz' Bosworth

Andrew 'Boz' Bosworth is CTO of Meta and Head of Reality Labs — one of the company's first ~10 engineers, he joined in 2006 when it had a few million users, and keeps a personal blog at boz.com on leadership and professional growth.

Boz joined Meta (then Facebook) in January 2006 when it was a sub-100-engineer startup with a few million users — he was roughly the tenth engineer through the door. Before that, he did a brief stint as a Software Design Engineer at Microsoft Visio from July 2004 to December 2005, right after graduating from Harvard with an AB in Artificial Intelligence in 2004. At Harvard he met Mark Zuckerberg, a connection that would define his career. Inside Facebook he moved through the stack: engineer on News Feed, Engineering Director overseeing Events, Places, Photos, Videos, Timeline, and Privacy, then Head of Ads and Business Platform, then VP of Augmented and Virtual Reality, Head of Reality Labs, and finally CTO — a title he holds today alongside the Reality Labs mandate. The through-line is a single bet made early and compounded for two decades: stay at the inflection point, own the hard problems, and grow with the company. He writes at boz.com on professional growth, technology, and career development, and is a sought-after podcast guest — 2024 appearances with Lenny Rachitsky on 'Making Meta,' Ben Thompson on Stratechery about Orion and Reality Labs, and Matthew Ball on metaverse, VR/AR, AI, and billion-dollar investment timelines. Possibly — the f7 Ventures fund attributed to him ($50 million pre-seed/seed focused on decade-defining entrepreneurs) reflects a side investing interest, though the founding year and current status are not fully sourced.

The most live story at Meta right now is a massive AI restructuring: in May 2026 the company cut approximately 8,000 employees, cancelled 6,000 open roles, and reassigned roughly 7,000 staff into AI-focused teams, all to fund a 2026 capital expenditure forecast of $125 billion–$145 billion — nearly double 2025's $72 billion capex. Revenue is not in question: Q1 2026 came in at $56.3 billion, up 33% year over year, and full-year 2025 hit $201 billion. On the product side, Meta launched new AI Glasses with EssilorLuxottica in July 2026, released Muse Spark 1.1 (multimodal reasoning model) and Muse Glimmer (a 30-billion-parameter open-weight model for local agent workflows) in July–August 2026, and introduced Muse Code, a terminal coding agent, in August 2026. The company is also testing subscription tiers — Meta One Plus at $7.99/month and Meta One Premium at $19.99/month — and launched an AI-powered Business Agent across WhatsApp, Instagram, and Messenger to diversify beyond advertising. Meta Connect is scheduled for September 23–24, 2026, where hardware and AI announcements are expected.

Meta is the dominant social advertising platform globally — 3.56 billion daily active users across its family of apps as of Q1 2026, and projected to surpass Google in global digital ad revenue for the first time by end of 2026. Its main rivals span multiple fronts: TikTok and YouTube Shorts in short-form video, Google and Amazon in digital advertising, Apple and Sony in hardware/VR, and OpenAI and Alphabet in AI model competition. The regulatory environment is the sharpest headwind: the EU's Digital Markets Act, ongoing FTC antitrust proceedings seeking to unwind the Instagram and WhatsApp acquisitions, a €1.2 billion GDPR penalty, and — most recently — China ordering Meta to unwind its $2 billion acquisition of AI startup Manus in April 2026, an unprecedented enforcement under China's Foreign Investment Security Review mechanism that has broader implications for cross-border AI M&A.

Boz's most important relationship is with Mark Zuckerberg — they met at Harvard and Zuckerberg is his direct principal as CEO; the two have collaborated continuously since 2006. In 2024 Boz was interviewed by Lenny Rachitsky (Lenny's Newsletter) and Matthew Ball, both of whom engage seriously with Meta's hardware and AI strategy, making them informed interlocutors in his extended network. Alexandr Wang, former Scale AI CEO now leading Meta's Superintelligence division after a $14.3 billion deal, is a significant new internal colleague.

  • Nearly two decades at Meta, from ~10th engineer through CTO — thinks in multi-year platform bets, not quarterly cycles.
  • Moved from News Feed engineer to Ads to AR/VR to CTO — broad functional range signals comfort with ambiguity and a preference for owning problems end-to-end rather than specializing narrowly.
  • Active public voice across boz.com, LinkedIn, and high-profile podcast appearances (Stratechery, Lenny's Newsletter, Matthew Ball) — comfortable being the public face of hard, contested decisions and responds well to direct, substantive questions.
  • Led Reality Labs through years of heavy losses and public skepticism before the AI pivot — suggests high tolerance for long-horizon, capital-intensive bets and resilience under external criticism.
  • Harvard AI background (2004) applied inside one company for his entire career — depth-over-breadth operator who goes very deep on a domain rather than collecting roles.
  • Possibly — the f7 Ventures side project suggests he enjoys mentoring early-stage founders, which may mean he engages more openly with builders than with analysts or press.

Conversation tips

  • Reference a specific post from boz.com or a line from the Stratechery Orion interview — he will immediately know you actually read it, not just the headline.
  • Come with a point of view on the Reality Labs / AI trade-off — he's spent years defending the long-term hardware bet and will engage more substantively with someone who has a real position than with neutral questions.
  • Don't ask him to justify the $125–145 billion capex in abstract terms — instead, ask what the ROI model looks like at the level of a specific product (e.g., AI Glasses daily active users, Muse Spark API adoption), which is where he'll have real numbers.
  • The Meta Connect timing (September 23–24, 2026) makes hardware and AI model roadmap a live topic — asking what 'Connect' is meant to signal to developers vs. consumers will get a more specific answer than general metaverse questions.
  • Open on the Muse Glimmer release — Meta just shipped a 30-billion-parameter open-weight model for local agent workflows under Apache 2.0, days before this meeting. It's a pointed bet: open-weight local inference as a counterweight to closed frontier models.
  • Reference the Stratechery interview on Orion — Boz went deep with Ben Thompson on Reality Labs investment timelines and the Orion AR platform, which is a rare level of transparency about a bet that's still losing money. It signals he's willing to defend the long-game in public.
  • Mention boz.com by name — he keeps a personal site on leadership and professional growth that predates the current CTO era. Naming it specifically shows you've done more than skim his LinkedIn.
  1. With Yann LeCun departing and Alexandr Wang now leading the Superintelligence division, how has the internal research culture shifted — is Meta more applied-product-first now, or still running a basic-research track in parallel?
  2. The Manus acquisition being unwound by Chinese regulators is the first enforcement of China's Foreign Investment Security Review in this way — how does that change how you think about cross-border AI talent acquisition going forward?
  3. Reality Labs is separating Quest and Horizon Worlds by June 2026 — what does that architectural split signal about where you think immersive computing actually lands for consumers in the next three years?

Don't lead with generic questions about 'the metaverse vision' — he's been defending that bet for years against sustained skepticism and will disengage quickly unless you bring a specific product, metric, or timeline to anchor the conversation.

Make it yours

Tailor these openers to what you sell

These openers are generic. Sign in and tell Brief what you sell — it rewrites the hooks and questions around your pitch.

Brief on your next meeting?

Type any name. Get a structured pre-meeting brief in seconds.

Try Brief →

Generated by briefthecall.com from public web sources on August 22, 2026. Each claim is linked to its source above.

Automatically generated by AI from public sources. May be inaccurate or out of date. Remove or correct this profile →