Andrew 'Boz' Bosworth

Who they are

Andrew 'Boz' Bosworth is CTO of Meta — joined Facebook in 2006 when it was a pre-IPO social network and has been Zuckerberg's technical right hand through every major platform shift since.

Person

Boz joined Facebook in 2006 when it was still a pre-IPO social network — and he's been there ever since, one of the longest-tenured executives in the company's history. He came out of Harvard in 2004 with an AB in Artificial Intelligence, then spent about 18 months as a Software Design Engineer at Microsoft before Facebook pulled him in. He rose through the VP ranks, overseeing large pieces of the ads and hardware businesses, before being named CTO in 2022. The through-line is a technically-trained operator who chose depth over breadth — two decades at one company, building through every major transition from News Feed to VR to the current AI pivot. He writes at boz.com, covering professional growth, technology strategy, and leadership — practical rather than philosophical. In 2024 alone he did extended interviews with Ben Thompson on Stratechery (on Orion and Reality Labs), Matthew Ball on the metaverse and VR/AR investment timelines, and Lenny Rachitsky on how Meta is built — a level of external visibility that signals someone who shapes narrative, not just ships product.

Company

The most pressing current moment at Meta is a massive restructuring: in May 2026 the company laid off approximately 8,000 employees and cancelled 6,000 open roles, shifting headcount into AI teams. This came alongside Meta raising its 2026 capital expenditure forecast to $125 billion–$145 billion — nearly double the $72 billion spent in 2025 — for AI infrastructure including custom MTIA chips and server farms. The financial backdrop is strong: Q1 2026 revenue hit $56.3 billion (33% YoY growth), and full-year 2025 revenue came in at $201 billion. Meta is also testing paid AI subscription tiers — Meta One Plus at $7.99/month and Meta One Premium at $19.99/month — powered by its Muse Spark model, and launched an AI-based creator assistant in June 2026. A $25 billion bond issuance in April 2026 raised total debt to $83.7 billion, reflecting how aggressively the company is financing this AI build-out.

Market

Meta is the dominant global social media advertising platform — 3.54 billion daily active users across its Family of Apps, commanding 67.3% of global social media ad spend, and projected to surpass Google in total digital ad revenue for the first time by end of 2026. Its primary AI competitors are Alphabet, Microsoft, and OpenAI; in social/short-form video, TikTok (ByteDance) remains the sharpest rival despite U.S. regulatory uncertainty, which is why Meta launched a standalone Reels app in March 2026. Regulatory headwinds are significant: EU antitrust probes on WhatsApp's AI gating, lawsuits over Llama training data, and China's April 2026 order to unwind Meta's $2 billion acquisition of AI startup Manus — an unprecedented enforcement under China's Foreign Investment Security Review mechanism — all add friction to Meta's cross-border AI expansion.

Network

Bosworth's most visible relationship is with Mark Zuckerberg — he reports directly to the CEO and is consistently described as Zuckerberg's technical right hand, including through the current AI restructuring. In 2024 he engaged publicly with three prominent tech journalists and analysts — Ben Thompson (Stratechery), Matthew Ball, and Lenny Rachitsky — suggesting he is comfortable being the external voice for Meta's most contested strategic bets. Alexandr Wang, former Scale AI CEO, now leads AI integration inside Meta after Meta's $14.8 billion Scale AI investment in June 2025.

Mark Zuckerberg· CEO, Meta
Alexandr Wang· Head of AI Integration (former Scale AI CEO), Meta
Ben Thompson· Founder, Stratechery
Lenny Rachitsky· Founder, Lenny's Newsletter
Matthew Ball· Author and analyst, matthewball.co

How they likely show up

Long tenure at Meta (joined 2006, now CTO — a multi-decade operator track) → thinks in company-scale arcs, not product cycles; unlikely to be impressed by short-term framing.
Harvard AB in Artificial Intelligence in 2004 → the AI pivot isn't a new language he's learning; it's the domain he trained in, which probably makes him impatient with surface-level AI commentary.
Three high-profile external interviews in 2024 alone (Thompson, Ball, Rachitsky) → comfortable on the record, uses public appearances to shape narrative; responds well to substantive, well-researched interlocutors.
Rose from engineer to CTO at the same company → operator disposition, not a generalist executive; will probe the technical specifics behind any claim.
Writes at boz.com on professional growth and leadership → he processes and codifies his thinking in public, which means his views on management and technology are documented; referencing a specific post signals real preparation.
Oversees Reality Labs and the AI pivot simultaneously → used to defending long-duration, high-capex bets against skepticism; probably has a practiced, structured answer for ROI-timeline questions.

Conversation tips

→ Reference a specific post from boz.com — he'll immediately know whether you've actually read it or are pattern-matching on his name.
→ Come prepared on the AI infrastructure economics: the capex jump from $72 billion in 2025 to $125–$145 billion in 2026 is the defining tension in his job right now, and he'll want to talk to people who understand the scale.
→ Ask about the tradeoffs between open-weight Llama and the closed Muse Spark model — it's a live strategic question with real stakes and he's been publicly articulate about it.
→ Don't conflate the Reality Labs story with the AI story — he runs both, but they're distinct bets with different timelines and he'll notice if you blur them.
→ He's a two-decade insider; don't present conventional wisdom about Meta as insight — he has more context than almost anyone alive on how the company actually works.

Toolbox

Openers

Open on the Stratechery interview about Orion and Reality Labs — he gave Ben Thompson a detailed technical breakdown of the bet, and engaging with a specific claim from that conversation signals you've done the work.
Reference the June 2026 creator AI assistant launch — it's one of the freshest product moves under his technical leadership and opens a direct conversation about where agentic AI fits into Meta's platform stack.
Bring up the Manus acquisition being unwound by Chinese regulators in April 2026 — it's the kind of geopolitical constraint that directly shapes what a CTO can build, and he's unlikely to have been asked about it by many people outside Meta.

Discovery questions

Meta is running Llama as open-weight and Muse Spark as closed — how do you decide which side of that line a model belongs on, and does that calculus change as capabilities increase?
You've overseen Reality Labs through a period where capex stayed high and timelines extended — how do you think about maintaining institutional conviction on a bet that the market prices as a drag?
The $125–$145 billion capex guidance for 2026 is an extraordinary number — at that scale, what does a good ROI signal actually look like, and over what time horizon are you measuring it?

Avoid

Don't treat the AI pivot as a recent conversion — he studied AI at Harvard in 2004 and has been building toward this for two decades; framing it as a new direction for Meta will read as uninformed.

Make it yours