Meta Ex-CFO Dave Wehner Jumps to Fireworks, a $17.5 Billion AI Inference Startup Eyeing a $40 Billion Round

Dave Wehner, the executive who was Meta's chief financial officer through eight years of its most explosive growth, is leaving the company to become CFO of Fireworks, an AI inference startup valued at $17.5 billion in July. Wehner announced the move on X on Friday, capping a 14-year run at the company he joined in 2012, when it was still called Facebook.

Wehner served as Meta's CFO from 2014 to 2022, steering the company through the Instagram and WhatsApp integrations, the pivot to mobile advertising and the early Reality Labs bet, before moving to the chief strategy officer role. Meta COO Javier Olivan said in a statement to CNBC that Wehner "helped guide Meta through tremendous growth and complexity over 14 years, and brought disciplined thinking to some of our hardest company-level decisions."

His destination is a very different kind of company. Fireworks, founded in 2022 by former Meta and Google employees and led by CEO Lin Qiao, runs an inference and model-training platform that lets enterprises customize and serve open-weight models. In July it raised a $1.505 billion Series D led by Atreides Management, Index Ventures and TCV at a $17.5 billion valuation, with NVIDIA and Sequoia Capital among its backers. Business Insider reports the company is already in early talks for another round at a possible $30 billion to $40 billion valuation.

The startup is not a toy: by its own July disclosure, Fireworks had passed a $1 billion annualized revenue run rate and was processing more than 40 trillion tokens daily, over 95% of that from models specialized on customers' proprietary data. Those are company-reported figures and have not been independently audited, but if accurate they put Fireworks in the same conversation as the frontier labs it competes with, including OpenAI and Anthropic, at a fraction of the headcount.

The move is not without baggage. In October, Fireworks disclosed a security incident in which an unauthorized third party used internal credentials to access environment variables in evaluation jobs across a limited number of customer accounts. The company said its investigation found no evidence of unauthorized access to customer models, datasets, training data, personal information or inference traffic inside its security perimeter — a company statement, not an external audit. The episode has been cited in coverage of this autumn's chain of AI-industry breaches.

Wehner framed the jump as a return to builder mode. "Fireworks has the right mission for the moment," he wrote on X, adding that he had "become increasingly bullish on what they are building" and citing the startup's technical depth and customer focus. Qiao wrote on LinkedIn that Wehner told her he "wanted to get back to a smaller company where he could build toward a big impact again."

The signal is bigger than one resume line. When the finance chief who navigated a $1.56 trillion platform's hardest calls chooses to run finance and operations at a four-year-old open-model startup reportedly prepping a round at up to $40 billion, it reads as a verdict on where enterprise AI value is concentrating: the inference and serving layer, where open weights meet real customer workloads. Big Tech is no longer just the talent pool AI startups hire from — it is increasingly the place senior executives leave.