In recent weeks, Meta Platforms announced a plan to launch “Meta Compute” to rent out excess AI capacity, accelerate production of its in-house Iris data-center chip with Broadcom and TSMC from September, and roll out advanced AI models like Muse Spark 1.1 and Muse Image alongside a new chief data officer appointment. These moves signal Meta’s push to turn its vast AI infrastructure and models into an enterprise-facing cloud and tooling business, even as it balances cost efficiency gains...