Apple and Google have officially confirmed a multi-year collaboration under which the next generation of Apple Foundation Models will be based on Google’s Gemini models and cloud technology, enabling future Apple Intelligence features, including a significantly upgraded Siri, to be launched later this year.

This partnership marks a major strategic shift for Apple, which had previously relied on OpenAI’s ChatGPT for some AI functions, and signals a decisive move to leverage Google’s advanced AI capabilities to accelerate its own AI ambitions while maintaining strict privacy standards.

The collaboration centers on Apple using Google’s Gemini models as the foundational technology for its own Apple Foundation Models, which will power a range of AI-driven features across Apple’s ecosystem.

The new Siri, expected to debut in the OS 26.4 update in March or April 2026, will be built on this Gemini foundation, offering enhanced personalization, in-app actions, and on-screen awareness.

Despite the deep integration of Gemini technology, Apple will customize the models for its own use, meaning the resulting Siri will not be identical to Google’s Gemini or Google Assistant.

Apple emphasized that all Apple Intelligence features, including the new Siri, will continue to run on Apple devices and its Private Cloud Compute architecture, ensuring user data remains protected and does not flow to Google’s servers.

A joint statement by both companies stated that after a thorough evaluation, Google’s AI technology was determined to provide the most capable foundation for its models, a decision that has been viewed positively by Wall Street and has contributed to Alphabet’s market capitalization surpassing $4 trillion.

"Apple and Google have entered into a multi-year collaboration under which the next generation of Apple Foundation Models will be based on Google's Gemini models and cloud technology," the statement reads on X. "These models will help power future Apple Intelligence features, including a more personalized Siri coming this year.

"After careful evaluation, Apple determined that Google's Al technology provides the most capable foundation for Apple Foundation Models and is excited about the innovative new experiences it will unlock for Apple users. Apple Intelligence will continue to run on Apple devices and Private Cloud Compute, while maintaining Apple's industry-leading privacy standards."

The deal follows speculation that Apple might pay Google around $1 billion annually to use its models, though the financial terms of the agreement have not been publicly disclosed.

This partnership builds on Apple’s existing relationship with Google as the default search engine on its devices, a long-standing arrangement that generates significant revenue for Apple.

Commenting on the collaboration, xAI and Tesla CEO Elon Musk wrote on X, "This seems like an unreasonable concentration of power for Google, given that the also have Android and Chrome."

A SpaceX Falcon 9 rocket launched the first "Twilight" rideshare mission on Sunday, (Jan. 11), from Vandenberg Space Force Base’s Space Launch Complex 4 East at 13:44 UTC, carrying  40 payloads into a dusk-dawn sun-synchronous orbit, a trajectory that allows satellites to maintain consistent lighting conditions for observations.

The Falcon 9’s first stage booster (B1097), completed its fifth flight and successfully landed back at Landing Zone 4 (LZ-4) at Vandenberg Space Force Base about 8.5 minutes after liftoff. This marked the 32nd landing at LZ-4 and the 557th booster landing in SpaceX’s history. The first stage had previously launched three batches of Starlink V2 Mini Optimized satellites and the Sentinel-6B spacecraft.

The 40 payloads were deployed over a 90-minute window, beginning about 61 minutes after liftoff and concluding more than 2.5 hours into the mission. The mission was named "Twilight" due to its deployment into a dusk-dawn sun-synchronous orbit, which straddles the terminator—the line between day and night on Earth—providing stable lighting conditions ideal for long-duration observations.

The primary payload was NASA’s Pandora exoplanet observatory, a 325-kilogram small satellite developed under NASA’s Astrophysics Pioneers program and led from Goddard Space Flight Center, with mission operations based at the University of Arizona in Tucson.

Pandora is equipped with a 0.45-meter (17-inch) Cassegrain telescope developed jointly by Corning Incorporated and Lawrence Livermore National Laboratory, and its spacecraft bus was supplied by Blue Canyon Technologies.

The mission’s core objective is to disentangle exoplanet atmospheric signals from stellar variability by repeatedly observing starlight passing through planetary atmospheres during transits. It will conduct long-duration observations—approximately 24 hours per visit—on at least 20 known exoplanets, with some program materials suggesting up to 39 planets could be studied over a broader timeline.

Most of the more than 6,000 alien planets we know of have been discovered via the "transit method."

Transits also allow astronomers to characterize known exoplanets, especially their atmospheres. Different elements and molecules absorb light at specific wavelengths, so studying the spectrum of starlight that has passed through an atmosphere can reveal a great deal about that atmosphere's composition.

However, such work is complicated by stellar complexity. Star surfaces are not uniform; they often feature patches of varying brightness, like the sunspots that speckle our own star. Pandora will help astronomers account for such complexity, if all goes to plan.

Daniel Apai of the University of Arizona, a lead scientist on the Pandora team, described the mission as "a bold new chapter in exoplanet exploration," emphasizing its potential to refine techniques for detecting atmospheric signals and guide future searches for habitable worlds.

"Pandora aims to disentangle the star and planet spectra by monitoring the brightness of the exoplanet's host star in visible light while simultaneously collecting infrared data," NASA officials wrote in a mission description. "Together, these multiwavelength observations will provide constraints on the star's spot coverage to separate the star's spectrum from the planet's."

Tjhe satellite will focus on planets with atmospheres that are dominated by water or hydrogen, agency officials added. The mission includes a one-month commissioning phase followed by a one-year primary science campaign, with all data to be made publicly available.

The mission’s $20 million budget cap reflects its role as a cost-constrained, high-impact experiment designed to complement larger observatories like the Kepler and James Webb Space Telescope (JWST).

The Twilight mission also carried a diverse array of secondary payloads. Among them were three NASA CubeSats: the Star-Planet Activity Research CubeSat (SPARCS), developed by Arizona State University, and the Black Hole Coded Aperture Telescope (BlackCAT), operated by researchers at Penn State University.

Additionally, the mission included 10 Aether spacecraft from Kepler Communications, which form part of a low Earth orbit optical data relay network. Capella Space contributed two advanced Acadia Earth-imaging radar satellites for high-resolution imaging. Exolaunch managed the deployment of 22 satellites, including Dcubed-1/Araqys-D1, which will demonstrate on-orbit 3D printing of a boom structure.

Other payloads included nine Lemur satellites from Spire Global for weather and maritime monitoring.

Sunday's mission is part of SpaceX’s broader rideshare program, distinct from its Transporter and Bandwagon series, and marks the first dedicated flight under the Twilight banner. To date, the company has launched 15 such flights in its Transporter series and four via Bandwagon.