Google Cloud Next 18 Highlights: TPU 3.0, AutoML Breakthroughs, and AI Strategy

Google Cloud NEXT 18 took place in San Francisco from July 24‑26, showcasing the latest advances in Google’s AI‑focused cloud platform.

CEO Diane Greene emphasized that AI and security are the company’s biggest investment areas, with AI offering the greatest opportunity.

Chief Scientist Fei‑Fei Li announced three major releases:

Cloud TPU 3.0 entered alpha testing, promising up to eight times the performance of the previous generation and up to 100 petaflops; early adopters such as eBay and a medical‑genomics partner reported training time reductions from months to days and from a week to one day, respectively.

AutoML expanded with AutoML Natural Language and AutoML Translation, enabling users without machine‑learning expertise to build language and translation models, furthering the goal of AI democratization.

Contact Center AI, an industry solution combining a virtual assistant with intelligent information extraction and sentiment analysis, aims to improve contact‑center efficiency, though it currently lacks Chinese language support.

Additional updates were made to Cloud Vision API, Cloud Text‑to‑Speech API, and Cloud Speech‑to‑Text API.

AutoML now serves over 18,000 registered users across media, retail, finance, insurance, energy, healthcare, and environmental sectors.

Google Cloud’s financial performance highlighted rapid growth: Q2 “other revenue” reached $44.25 billion, a 37 % year‑over‑year increase, outpacing the advertising business, while capital expenditures rose to $54.8 billion, nearly double the prior year, underscoring the strategic focus on cloud expansion.

Alphabet CFO Ruth Porat noted cloud computing as a key growth engine for Google, linking it to future search, advertising, and machine‑learning initiatives.