Alibaba on Thursday introduced Qwen VLo, a multimodal model that generates images from text or existing pictures and displays each step of the creative process through so-called progressive generation. The release expands the Hangzhou-based company’s Qwen family of large language models and underscores Chief Executive Officer Eddie Wu’s declaration that developing artificial general intelligence is Alibaba’s “primary objective.” Analysts position Qwen VLo as a direct challenger to OpenAI’s GPT-4o and Google’s Imagen 4. Commentators who have tested early versions say the model matches GPT-4o-level creativity while working across multiple languages. Its debut follows Tencent’s launch of Hunyuan-A13B, a compact model that blends rapid inference with advanced reasoning on a single graphics processor, highlighting Chinese vendors’ rapid progress in generative AI. The competitive pressure extends beyond model releases. In the United States, Meta Platforms has spent US$14.2 billion for a 49 percent stake in data-labeling firm Scale AI and recruited four senior researchers from OpenAI, luring them with signing bonuses that Chief Executive Sam Altman says exceed US$100 million each. Meta CEO Mark Zuckerberg is assembling a new team to pursue so-called superintelligence after his Llama 4 model underperformed peers earlier this year. Industrywide, capital outlays are soaring. Meta, Microsoft, Amazon and Google together expect to spend about US$320 billion on infrastructure this year, more than double 2023 levels, as they race to secure data-centre capacity and top talent. The flurry of product launches and billion-dollar hiring packages signals that large technology groups view overspending as less risky than being overtaken in the next phase of AI development.