ByteDance just launched Seedream 5.5🚀 Outperforming GPT Image 2.5 Sunburst & Nano Banana 2.1 in both text rendering fidelity and visual realism. What's new: • Supports continued work on the same image across 100 rounds of editing. • Native transparent/translucent background export (Alpha channel) • Doubled usability for cinematic and office use cases vs. 5.0 Pro Seedream 5.5 is gradually rolling out across Dreamina, Doubao, Doubao Work, and Volcano Engine Ark.

ByteDance is reportedly considering a Brazil data center expansion that would go beyond its existing $10B, 300 MW project in Ceará, on top of the tens of billions it has already announced there. Brazil's data center capacity has passed 1 GW, doubling since 2022, and Brasscom expects about 20% annual growth through 2031, which means more demand for chips and servers from Nvidia $NVDA and $AMD. The AI buildout is far from over, and the catch is power, with a Brazilian energy lawyer warning the grid can't take all the plants trying to connect.

ByteDance Seed's Cola2 spotted on GitHub: an LM that diffuses latents first, tokens second it groups text into semantic spans, a diffusion transformer generates one latent per span, a decoder turns it into tokens 346.03B tokens, 60.10 avg on 8 tasks. weights coming soon

ByteDance released Seedance 2.5 4K, featuring richer visual detail. > Seedance 2.5 4K offers richer color gradations and smoother transitions thanks to 10-bit color depth. > Users can use Draft mode to experiment with low-cost 480p previews and then upscale them to 4K.

ByteDance, Alibaba and Tencent have ordered more than 3 million of Nvidia's new RTX PRO 5500 graphics cards, valued at over $35 billion, two sources said, in a rush for a less powerful but export-compliant chip amid US restrictions on trade with China

ByteDance could become MediaTek's next major AI chip customer after Google. Bloomberg reports advanced talks for custom data-center chips, though no deal is confirmed. Another huge AI buyer is looking beyond GPUs to build silicon tailored to its own workloads.