Search results for AI inference
AI & Enterprise
K-AI chips put to the test as global AI startups step up push into South Korea
South Korean neural processing unit makers are teaming up with domestic cloud companies to target the AI inference market, testing whether they can secure a foothold beyond Nvidia’s barriers. Samsung SDS launched an NPU-as-a-service offering on its Samsung Cloud Platform based on FuriosaAI’s second-generation chip. Rebellions is also partnering with KT Cloud and Gabia. Foreign AI startups are stepping up moves into Korea, including Genspark and China’s Kling AI.
Crypto
MEXC supports Bittensor TAO staking for global users
Cryptocurrency exchange MEXC is supporting staking for Bittensor’s native token, TAO. Cointelegraph reported that the integration allows MEXC users to participate in securing the Bittensor network and receive rewards. The service connects Yuma’s validator infrastructure to MEXC, and Yuma said 40 million MEXC users can access TAO staking. The companies said the integration aims to broaden access to the Bittensor ecosystem and increase network participation.
Crypto
RippleX engineering lead says AI agents buy and sell services with XRP on XRPL
Cases are rapidly increasing of AI agents on the XRP Ledger (XRPL) paying real service fees using XRP and Ripple\'s stablecoin RLUSD, as programs buy data, servers and inference services without human involvement. RippleX engineering lead Ayo Akinyele said AI agents trade technical services on XRPL. Transactions on t54.ai\'s on-chain dashboard hit 1,433,836, drawing attention. Uses include financial data, external information services, social-media APIs and AI inference services, with fees starting at 0.003 RLUSD per request.
-
AI & Enterprise
Samsung SDS teams up with FuriosaAI to launch Korean NPU-based NPUaaS
-
Industry
SpaceX orbital data centre plan may need up to 42 launches a day
-
AI & Enterprise
NIA conducts South Korea\'s first on-site trial of 5G AI-RAN and physical AI
-
AI & Enterprise
Altman revives debate on space data centres, industry says too early
-
Industry
FuriosaAI teams up with Equinix to target European AI infrastructure market
-
Industry
AI industry shifts toward in-house chips as Chinese firms join in
-
AI & Enterprise
IBM shrinks z17 mainframe to lower barrier to AI infrastructure
-
AI & Enterprise
Neocloud market shifts as SoftBank and Meta enter
-
AI & Enterprise
Dinotisia unveils KV cache compression technology to ease AI computing bottlenecks
-
AI & Enterprise
Qualcomm unveils \'High Bandwidth Compute\' as HBM alternative, enters AI chip race
-
AI & Enterprise
Nvidia reintroduces RTX 3060 amid surging memory prices, still top on Steam
-
AI & Enterprise
Logpresso, FuriosaAI partner on sovereign AI-based agentic SOC
-
Telecommunications & Media
MSIT to launch AI platform for compound disaster response with MOIS and MOLIT
-
AI & Enterprise
Rebellion to acquire Squeezebits, expand with AI inference optimisation
-
AI & Enterprise
Neologic unveils AI inference-focused server CPU technology at VivaTech 2026