守法公民

Pres Biden is naming 'tech hubs' for 32 states, Puerto Rico to help industry_我的网站

尼基塔第三季

一 | AI浪潮正在重塑全球劳动力结构，但对科技硬件与半导体行业而言，这场变革带来的是净收益而非冲击——算力需求的持续扩张，正将供给侧的结构性约束推向前台。追风交易台消息，据摩根士丹利最新发布的AI采用与未来工作调查报告，科技硬件与半导体行业在过去12个月录得约8.6%至8.2%的净生产率提升，同时自身也面临约5%至8%的净岗位净流失。 REHOBOTH BEACH, Del. -- The Biden administration is designating 31 technology hubs touching 32 states and Puerto Rico to help spur innovation and create jobs in the industries that are concentrated in these areas.President Joe Biden is set to announce the hubs on Monday at the White House with Commerce Secretary Gina Raimondo. “I have to say, in my entire career in public service, I have never seen as much interest in any initiative than this one," Raimondo told reporters during a Sunday conference call to preview the announcement. Her department received 400 applications, she said. “No matter where I go or who I meet with — CEOs, governors, senators, congresspeople, university presidents — everyone wants to tell me about their application and how excited they are,” said Raimondo.The tech hubs are the result of a process Raimondo's department launched in May to distribute a total of $500 million in grants to cities.The $500 million came from a $10 billion authorization in last year’s CHIPS and Science Act to stimulate investments in new technologies such as artificial intelligence, quantum computing and biotech. It’s an attempt to expand tech investment that is largely concentrated around a few U.S. cities — Austin, Texas; Boston; New York; San Francisco; and Seattle — to the rest of the country.The program, formally the Regional Technology and Innovation Hub Program, ties into the president's economic argument that people should be able to find good jobs where they live and that opportunity should be spread across the country, rather than be concentrated. The White House has sought to elevate that message and highlight Biden's related policies as the Democratic president undertakes his 2024 reelection bid. “These Tech Hubs will catalyze investment in technologies critical to economic growth, national security, and job creation, and will help communities across the country become centers of innovation critical to American competitiveness,” the White House said Monday in an emailed statement.The 31 tech hubs reach Oklahoma, Rhode Island, Massachusetts, Montana, Colorado, Illinois, Indiana, Wisconsin, Virginia, New Hampshire, Missouri, Kansas, Maryland, Alabama, Pennsylvania, Delaware, New Jersey, Minnesota, Louisiana, Idaho, Wyoming, South Carolina, Georgia, Florida, New York, Nevada, Missouri, Oregon, Vermont, Ohio, Maine, Washington and Puerto Rico.。然而，该行分析师明确指出，这一行业是AI浪潮的"净受益者"——既通过内部运营效率改善降低成本，又因外部AI基础设施需求激增而直接受益于芯片及相关设备的订单增长。

二 | 摩根士丹利分析师认为，AI模型效率的持续提升将进一步刺激企业加大AI基础设施投入，从而为半导体需求提供持续支撑。下一阶段AI硬件的真正瓶颈将从芯片设计转向能源、电力、数据中心容量及制造产能，这些供给约束或成为AI基础设施扩张的关键限制。硬件与半导体：AI的"镐与铲" 摩根士丹利报告将科技硬件与半导体行业定性为AI生态中提供"镐与铲"的核心供给方。

三 | 调查数据显示，该行业过去12个月净岗位损失约为5%（科技硬件）和8%（半导体），但这一数字并不代表行业受损——裁员主要集中在早期阶段，且整体行业影响为正面。从具体应用场景来看，欧洲半导体企业的AI落地已相当多元。ASML将AI用于计算光刻解决方案的精度与效率提升，STMicroelectronics则应用于缺陷分类与良率优化。此外，23%的英国和德国科技硬件受访企业表示已部署制造设施数字孪生技术，供应商涵盖EDA领域的Cadence和Synopsys，旨在缩小仿真与现实之间的差距。在人员结构调整方面，欧洲半导体企业的裁员尚处早期阶段。 ASML、Infineon和Aixtron均于2026年宣布裁员计划，但管理层并未将此直接归因于AI。相对明确的案例是ams OSRAM，该公司于2026年2月宣布计划在未来三年削减逾2000个岗位，明确将欧洲业务的AI自动化列为裁员理由之一，同时将部分劳动力向亚洲转移。

四 | AI采用加速推动芯片订单持续扩张摩根士丹利报告的核心逻辑在于：AI效率收益越显著，企业对AI基础设施的投入意愿就越强，进而形成对半导体需求的正向循环。调查数据显示，全行业平均净生产率提升为9.6%，其中银行、软件等高强度AI采用行业的生产率增益尤为突出，这些行业的持续投入将直接转化为对算力的刚性需求。从资本开支角度观察，摩根士丹利援引的数据显示，摩根大通2026年科技预算接近200亿美元，其中约23亿美元（占总投资预算的25%）直接用于AI投资。这一量级的企业级AI支出，最终将沿供应链传导至芯片设计、晶圆制造及半导体设备环节。摩根士丹利分析师在报告中进一步指出，随着AI从生成式模型向智能体（Agentic AI）演进，CPU、半导体设备及存储器将成为关键受益方向。

五 | 智能体AI的大规模部署对算力密度、推理效率及内存带宽提出更高要求，这意味着现有硬件架构面临新一轮迭代压力，也预示着下一个供给瓶颈正在形成。

六 | 能源、算力与供应链约束尽管需求前景明朗，摩根士丹利的调查也揭示了制约AI硬件扩张的多重结构性障碍。在软件行业受访者列出的未来12个月主要挑战中，"基础设施与能源约束（算力容量、电力可用性）"被明确提及，这一问题同样适用于为AI工作负载提供底层支撑的硬件供应商。从供应链层面看，调查显示"遗留系统集成"和"数据准备度"是各行业AI部署的普遍痛点，这对需要将AI嵌入既有制造流程的硬件企业构成额外挑战。与此同时，"AI技能短缺"在多个行业均位列前三大障碍，半导体行业高度依赖专业工程人才的特性使其在人才竞争中面临更大压力。澳大利亚软件行业的观察提供了另一个维度的参照：随着AI采用向更广泛部署推进，对数据中心和算力基础设施的需求将显著增强，但资本密度和电力强度也将同步上升。这一判断对全球半导体及硬件投资者均具有直接参考价值——下一阶段的核心瓶颈，或许不在于芯片设计本身，而在于支撑大规模AI推理所需的能源供给与制造产能。