内容来源：https://www.geekwire.com/2025/go-inside-zap-a-seattle-area-company-trying-to-build-a-star-in-a-jar-to-unlock-abundant-clean-energy/

内容总结：

【本报综合报道】在美国华盛顿州埃弗里特市，扎普能源公司的七人工程师团队正在一间布满巨型计算机显示屏的实验室内，开展一项仿若"将恒星装入容器"的前沿研究——核聚变能源开发。近日，该公司向媒体展示了其最新研发进展。

在犹如航天发射指挥中心的实验室内，工程师们通过精密协作启动核聚变装置。随着倒计时结束，室外传来等离子体成功激发的闷响，伴随紫色闪光掠过。这套被研究人员形象称为"驯服闪电"的系统，本年度已成功实现连续产生超1000个等离子体的突破，迄今累计完成上万次不同配置下的等离子体生成实验。

与采用超导磁体或激光约束的主流技术路线不同，扎普公司独创通过超高电流形成自约束磁场的技术方案。其研发副总裁本·莱维特解释："这种自组织等离子体能在微秒级时间内自主生成并消散磁场"。该技术使反应堆体积仅相当于家用热水器，兼具成本优势与扩展潜力。

目前全球约45家企业正竞相攻克核聚变商用难题。随着数据中心与人工智能产业对清洁能源需求激增，该领域近年已获数十亿美元投资，包括比尔·盖茨旗下基金及OpenAI首席执行官萨姆·奥尔特曼支持的相关企业。

扎普公司采用双轨研发策略：一组团队专注提升等离子体性能，另一组同步攻关发电系统集成。其最新一代"世纪"系统已实现39千瓦的功率输出，正朝着100千瓦、1兆瓦直至商用所需的10兆瓦目标迈进。尽管当前放电频率已提升至每5秒一次，但距离最终商用所需的每秒10次仍有技术瓶颈待突破。

这家成立于2017年的企业迄今累计获得3.3亿美元投资，其发展轨迹折射出全球清洁能源竞赛中创新科技与资本深度融合的新趋势。

中文翻译：

华盛顿州埃弗里特市——一间排列着巨型计算机显示屏的静谧房间里，七名扎普能源公司的工程师正准备在一台演示型核聚变装置中生成超高温等离子体。如同航天发射任务控制中心，操作员逐一向队友确认系统状态。一名工程师监控着黑色集装箱内的电容器——这些设备从电网充电并向聚变反应堆输送巨大能量脉冲；另一名工程师则确认冷却反应堆堆芯的银色液态金属循环正常。每个部件都必须精准配合。

"启动倒计时：三、二、一。"
"开始充能。"

室外传来等离子体成功点燃的"砰"声，随即闪过一道紫光。扎普公司是太平洋西北地区四家核聚变企业之一，全球约45家正执着攻坚同一前无古人的目标：在地球上复现太阳与星辰的能量反应，获取清洁近乎无限的电力。

今年初，这家总部位于埃弗里特的公司创下新里程碑：三小时内连续生成超一千个等离子体——这是核聚变必需的物质状态。此后，扎普的"世纪"系统在不同配置下已完成上万次等离子体生成实验。在近日的参观中，工程师向少数记者展示了这项技术。

每次运行都能获取更多数据，让科学向"瓶中之星"的目标更近一步——研发副总裁本·莱维特如此形容该反应。

『受控闪电』
核聚变概念本身简明：反应堆产生包含离子的等离子体，使其达到足够高温、稠密且稳定的状态，迫使本不相容的原子结合释放能量。但物理学家耗费数十年仍未实现能量净增益，无人知晓能否或何时能突破。

尽管存在长期不确定性，数据中心与人工智能运营导致的电力需求激增，仍使这种清洁能源重获关注，近年该领域已吸纳数十亿美元投资。科技巨头纷纷注资：克里斯·萨卡的Lowercarbon资本与比尔·盖茨的突破能源基金投资扎普，而OpenAI首席执行官萨姆·奥尔特曼则支持同处埃弗里特的赫利昂能源公司。

物理学家正用不同类型反应堆追逐聚变梦想，或采用高功率磁体，或使用激光生成约束等离子体。扎普的方案是让超强电流穿过反应堆中的等离子体，形成压缩物质的磁场。"这简直就是被驯服的闪电，"系统工程副总裁马修·汤普森表示。实际上其强度更甚——等离子体内电流强度达到普通闪电的20倍。

扎普团队宣称其方案更具成本效益与扩展性，因占地面积远小于多数方案。反应腔仅热水器大小，且无需超复杂磁体与激光。"这是所谓的自组织等离子体结构，用自身磁场实现约束，"莱维特解释，"它仅在需要时的几微秒内生成磁场，随即消失。"

双轨并行
扎普公司2017年脱胎于华盛顿大学研究项目，已获3.3亿美元投资及1300万美元美国能源部拨款，现有员工150人。研发采用双轨制：部分团队专注优化等离子体与聚变反应堆，另一团队则整合发电系统组件——既为反应堆供能，又捕获聚变能量输入电网。

"这是并行工程，"莱维特说明，"在完善等离子体性能的同时，同步推进电站技术研发。"2024年投用的"世纪"系统已将等离子体生成平均功率提升20倍至39千瓦，下一目标是100千瓦，继而1兆瓦，商业级系统需达10兆瓦。

速度是另一挑战。"世纪"初始每10秒发射一次等离子体，现速率已翻倍。但进展需呈指数级增长而非线性。"最终电站需每秒发射10次，速度要快得多，"莱维特坦言，"不过我们目前暂不聚焦于此。"

英文来源：

EVERETT, Wash. — In a quiet room lined with giant computer monitors, a team of seven Zap Energy engineers prepares to generate super-heated plasmas in a demo fusion device.
Like mission control at a space launch, a Zap operator checks with her teammates one-by-one to ensure the systems are ready. One engineer oversees capacitors stored in black shipping containers that charge from the grid and send a massive energy surge to the fusion reactor. Another verifies the silvery liquid metal cooling the reactor core is circulating properly. Each component must play its role on cue.
“Starting sequence, and three, two, one.”
“Charging.”
Outside the room one hears the successful “thump” of the plasma firing, followed by a flash of purple-hued light.
Zap is one of four fusion companies in the Pacific Northwest and roughly 45 worldwide that are doggedly working to do what’s never been done before: replicate on Earth the reactions that power the sun and the stars in pursuit of clean, nearly limitless electricity.
Earlier this year, the Everett, Wash.-based company hit a new milestone, creating more than 1,000 consecutive plasmas — the state of matter required for fusion — over three hours. Since then, Zap’s Century system has delivered more than 10,000 plasma-forming shots under different configurations. On a recent tour of the site, the engineers demonstrated the technology for a small group of journalists.
Each run provides a little more data, nudging the science another step closer to the goal of capturing “a star in a jar,” as Ben Levitt, Zap’s vice president of R&D, describes the reaction.
‘A tame lightning bolt’
The concept of fusion energy is simple enough. Reactors generate ion-containing plasmas that are sufficiently hot, dense, and long lasting to create conditions in which atoms that don’t want to combine are forced together and release energy. But physicists have spent decades trying to create fusion and produce more power than is required to run the devices, and no one knows if or when it will be accomplished.
Despite the long-standing uncertainty, skyrocketing demand for electricity to fuel data centers and AI operations has stoked interest in the clean power source and billions of dollars have flowed into the sector in recent years.
Fusion has attracted investments from deep pockets in the tech sector, including Chris Sacca’s Lowercarbon Capital and Bill Gates’ Breakthrough Energy Ventures investing in Zap, while OpenAI CEO Sam Altman is backing Helion Energy, also located in Everett.
Physicists are chasing fusion with different kinds of reactors using high-powered magnets and lasers to create and hold plasmas. Zap’s solution is to run a super high current through the plasma in its reactor, which produces a magnetic field that compresses the matter.
“This is literally a tame lightning bolt,” said Matthew Thompson, Zap’s vice president of systems engineering.
In truth, it’s even more intense — the current inside the plasma is cranked up 20 times higher than a bolt of lightning.
The Zap team touts its approach as more affordable and scalable than other strategies given that it has a much smaller footprint than most. Its reactor chamber is about the size of a hot water heater and doesn’t require ultra-complex magnets and lasers.
“It’s this so-called self-organized plasma structure which confines itself with its own magnetic field,” Levitt said. “And it makes its own magnetic field just when it needs it, for those few microseconds, and then it goes away.”
Parallel tracks for progress
Zap launched in 2017 with research out of the University of Washington. It has raised $330 million from investors and $13 million in U.S. Department of Energy grants. Zap’s headcount totals 150.
Its R&D approach is two-pronged, with part of the team working on improving the plasma and fusion reactors while the other works on integrating the other components needed to produce the power for the reactors and capture the energy generated by fusion to put it on the grid.
“These are parallel efforts,” said Levitt. “It’s getting the plant technology ready while we perfect the plasma performance.”
The company commissioned Century in 2024 and has already increased 20-fold the average power it can deliver to the system for plasma creation, reaching 39 kilowatts. The next target is 100 kilowatts, then 1 megawatt, with a commercial-scale system requiring 10 megawatts.
Speed is another challenge. Century initially fired one plasma shot every 10 seconds, and has since doubled that rate. But the progress needs to be exponential, not linear.
“Ultimately, for a power plant, you’re going to need to do that [shot] 10 times a second, so much faster,” Levitt said. “But we’re not worrying about that right now.”