別講：makes（說 building）· 一百倍、付不起 · 國防 · 路線圖過去式 · joint development · the world's first · in TSMC · 開場即興與 bear with me。黃字 ⟨ ⟩ 是超時可跳過的句子，①②③ 是砍序
P1The photon engine for civilization.
40 秒 ｜ 上台累計 0:40
▷ 封面，看鏡頭；問句停住，換頁再答
Good afternoon, everyone. I'm Jim, founder and CEO of Holonomy Systems.
Photons are particles of light.
We are building a way to make high-power photons on a wafer, the way semiconductor chips are made.
To ignite nuclear fusion, the clean, limitless energy source for civilization.
Why photons, and why now?
P2The photon shift is already happening.
59 秒 ｜ 上台累計 1:39
▷ 四張卡由左到右；功率軸 milliwatts → megawatts
Because a paradigm shift has already started, climbing a billion-fold, from milliwatts to megawatts.
In AI data, photons replace copper because they carry more data with less heat.
In space, photons replace radio because a narrow beam carries far more data.
In energy, photons replace the wire because they carry power through air and vacuum.
In fusion, photons replace the sun's gravity ⟨because they press fuel harder than the center of the sun⟩③.
P3Why Now
54 秒 ｜ 上台累計 2:33
▷ 左：增益圖；右欄三行（once a day · ten times a second · tens of gigawatts）
The National Ignition Facility, in the US, is the world's largest laser, built to ignite fusion.
In 2022, they succeeded: more energy came out than the photons put in.
But it fires about once a day.
A power plant must fire ten times a second, and the leading designs use chips that turn electricity into photons.
One plant needs tens of gigawatts of peak power from these chips.
P4The Problem
55 秒 ｜ 上台累計 3:28
▷ 六道工序沿圖上 1 到 6 走；最後一句指四張文獻卡
For one plant, that means three to six years of the world's output of these high-power chips, at 2016 capacity.
And today they cost thirty to a hundred and thirty times what plant studies assume.
Why? Because each chip is cleaved, soldered, sorted, lensed, stacked and tested, one at a time.
Last year, the US lab that ignited fusion and Germany's leading laser-chip institute published the same diagnosis.
P5The Solution
69 秒 ｜ 上台累計 4:37
▷ 左欄 CCD → CMOS 小圖；右欄 PCSEL 元件圖
So we change how the chip is made. This has happened before.
CMOS camera sensors started behind CCDs.
⟨But they were made on standard wafers, so they reached volume first, iterated faster, and overtook CCDs.⟩②
Our chip is a photonic-crystal surface-emitting laser, a PCSEL.
A pattern of tiny holes makes the whole chip surface emit photons as one beam, all made and tested on the wafer.
Like CMOS, it starts behind today's best, and it scales the same way.
Our first devices are targeted for 2027.
P6Market and Customers
55 秒 ｜ 上台累計 5:32
▷ 左：買家曲線由左到右；右欄四步；市場數字不念
Our first customers are laser makers, who need these chips long before the first fusion plant is built.
From there, we climb in power density and efficiency, through the markets you just saw, to the fusion plant.
The plan with each customer is four steps: an assessment, a paid study, samples, then volume.
So revenue can start early, with the study, and grow with every chip they buy.
P7Product and Business Model
49 秒 ｜ 上台累計 6:21
▷ 左欄 Product、右欄 Business model，照三列由上往下
What do we sell? First, custom design: a paid study scoped to the customer's specification.
A fixed fee, paid against milestones, with acceptance criteria agreed up front.
Second, devices: engineering samples quoted per project, then chips and modules in volume, priced per watt.
Every chip ships with its model.
⟨Later, a process kit with a commercial foundry, licensed to other designers.⟩①
P8Team and Traction
91 秒 ｜ 上台累計 7:52
▷ 左欄三人；右欄四條，UA 那條說 here
We have the right people for this.
I am a serial founder trained in nuclear engineering.
Eric Chen was a process integration engineer at TSMC.
We design with the NYCU lab that built one of the first electrically driven blue surface-emitting lasers.
A Taiwan epitaxy foundry has taken on our first process run.
A European laser research center has offered to test our first chips.
And here at the University of Arizona, a research group has been under NDA with us since June, working on modeling and the test plan for our first devices.
Three laser-fusion developers are in talks with us, and we plan to share a system-level assessment with one of them.
P9Where UACI moves our roadmap.
84 秒 ｜ 上台累計 9:16
▷ 路線五站由左到右，全用未來式；三張卡 First、Second、Third；收尾看鏡頭
This year, we start our first process run at the foundry.
In 2027, we will tape out, measure our first chips, and sign our first paid design study.
In 2028, we will ship evaluation chips to R&D centers.
We'd love UACI to help on three fronts.
First, funding support: introductions to U.S. investors, and help navigating DOE programs like ARPA-E and INFUSE.
Second, the University of Arizona: research groups and labs to test our first chips.
Third, a home in Tucson: a private office from November, and mentorship on selling to U.S. laser makers.
Join us in building the photon engine for civilization. Thank you.