Why is Google willing to sign a 20‑year‑long agreement? The reason is straightforward. Its data centers demand power — long‑term, stable, and large‑scale electricity supply.

Over the past two years, whenever discussions revolved around the race for AI talent, conversations inevitably centered on large language models, GPUs and algorithm engineers.

Yet a recent agreement signed by Google has brought another critical question to the forefront:

With chips readily available, where will the power come from?

On October 6, Google and U.S. energy firm Constellation announced a long‑term partnership. The two parties entered a 20‑year power‑purchase agreement to upgrade six operational nuclear power plants across Illinois, Pennsylvania and New Jersey, adding 890 megawatts of generating capacity to the PJM power grid.

890 megawatts is roughly equivalent to the output of one large‑scale nuclear reactor.

This partnership targets 11 reactors within the six existing nuclear plants. Upgrades include turbines, steam generators and digital control systems. Simply put, existing facilities will be retrofitted to boost their power output.

Why is Google willing to sign a 20‑year‑long agreement?

The reason is straightforward. Its data centers demand power — long‑term, stable, and large‑scale electricity supply.

01 AI Firms Have Secured Chips, Now They Worry About Power

Training and running AI models consumes massive computing resources housed inside data centers. Whether a data center can go live depends not only on sufficient chips, but also on adequate local power availability, grid access and capable cooling systems.

GPU shortages dominated headlines in the past. Today, power constraints are becoming increasingly unavoidable.

Even if a tech company secures chips and land, projects can still get stalled if the local power grid cannot deliver enough electricity in time.

Constructing new power plants and transmission infrastructure moves at a glacial pace. Permitting, construction and grid interconnection often take multiple years.

By contrast, retrofitting already‑operational nuclear power plants is faster. According to Google, these projects will be completed in phases by the end of 2032, delivering 890 megawatts of round‑the‑clock power to the PJM grid.

Google has previously taken part in other U.S. nuclear‑energy initiatives. Combined with this new partnership, the total new capacity from its supported nuclear‑expansion and restart projects exceeds 1.5 gigawatts.

From a job‑hunting perspective, the noteworthy takeaway is not that Google is suddenly “dabbling in nuclear power”. Rather, AI‑driven corporate demand is moving upstream along industrial supply chains.

Previously, everyone’s focus was on models and chips. Now power supply, power grids, cooling and data‑center construction have become mandatory challenges to solve.

Recruiting demand will follow suit.

02 These Opportunities Will Not Be Concentrated Exclusively at Big Tech Companies

Figures released by Google and Constellation show this partnership represents over $4.3 billion in new investment. It is expected to sustain approximately 4,400 existing jobs and create around 7,200 new positions during the construction phase.

These roles are largely tied to nuclear‑plant upgrades and associated engineering construction. Actual hiring organizations may include Constellation, engineering contractors, equipment vendors and technical‑service providers.

Some positions exist only for the construction period and are not permanent long‑term roles.

Even so, it is clear that job openings will spread across the AI‑infrastructure industrial chain.

Electrical engineering students may focus on power systems, control systems, power‑distribution design, grid planning and load forecasting.

Mechanical engineering students can explore thermal management, cooling systems, facility maintenance and equipment reliability.

Civil, architecture and project‑management majors may participate in data‑center construction, construction management and infrastructure planning.

CS and data‑science students do not need to limit themselves purely to internet‑product roles.

Google and Constellation plan to leverage cloud computing and Gemini for site‑selection workflows, power‑flow modeling, equipment‑health monitoring, outage management and critical‑infrastructure protection.

Translated to the job market, new demand may emerge in industrial software, energy‑data analytics, automation and cybersecurity.

Business‑major students also have more viable paths than many realize: energy procurement, project financing, supply‑chain operations and infrastructure investment, for instance.

If you unpack this value chain, AI companies sit merely at the origin of demand. Power‑generation firms supply energy; power grids deliver it; engineering firms build and retrofit facilities; equipment suppliers deliver cooling, electrical and control hardware; only then do we reach data‑center operations.

Not every job created by AI will carry the title of an “AI role”.

03 Many Opportunities Lie Within the News‑Maker’s Supply Chain

When reading industry news, students tend to fixate on the headline‑making company.

Upon seeing Google invest in nuclear power, they immediately open Google’s career portal to search for Software Engineer openings. When a firm announces new data‑center builds, they only browse that company’s job board.

Industrial investment does not translate to hiring in such a direct fashion.

When Google needs more power, hiring growth may first appear among power utilities, engineering‑design firms, cooling‑equipment suppliers or construction contractors.

When Google expands data‑center capacity, opportunities may sit with partner firms handling site selection, construction, equipment installation and facility operations.

These can be thought of as second‑order opportunities: openings reside not with the headline company itself, but with vendors enabling its expansion.

Adjust your search keywords accordingly. Beyond AI, Machine Learning and Software Engineer, try:

data center

critical facilities

power systems

electrical design

energy analyst

infrastructure

industrial automation

Industry demand for talent does not mean every opening is accessible to international students.

Roles in energy, nuclear power and critical infrastructure vary widely. Some may require professional licenses, U.S.‑citizenship requirements, compliance screenings or work‑authorization restrictions. Other positions in software, data analytics and general engineering carry comparatively fewer barriers.

Therefore, growing industry investment does not automatically mean you qualify to apply.

Review each job description individually. Verify work responsibilities, immigration‑related requirements, employment type and whether the employer has a documented track record of visa sponsorship.

04 Should You Switch Majors to Chase This Trend?

Generally speaking, no.

A more practical approach is assessing how your existing major can connect to this industrial chain.

CS students can add projects focused on energy data, industrial systems or cybersecurity.

Electrical‑engineering students can strengthen skills in power systems, controls and simulation.

Mechanical‑engineering students can dive deeper into cooling and thermal management.

Business‑major students can learn about energy markets and infrastructure investment.

Do not rush into new courses, certifications or second‑degree programs purely after reading one industry article.

First pull 20‑30 real‑world job descriptions. Compile recurring skill requirements, then map them against your coursework, personal projects and internships.

Identify what you are missing: is it a software tool, relevant internship experience, insufficient technical skills, or fundamental eligibility constraints tied to immigration status?

Answer these questions before deciding what upskilling to pursue.

Google’s agreement does not prove nuclear power will become a universal hot sector for international students, nor does it signal broad‑scale hiring surges across big‑tech firms.

What it illustrates is capital from AI enterprises is flowing into energy and infrastructure.

In the coming years, some AI‑enabled jobs will bear no “AI” label at all. They will appear on career pages of power utilities, data‑center operators, equipment manufacturers and engineering firms, under titles such as Power‑Systems Engineer, Facilities Engineer, Energy Analyst or Industrial Cybersecurity Engineer.

When job hunting, avoid chasing trendy job titles alone.

Follow where corporate capital flows and where project bottlenecks emerge. Identify which organizations will ramp up hiring as a result, and you will uncover far more options.