Dr John Sullivan Talent Management Thought Leadership
Most are surprised to learn how difficult it is to staff new AI data centers (AIDC). That’s important because the billions budgeted for new AIDCs will be wasted if the new centers cannot maintain continuous operations due to serious recruiting and retention issues. Staffing these AIDC jobs will be exceptionally challenging.
Primarily because the current and future supply of experienced AIDC talent is likely to be minuscule. So, everyone involved in any aspect of AI must first learn that these AIDCs are likely the weak link in your overall AI effort.
To prevent that important link from breaking, those involved in AI will need to take partial ownership of the recruiting problem to the point that they immediately begin collaborating with and sharing ideas with the recruiting function.
An AI data center – is a specialized, high-performance facility that’s designed to train, deploy, and manage large-scale AI models. These data centers are currently undergoing a construction boom. So, when the boom is over, the number of operational data centers will have increased by 57 %. Currently, most AI data centers are built in rural locations because these are the only remaining geographic areas that offer sufficient affordable land and electric power.
A List Of The Talent Problems That AI Data Centers Must Overcome
The remainder of this article highlights the multitude of recruiting, staffing, and retention problems that must be solved to prevent AIDC talent issues from limiting the overall success of your AI effort. The most damaging problems occur early in the list.
- You will face a dramatic shortage of experienced workers – because the flawless operation of an AIDC doesn’t allow for much “learning on the job”. Your recruiters will be forced to target only candidates with recent experience in AIDC roles. And because almost everyone who meets that standard is currently employed, your recruiters will be charged with the difficult task of convincing people who already have a perfectly good job to leave it for a new one that will likely require an interstate relocation.
- A poaching recruiting strategy will be required – because most in the pool of qualified AIDC candidates will already have a data center job. Your recruiting effort will have to adopt “a poaching strategy” that will focus on hiring the employed workers at your competitor’s data center. Unfortunately, this type of poaching can be problematic because it requires your corporate recruiters to have the selling skills necessary to convince your competitor’s AIDC employees, who probably have not been actively looking, to change their minds and proactively apply for your job. Adopting this poaching strategy will also be problematic at many corporations because many talent leaders are reluctant to directly poach. As a result of their unreasonable fear of retaliation (even though your competitors will poach you anyway, regardless of your volume of poaching). Outside of direct poaching, you are most likely to find the most qualified AIDC candidates through employee referrals, boomerang rehires, and professional conferences.
- Your rural location will increase relocation issues – because these new data centers are now located in rural areas far away from population centers. Most new hires will require an expensive relocation package. And because your new hires will be new to this rural area, many candidates will seriously consider a job only after learning about housing availability, school quality, amenities, and employment opportunities for their spouses.
- Realize that training new-hires won’t be a viable option – you may be wondering why you can’t just hire recent college graduates or other types of engineers and train them, instead of recruiting those with AIDC experience. But unfortunately, few four-year colleges even offer an AIDC degree option. And there are several reasons why you can’t just train newly hired engineers. The first reason is that even if your L&D department has the capacity to train for AIDC jobs, it will likely be busy updating the training of your experienced new hires. It’s also important to realize that relying on training, in lieu of proven experience, is risky. Because you can’t know in advance what portion of your newly trained employees will turn out to be as successful as your experienced hires. And finally, your data center training will likely take months to complete. This will be unacceptable to most managers, who will need their complete AIDC capability almost immediately.
- Recruiting AIDC talent will be more costly – recruiting data center talent will be more expensive for several reasons. First, because those with AIDC experience are hard to find, you will have to invest significantly more money and resources in your sourcing effort. Next, to increase their willingness to relocate, you may have to literally fly each of your finalist candidates in (and perhaps their spouse) for their final on-site interview. And because your candidates have a current job to return to, you will need to plan on having a much lower offer acceptance rate. And finally, you must realize that most new hires will insist on full coverage of relocation costs.
- You must effectively counter local and environmental resistance – you, of course, need to be aware that in order to attract and close the best candidates. You will need to overcome several types of public resistance to the opening of your data center. Many will speak out against the opening because it will change the rural nature of the area. While others will argue that your center will be contributing to higher local electric rates. Some of your local employers will argue that your higher pay rates will put pressure on them to raise their local wages, while still others, out of fear, will resist anything that is related to AI. To minimize candidate resistance, your recruiting function must assemble and thoroughly test the effectiveness of your counterarguments to each top candidate concern. And if you want a high offer acceptance rate, you must provide each candidate with effective counterarguments they can use when confronted by family, friends, and colleagues about the ethics of their decision to pursue the job or accept your offer.
- You will likely need global recruiting – because many corporations are opening up AI data centers around the world. Your recruiting function will need the capability to recruit AIDC talent from around the world.
- Even if you lease AIDC space, you should be concerned about recruiting – if instead of building your own AIDC, you decide to lease space in a vendor’s data center. You should use a vendor’s capability to recruit and retain experienced talent as a primary deciding factor in your decision
- Be aware that employee retention will be a constant problem – because there will be an ongoing war for AIDC talent. You must assume that your own employees will be continually targeted by recruiters from your talent competitors. And while you can’t completely stop this raiding. You can, however, make the raiding less effective. By developing a data-driven retention program. That provides each of your AIDC employees with a personalized and frequently updated retention plan.
Note: I have previously written related articles covering difficult recruiting areas, including how to win the war for AI talent. And why recruiting for Quantum Computing jobs will be another recruiting battleground. Because both of these recruiting areas offer an opportunity for your recruiting function to provide the company with a competitive advantage in critical technical areas.
Final Thoughts
If you assume the big picture perspective of the recruiting function. You will quickly realize that the number of AIDC employees that you will be hiring will be a relatively small percentage of your total hiring. However, it’s equally important to note that the ability to fill this small volume of jobs with quality, experienced talent may be the weakest link in your entire AI effort. Because, unfortunately, failing to have a capable, effective, and continuously operational AI data center will likely be enough to slow or even damage your entire AI effort.
Thanks for finding the time to read and share this article.
Notes for the reader
- This is the latest article from Dr. Sullivan, who was called “the Michael Jordan of Hiring” by Fast Company.
- You can subscribe to his Aggressive Talent Management newsletter (which focuses on recruiting tools, current recruiting opportunities and recruiting trends). Either here or by following him on LinkedIn.
