It’s been a bumpy year for tech, with mass layoffs and a struggling economy indicating a shift in the once thriving industry. While we saw some companies flourish during the pandemic, thanks in large part to their ability to operate remotely while also providing value to consumers’ isolated lives, the post-pandemic return to normal has forced companies to reevaluate their growth and spending.
I spoke with Raul Martynek, CEO at DataBank, to explore the current state of the industry and investigate why this is happening, and what he thinks the future holds for the tech space. Raul has a unique perspective from behind the scenes of the industry, as DataBank operates the data centers that power the technological innovations, we’ve come to take for granted.
Gary Drenik: How has the tech industry changed in the past two or three years?
Raul Martynek: The most obvious point is AI – in just the first half of this year, the AI “gold rush” exploded. According to a recent Prosper Insights & Analytics survey, 59.6% of US adults have heard of Chat-GPT to some degree, with 11.8% already using it for various tasks, including research, writing assistance, and customer support.
What’s less obvious is the infrastructure that supports all of this in the background. The past few years have seen the development of edge and the decentralization of the Internet. Slowly, data applications are moving closer to the end user to decrease latency and increase scalability, and we’re seeing a fundamental rearchitecting of the Internet at every level to adapt to new technologies like Chat-GPT. Everything – fiber, networks, data centers – is becoming more distributed and increasingly replicated.
Chat-GPT, and its ripple effects in the generative AI world, has led to intensified demand for compute power, on top of normal growth from enterprises and hyperscalers, that affects everything from the semiconductor chips that power supercomputers to the designs for new data center builds.
Drenik: What was the catalyst behind this “Great Reset,” as you call it?
Martynek: Silicon Valley layoffs have created a huge stir this year, adding to the economic worry we’re experiencing on a societal level. But unlike in the late 1990s, we’re not experiencing another dot-com bubble; it’s a rather healthy pruning in tech. The layoffs over the past few months have brought to light that tech was growing too quickly from our consumption habits during the pandemic. As a result, companies bagged high valuations and impressive growth rates but ultimately saw no profit because this simply wasn’t sustainable post-pandemic.
Now, the industry’s going through a reset, and this is what I view as the “Great Reset” – a recalibration of growth versus spending as investors and consumers grow more cautious while the economy recovers.
The “grow at all costs” mindset we’ve seen in the past is no more – leaders will need to be more disciplined, ensuring that ROI is understood and knowing that capital won’t be as readily available. Despite recent investments in AI, global venture funding is still down a significant amount compared to last year, and the same Prosper Insights & Analytics survey shows that half of US adults are still spending less overall in the last month due to the state of the economy.
Drenik: What’s an example of the “grow at all costs” mindset failing and how do we overcome it?
Martynek: The “grow at all costs” mindset, where founders and investors pour money into new tech hoping to drive high growth rates and expedite innovation, fails to deliver because the nature of technology means it takes time to develop. New tech always goes through a crawl, walk, then run phase.
For example, everyone is disappointed in 5G right now, because it hasn’t delivered on the high speeds and faster responsiveness that were promised by the newest generation of networking. While it’s true that we haven’t realized the dreams of “true 5G,” it’s also still too early to see real returns on the investments we’ve made into bringing 5G to life.
Getting to a “standalone” 5G network will require billions of dollars to build the network architecture needed to support it. Even if carriers can provide 5G, the ecosystem of applications and devices that can effectively use it still needs to be built in addition to the IT infrastructure at the foundation of the entire network.
This is just an example of how patience and continuous investments are needed before we can expect to see real returns. We have to be realistic about technology’s evolution rather than expecting growth to happen instantaneously.
Drenik: Can you share more on the current IT infrastructure behind technological innovations like generative AI?
Martynek: There’s a lot that goes into creating artificial intelligence, which has become slightly more well-known thanks to the scrutiny of Chat-GPT. You need more than just data sets to train and run AI, you also need the computing power to process the massive amounts of data, the space to store it, and the power to, well, power it.
Training large AI models is incredibly data-intensive – training OpenAI’s GPT-3 required 936 MWs of data center power, equal to 3 hours in Dallas (one of the largest tier-one data center markets) if the entire market’s power resources were dedicated solely to training GPT-3. This presents a problem for the industry. Considering data centers can take anywhere from 24 to 36 months to complete a new build, that means the new space coming online now was scoped based on insights from two or three years ago, before the current AI rush.
A data center bottleneck would be a real hindrance to the development of AI. The good news, though, is that infrastructure investors, who are drawn to the stable, long-term returns from bridges, ports, airports, energy, and the like, saw an opportunity in digital infrastructure such as data centers, fiber, and wireless towers when commercial real estate market faltered during the pandemic shutdown. It’s opened the door to capital for data centers that once were reliant on public REITs to raise more funding to meet new infrastructure demands.
Drenik: Where do you see the tech industry going from here? What is on the horizon for technology adoption?
Martynek: For something like 5G, even though it seems like a disappointment right now it’s actually following the pattern of previous technological adoptions. We’re even seeing it a little with Chat-GPT, too, as people come to understand what “generative AI” really means and the limits to its generative capabilities, such as its tendency towards hallucination. At the same time, we’re fortunate enough to witness the crawl-walk-run progression of generative AI, as GPT-4 is said to be leagues ahead of its predecessor.
Despite the current recalibration of the tech industry, I anticipate that everyone will have forgotten about this Great Reset in a few years, and we’ll be pouring right back into tech like it’s 2019 again, only with slightly better technology. I suspect we’ll even see further development on the 5G front as the Internet moves more toward edge computing. The tech world is always innovating, and we will continue to overcome each new obstacle, be it data center bottlenecks or 5G letdowns, so I’m always excited to see what we’ll think of next.
Drenik: Thanks for sharing your insights on the state of the tech industry and technological innovations, Raul.
Read the full article here
