Best of LinkedIn: Sustainability & Green ICT CW 08/ 09

00:00:00: This episode is provided by Thomas Allgaier and Frennis, based on the most relevant LinkedIn posts about sustainability in green ICT in CW eight-and-nine.

00:00:08: Frennes supports ICT enterprises with market and competitive intelligence decoding green software developments benchmarking emerging standards tracking regulatory shifts and analyzing competitor strategies.

00:00:20: Right and just to set the expectations right up front for you today.

00:00:24: Yeah, we're focusing on sustainability in green ICT trends seen across LinkedIn exactly

00:00:30: And welcome if you are joining us today.

00:00:32: It means you care about the actual mechanics of the tech We use every day.

00:00:36: our mission for you on this deep dive is simple?

00:00:39: We are we're cutting straight through the corporate fluff To uncover the operational realities of sustainable digital transformation.

00:00:46: yeah combining the engineering The government

00:00:48: market signals Based entirely on curated insights from real practitioners of the last two weeks.

00:00:54: So if you're a professional navigating this space, You aren't exactly the right place.

00:00:59: it really has been a highly revealing couple of weeks.

00:01:02: What stands out from the sources is a clear maturation in the industry.

00:01:07: Oh absolutely

00:01:08: like sustainability and IT Is shedding its skin as just a marketing slogan or an isolated side project, it's becoming a rigorous integrated agenda where we are finally seeing real structural trade-offs being debated openly.

00:01:24: Okay, let's unpack this.

00:01:25: The best place to start is the sheer physical weight of the cloud specifically...the massive footprint of artificial intelligence.

00:01:33: The green AI paradox?

00:01:35: Yes!

00:01:35: We are constantly calling it the Green AI Paradox.

00:01:38: It is so easy to think a generative AI prompt as weightless but the physical reality behind screen its incredibly heavy.

00:01:47: To give you a sense of the scale that weighed, Burton Kelso shared a post noting that by twenty-twenty six The data center is powering all those supposedly simple AI requests.

00:01:55: Like generating a dinner recipe?

00:01:57: Right?

00:01:57: or summarizing in meeting they are now using more electricity than the entire country of Japan.

00:02:02: think about for second an Entire industrialized G seven nations power grid completely eclipsed by server farms.

00:02:12: That's just difficult to wrap your

00:02:14: head around.

00:02:14: and it isn't Just Power To keep those supercomputers from melting down, they are drinking a staggering amount of water.

00:02:21: Yes often competing directly with local farmers and residents for shared reservoirs especially during droughts.

00:02:28: The water issue is what really caught my eye because it's so hyper-localized.

00:02:33: Chalet Ren shared some research recently regarding the US Water Infrastructure Mismatch.

00:02:38: that explains this perfectly.

00:02:39: Ren points out a single data center might request six million gallons of water per day from local authority.

00:02:46: Crazy.

00:02:47: Right, and the bottleneck isn't the national water totals.

00:02:51: there is no giant National Reservoir that data centers can just plug into.

00:02:55: Exactly.

00:02:55: you cannot easily route water across state lines The way you know.

00:03:00: wheel electricity across a power grid.

00:03:02: It's like trying to fill an olympic swimming pool using single garden hose connected To neighborhood thats already experiencing severe drought.

00:03:09: Water might exist somewhere in country but local municipal pipes And community systems

00:03:15: Burst

00:03:16: under that kind of concentrated strain.

00:03:18: The crisis is entirely about local peak demand.

00:03:21: So the immediate question that surfaces, Is how we mitigate this?

00:03:25: Abhishek Karol pushed a very practical solution in his recent post.

00:03:28: What he argues?

00:03:29: that green AI has become measurable down to the individual application level.

00:03:33: You cannot just look at the data center as a monolithic entity anymore,

00:03:37: right?

00:03:37: He needs specifics.

00:03:38: he says you need telemetry to definitively state This specific AI app.

00:03:43: use this much energy and produce this much carbon.

00:03:47: He's advocating for organizations to introduce hard-carbon budgets directly into their procurement decisions.

00:03:52: So if enterprise buyers demand sustainability metrics alongside cost in latency suppliers

00:03:57: will have no choice but to optimize for it.

00:03:59: I understand the logic there But doesn't that create a tracking nightmare?

00:04:03: And Curry offered a perspective that seems to directly challenge the granular approach.

00:04:08: She argues trying to calculate the exact energy cost of single AI query is actually the wrong question altogether.

00:04:15: What's fascinating here, she isn't saying we shouldn't track it... Right!

00:04:19: ...she says we are looking through the WRONG end.

00:04:23: She calls it the ultimate, It-Depends scenario.

00:04:26: On

00:04:26: steroids basically?

00:04:27: Exactly!

00:04:28: The carbon cost of a query depends on this specific model —the time of day... ...the ambient temperature outside the data center… …the phrasing of the prompt—It's

00:04:37: impossible to standardize

00:04:38: Yes.

00:04:39: So instead of microtracking queries she says you should be looking at the macro level – the provider's overall commitments and their grid impact.

00:04:47: Like Google versus XAI

00:04:50: Right She contrasted Google which is aggressively pushing for two hundred forty seven carbon-free energy by twenty thirty four other compute against ex AI, which currently lacks public sustainability targets.

00:05:02: So her point that enterprise leaders should use their consumer power to choose providers with systemic clean energy policies rather than stressing over the wattage of a single chat prompt.

00:05:12: in speaking of Google's systemic policies and on learner Nesbit highlighted there approach infrastructure recently Google is planning a new, eight hundred and fifty megawatt data center powered primarily by co-located solar in wind.

00:05:26: Which

00:05:26: is massive!

00:05:27: It isn't what's notable here that this avoids the typical behind the meter gas turbines?

00:05:32: Just to clarify for anyone listening who might be unfamiliar Behind the Meter usually means private localized power generator

00:05:38: Often a dirty gas or diesel turbine.

00:05:40: Right,

00:05:41: that data center builds on its own property to bypass delays and getting connected to the main public grid by integrating renewables And battery storage directly.

00:05:50: Google is projecting an eighty three percent reduction in emissions.

00:05:54: That

00:05:54: Is A massive Reduction On Paper.

00:05:56: But Infrastructure Decisions Do Not Happen In A Vacuum.

00:05:59: They Happened In A Highly Complex Political And Regulatory Context.

00:06:03: Boris Gamazachikov highlighted a very specific dynamic recently regarding the Ratepayer Protection Pledge.

00:06:09: This is voluntary commitment signed by major tech firms regarding power for data centers.

00:06:13: And just to pause here, to be absolutely clear We are simply reporting the observations from source material here without endorsing any political viewpoint.

00:06:23: Absolutely, we're completely impartial on this.

00:06:25: We just looking at market signals

00:06:27: Exactly!

00:06:28: So Boris points out that while clean energy could technically meet commitments of this pledge The specific political framing of a pledge targets quote-unquote unreliable energy subsidies.

00:06:40: In practice, he notes this particular framing might actually favor the expansion of fossil fuels.

00:06:47: Oh so

00:06:47: because it essentially provides cover to build more of those behind-the-meter gas and backup diesel generators you just mentioned rather than waiting for renewable infrastructure to scale.

00:06:56: Wow

00:06:57: that completely alters The trajectory of those emission reduction goals.

00:07:00: It does And we are seeing these tensions play out globally.

00:07:04: Dr Daniel CFN and Maria Apple Morales highlighted the core paradox playing out in Southeast Asia right now.

00:07:11: You have Singapore, Malaysia and Indonesia racing to build AI data centers to capture market share.

00:07:17: this is heavily straining local grids that are still deeply reliant on fossil fuels.

00:07:23: but The flip side the paradox This very same AI is being deployed to optimize those regions.

00:07:30: Like running smart building systems?

00:07:32: Exactly, which radically cuts urban energy use.

00:07:36: and it's driving precision agriculture that reduces water waste in farming.

00:07:40: It's a delicate balancing act And we have to be incredibly careful about assuming those optimization gains will just save us Right.

00:07:48: Dr.

00:07:48: Sascha Lucioni brought up critical warning of the Jevons Paradox.

00:07:51: If you aren't familiar with the Jeffens paradox as an economic principle where technological progress increases the efficiency of a resource, but because it becomes cheaper and more efficient.

00:08:01: The overall rate of consumption actually explodes at backfires exactly.

00:08:05: Dr Lucioni points out that the efficiency gains in AI hardware and software are currently being entirely swallowed up by the massive energy rebound effects of generative AI.

00:08:15: we make them models ten percent more efficient

00:08:17: so people use him fifty percent more often

00:08:19: which means we cannot rely solely on the hardware engineers to fix this.

00:08:23: The software developers have to step up.

00:08:26: if

00:08:26: the code running on those servers is bloated, it doesn't matter how green the power source is

00:08:31: right?

00:08:31: Which brings us to Yuna Morgenstern.

00:08:33: Yes.

00:08:34: You know Morgan Stern's thoughts on Green Software and engineering for efficiency.

00:08:39: she made a really compelling plea for minimalism in coding.

00:08:42: I loved her quote.

00:08:44: your business fits on a raspberry PI.

00:08:46: It is a great image.

00:08:47: A Raspberry Pi low-power computer, and she uses it to point out how incredibly bloated modern software has become.

00:08:56: Outside of the massive hyperscalars most companies deploy fleeps over engineered microservices just to render a few text boxes on screen.

00:09:04: Exactly!

00:09:05: She's championing green code not as some trendy ESG marketing gimmick but as core senior engineering discipline about returning real engineering optimizing compute cycles cutting computational waste.

00:09:17: The DEA supports her urgency.

00:09:19: Tim Prosser shares some sobering numbers.

00:09:21: The ICT sector currently emits two to four percent of global greenhouse gases.

00:09:25: That puts the tech industry roughly on par with the entire aviation industry.

00:09:30: Yeah, Prossor notes that fundamental architectural choices like choosing compiled languages over interpreted languages can drastically impact that energy footprint.

00:09:41: Let's break that down for a second because I think it is crucial distinction to any developer listening.

00:09:45: Why does the programming language actually dictate energy drawn from grid?

00:09:50: It comes down how computer processes instructions.

00:09:54: Think of compiled languages like translating book into another language once, printing and anyone can read whenever they want.

00:10:01: Highly efficient!

00:10:03: An interpreted language, on the other hand is like having a live translator.

00:10:06: follow you around translating your words On The Fly sentence by sentence every single time.

00:10:11: You speak

00:10:12: that sounds exhausting.

00:10:13: it Is for the machine?

00:10:14: It requires a processor to work much harder Continuously burning CPU cycles which translates directly into burning more electricity.

00:10:22: That makes perfect sense and Lenny Keeloman shared a tangible win regarding exactly this kind of optimization.

00:10:29: She highlighted a green coating pilot conducted by DSM Fermanic and Accenture.

00:10:34: Oh, right!

00:10:34: By stripping out the bloat in focusing on energy efficient software design they save the energy equivalent of driving an electric vehicle all the way from Paris to Bangkok.

00:10:44: That is incredible.

00:10:46: When you translate code efficiency into real world physical metrics like that it suddenly makes the abstract concept of Green Software something whole engineering team can rally around.

00:10:56: It anchors this concept reality And looking toward the immediate future, Anita Schuttler shared some predictions for twenty-twenty six that complicate this even further.

00:11:06: What

00:11:06: does she predict?

00:11:07: She anticipates the massive rise of agentic coding where AI agents autonomously write and speed up software development but she warns if we unleash AI to right code.

00:11:16: it requires strict resource guardrails.

00:11:18: We need systems that automatically select smallest most efficient AI model for a specific

00:11:24: task Yes rather than defaulting into the most power hungry model every time.

00:11:28: She also mentioned local AI,

00:11:29: didn't she?

00:11:30: Yes.

00:11:31: And that is a massive trend.

00:11:32: instead of sending every single prompt to a centralized... energy-hungry cloud data center.

00:11:38: Local AI means running smaller, highly specific models directly on edge devices

00:11:43: like your laptop or phone or a local server.

00:11:46: exactly.

00:11:47: it boosts data privacy reduces reliance on proprietary tech giants and utilizes whatever local decentralized green power you might have.

00:11:56: It essentially bypasses the massive datacenter transmission bottleneck entirely for everyday tasks.

00:12:01: Here's where it gets really interesting though.

00:12:03: You can have the leanest code the smartest local AI and the greenest data centers in the world.

00:12:09: But it all operates within human systems, right?

00:12:11: Technology requires rules accountability And above All people to drive The change

00:12:16: which is why the governance landscape Is rapidly shifting Right now from voluntary corporate pledges Demanded to our regulations.

00:12:22: Rebekah Lee Al posted about Singapore's proposed Digital Infrastructure

00:12:25: Act.

00:12:25: This is a major structural

00:12:27: shift.

00:12:27: It Is.

00:12:28: They are moving from encouraging best practices to enforcing legal standards for power efficiency and service resilience in data centers.

00:12:36: If you don't meet the standard, You Don't Operate

00:12:38: And that regulatory tightening is happening alongside massive shifts on the legal system itself as Morelda Colombo provided some deep insights into the evolution of net zero litigation.

00:12:49: Yeah, we are seeing courts actively shifting the burden of

00:12:52: proof.

00:12:52: It is no longer enough for a company to just publish a glossy sustainability report with a green claim.

00:12:59: Corporations are being held legally accountable for greenwashing and their voluntary commitments are being actively scrutinized under procedural law.

00:13:07: The liability's becoming very real And Nils Giesen offered a thought on this.

00:13:12: that pushes the boundary even further.

00:13:14: He argues that sustainable IT, two point oh needs robust social and governance layer

00:13:19: because it is no longer just an engineering question of how energy efficient our software is

00:13:23: exactly.

00:13:24: we have to ask The Governance Question what does This Software actually being used for?

00:13:28: That Is A crucial Point.

00:13:30: Giesen points out that in military and dual-use contexts, a hallucinating AI doesn't just result in a badly written newsletter.

00:13:37: It causes real physical harm.

00:13:41: If we are talking about sustainability in a holistic sense, responsibility and safety have to be part of the development chain.

00:13:48: It cannot just be about carbon!

00:13:50: And

00:13:50: if you connect us with that bigger picture... That sense of deep responsibility has been embedded into actual workplace culture.

00:13:57: You can't mandate transformation like this from top down and expect it work without team.

00:14:03: Drew Wilkinson outlined five key barriers to employee engagement and sustainability.

00:14:08: He lists lack of time, lack of education ,lack of place or channels to contribute .

00:14:14: Lack of aligning culture and crucially lack of relevance to their specific

00:14:18: role.

00:14:18: because if sustainability feels like extra work bolted onto a developer's day job that transformation stalls entirely.

00:14:24: you have to activate the ninety-nine percent employees who don't.

00:14:29: And speaking of the people driving this transformation, Dr.

00:14:33: Katya Shavan shared an incredible piece of historical context that we all need to hear.

00:14:38: She reminded us women have been a hidden backbone for IT from the very beginning

00:14:43: From Ada Lovelace To The Women Serving As Human Computers In World War II all

00:14:47: the way to modern CIOs and chief sustainability officers.

00:14:51: It is a lineage of innovation that often gets completely overlooked.

00:14:55: Completely,

00:14:55: Dr.

00:14:56: Chabon points out despite facing ongoing structural biases uneven access to venture funding in disproportionately small share overall climate tech investment women are currently at absolute forefront.

00:15:09: driving this climate aligned digital transformation.

00:15:12: it's powerful reminder.

00:15:13: diversity isn't just metric.

00:15:16: It directly improves innovation and performance in solving these massive sustainability challenges.

00:15:21: The human element truly is the engine, which leads us perfectly into the final area we need to cover—the physical lifecycle of our devices... ...and cold hard math running a climate tech business!

00:15:32: Let's shift to hardware life cycle.

00:15:34: I really liked Didier Appel's insight on this.

00:15:36: He suggested that we reframe the whole concept of IT asset disposition.

00:15:41: Even the phrase asset disposition sounds like you were just throwing things into a landfill.

00:15:45: Yeah,

00:15:46: he wants to reframe it into circular life cycle services.

00:15:49: It is a mindset shift prioritizing refurbishment redeployment and recycling over a linear buy use-and-replace model because

00:15:57: changing The wording changes the operational mind set

00:15:59: exactly And Brandon Smith shared a massive success story proving that this works in practice.

00:16:05: He outlined a partnership between human IT and national grid.

00:16:09: I read about that.

00:16:09: by collaborating and actually operationalizing circular IT principles, they rocketed their device reuse rates from roughly fifty percent to over ninety percent in a single year.

00:16:20: That's incredible progress!

00:16:21: It isn't just the theoretical goal that is measurable physical sustainability on this supply chain.

00:16:26: This raises an important question though for the businesses and startups actually trying to build these solutions.

00:16:33: How do they survive long enough to make an impact?

00:16:36: That is a real challenge!

00:16:38: Chris Wedding provided his stark reality check, he pointed out that climate tech doesn't die in the pitch deck it dies on the factory floor.

00:16:50: Can you clarify that for the listeners who might be more on this software side?

00:16:54: What do we mean when say it dies in the unit economics rather than market size.

00:16:59: Think of like trying to franchise a new restaurant, It doesn't matter if millions people want your burgers or total addressable markets.

00:17:05: If costs five dollars make burger and sell four dollars You lose money every single transaction.

00:17:11: That is bad.

00:17:12: unit economic

00:17:12: Exactly In climate tech are dealing with capex heavy assets.

00:17:18: That means massive upfront capital expenditures to build physical infrastructure.

00:17:23: When you have long deployment cycles, operational discipline is everything.

00:17:26: Process yields installation times

00:17:29: warranty reserves These are the metrics that determine survival.

00:17:33: as wedding puts it.

00:17:34: Operational Discipline Is The New Storytelling.

00:17:37: You don't scale a company.

00:17:38: use scale processes

00:17:39: and if the process isn't profitable at the unit level Scaling it just multiplies the financial damage.

00:17:45: Flashy market-sized projections cannot hide sloppy operational

00:17:50: math.

00:17:51: So what does this all mean for you listening?

00:17:53: Over the past two weeks, The conversation across LinkedIn has made it abundantly clear.

00:17:58: Sustainability and digital tech have grown up.

00:18:00: It really has!

00:18:01: It is a core operational engineering & governance discipline Whether its auditing local water strain of an AI data center

00:18:10: Choosing compiled language to write leaner code.

00:18:14: Navigating the new legal realities of net zero litigation or

00:18:17: proving the unit economics of a circular IT supply chain,

00:18:20: it is all one deeply integrated ecosystem.

00:18:23: if you enjoyed this episode New episodes drop every two weeks.

00:18:27: Also check out our other editions on cloud digital products and services artificial intelligence an ICT in tech insights health tech defense Tech.

00:18:35: Thank You so much for joining us on this deep dive And don't forget to subscribe So you never miss an update.

00:18:39: And before you go, I want to leave you with one final provocative thought to mull over.

00:18:44: Drawing on some brilliant insights shared by Karen van der Zanden.

00:18:47: We spend so much time in this industry optimizing for the bottom line But are we as a cultured too hyper focus on mere efficiency instead of just being passive consumers?

00:18:59: Of these digital tools how can you step up to become a co-architect at The Future Digital Market?

00:19:04: That's a great question.

00:19:05: How do we build a landscape where sustainability and digital sovereignty carry exactly as much weight in the boardroom, at speed and cost?

00:19:12: Something to think about until next time!