Best of LinkedIn: Sustainability & Green ICT CW 46/ 47

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-Forty-Six and Forty-Seven.

00:00:10: Frennis supports ICT enterprises with market and competitive intelligence, decoding green software developments, benchmarking emerging standards, tracking regulatory shifts, and analyzing competitor strategies.

00:00:23: So welcome to the deep dive this week.

00:00:25: We're really focusing on the most Immediate and impactful conversations happening right now around sustainability and green ICT and we're drawing everything from some of the sharpest insights shared on LinkedIn over the past couple weeks.

00:00:38: Yeah, and I think what's so critical about this batch of sources is that the whole industry conversation has matured.

00:00:44: It feels like we're finally moving past the abstract.

00:00:46: We need to be sustainable talk.

00:00:48: people are now diving into the specific technical levers, you know, the real world strains and the heart metrics.

00:00:53: you need to actually implement this stuff.

00:00:55: Exactly.

00:00:56: Our mission today is really to give you a shortcut to understanding the how of green ICT as it stands right now.

00:01:01: Absolutely.

00:01:02: The vibe is less about ambition and more about verifiable efficiency.

00:01:08: And nowhere is that tension clearer than in our first big theme.

00:01:13: The dual challenge of AI trying to square its staggering environmental footprint with its promised climate handprint.

00:01:20: The numbers are, well, they're hard to ignore.

00:01:22: We've seen just incredible growth from the companies driving this AI revolution, but it all comes at a physical cost.

00:01:29: A post from James Martin illustrated this perfectly.

00:01:32: He noted NVIDIA's emissions shot up by eighty seven percent in twenty twenty four eighty

00:01:36: seven percent.

00:01:37: That's a huge spike.

00:01:38: It is.

00:01:39: And it sort of confirms that the current trajectory for large scale AI is.

00:01:42: Frankly, incompatible with hitting any serious planetary goals in the short term.

00:01:46: That's not just a trend, that's a systemic risk if that energy curve keeps going up.

00:01:50: Right,

00:01:50: and the projections are even more alarming.

00:01:52: There's new research, which, Fankichi, you referenced, that estimates by twenty-thirty AI servers could add the co-equivalent of, get this, between five and ten million cars.

00:02:01: Wow.

00:02:02: And that's just the sea.

00:02:03: That's

00:02:03: not even counting the water.

00:02:05: The water usage could rival the household needs of up to ten million Americans.

00:02:09: So when we talk about planetary boundaries, these are the resource strains that need, you know, an immediate architectural intervention, not just sourcing more renewables.

00:02:19: So if bigger and hotter isn't the answer, then we have to be talking about smarter design.

00:02:23: Exactly.

00:02:24: And Frank Huber highlighted a pretty radical alternative.

00:02:28: Dresden's SpinCloud.

00:02:30: It's a neuromorphic supercomputer.

00:02:32: modeled after the human brain, which is the fascinating part.

00:02:35: It

00:02:35: claims to be ten times more efficient per watt compared to conventional AI systems.

00:02:40: So what makes it so efficient?

00:02:42: Why is that approach better?

00:02:43: Well,

00:02:43: it's why that's so interesting.

00:02:45: It's not just optimizing what we already have.

00:02:47: It's a fundamental change in approach.

00:02:49: By simulating neural networks directly, it drastically cuts down on the need to move huge amounts of data back and forth between memory and the processors.

00:02:57: Ah, so less data movement equals less energy.

00:03:00: Precisely.

00:03:01: It makes green AI a real technological upgrade, not just some, you know, green add-on.

00:03:06: That completely reframes the conversation.

00:03:09: But that efficiency, that only addresses the footprint, right?

00:03:12: The power it consumes.

00:03:13: Anna Lerner-Nesbitt brought up a really critical distinction.

00:03:17: We need to focus more on the handprint, which is how AI is actually used out in the world.

00:03:23: And that's the crux of the moral hazard here.

00:03:25: The handprint could be the positive stuff, you know, accelerating sustainability with AI.

00:03:30: but it also covers the negative acceleration.

00:03:33: Dr.

00:03:33: Sasha Lucioni noted that at COP-Thirty, a lot of the talk was dominated by these narratives of AI as a climate savior, basically climate washing.

00:03:42: But that overlooks its actual deployment, which in some cases includes accelerating oil and gas exploration, and that obviously just undercuts any efficiency gains you might make.

00:03:51: It's the classic paradox of any powerful new tech.

00:03:54: You can't just talk about the good it could do.

00:03:56: You have to look at what it's currently doing.

00:03:58: Exactly right.

00:03:59: And for practitioners, this brings the focus right back to design.

00:04:02: principles.

00:04:03: The Will Nordberg made a great point that responsible scaling really hinges on efficient architecture.

00:04:08: So smaller, more specific models.

00:04:10: Yes,

00:04:11: smaller purpose built models, optimizing data routing, smart caching.

00:04:15: It means designing for cost efficiency is now basically the same thing as designing for sustainability,

00:04:20: which is a perfect lead into our next theme, where cost and carbon savings really start to converge.

00:04:25: Yeah, green software, cloud optimization, and the rise of phenops principles.

00:04:32: A lot of organizations still see green IT as just a hardware problem, you know, e-waste, procurement standards, that sort of thing.

00:04:38: But as Anita Schittler pointed out, the software layer is this hugely undervalued strategic lever.

00:04:44: The old idea that sustainability is expensive just doesn't hold up in the digital

00:04:49: world.

00:04:49: Okay, but why is software different in that respect?

00:04:52: Because software optimization gives you instant financial returns.

00:04:55: When you optimize code or you streamline your cloud architecture, you can deprovision resources you're not using, like right away.

00:05:01: You save energy and you cut your cloud bill.

00:05:03: Right.

00:05:03: Christine Shields cited this fantastic example from Christopher Salerno, a project achieved a fifty-one percent carbon reduction alongside a forty-one percent cost reduction.

00:05:13: production.

00:05:14: That's proof.

00:05:15: The business case is ironclad.

00:05:17: It sounds less like a sacrifice and more like just smart management.

00:05:20: It

00:05:21: is.

00:05:21: And this is exactly why FanOps Cloud Financial Operations is merging with GreenOps.

00:05:26: Shruti Ranjan Ranjan highlighted a painful truth, which is that roughly thirty percent of all cloud spending is just wasted due to mismanagement.

00:05:34: And

00:05:34: that thirty percent waste is also wasted carbon emissions.

00:05:37: There are two sides of the same coin.

00:05:39: If we look at this through the lens of how organizations are changing, the convergence is inevitable.

00:05:45: Fine-offs and green-offs, as both Smriti Ranjan and Kevin Leslie argued, they share the same fundamental goal, eliminate waste.

00:05:53: Whether you're measuring dollars or kilograms of fuel.

00:05:55: Exactly.

00:05:55: The action is the same, optimization.

00:05:58: And the hyperscalers are recognizing this too.

00:06:00: We're seeing tools like Microsoft Emissions Impact Dashboard and the AWS Customer Carbon Footprint Tool becoming standard.

00:06:07: But the waste often starts even before you deploy anything.

00:06:10: It starts with a sheer amount of data we generate and just hoard.

00:06:15: You want to peer hone and made such an essential point.

00:06:18: Digital data is not immaterial.

00:06:20: Every single duplicated file, every log you keep, it all requires server space, infrastructure, cooling, electricity.

00:06:28: This is a critical point for any business managing large digital archives.

00:06:34: You want to up your own and challenge people to ask themselves a few key questions.

00:06:38: Are we just collecting data just in case with no clear use?

00:06:41: Is the exact same information duplicated across three or four different systems?

00:06:45: Right.

00:06:46: Getting rid of that kind of information blow is probably the lowest hanging fruit for cutting both costs and emissions.

00:06:52: Okay,

00:06:52: so let's move from the software layer to the concrete physical infrastructure that holds all that data.

00:06:58: Data centers.

00:06:59: This is an area where innovation is happening incredibly fast, but it's also where integrity is being challenged.

00:07:05: Yeah, the physical infrastructure needed for AI and all this data is really pushing the limits of engineering.

00:07:10: And we saw some really inspiring examples globally.

00:07:13: Tarik New described how the UAE is pioneering sustainable data centers, like in the middle of the desert.

00:07:18: In the desert.

00:07:20: They're using massive solar farms and some really advanced cooling, specifically AI optimized closed loose liquid immersion cooling.

00:07:28: It can cut energy use by up to fifty percent.

00:07:30: And it's not just there.

00:07:32: Alberto Zancanella shared news from China about some radical cross-functional innovation.

00:07:38: What were they doing?

00:07:39: Building a wind power data center.

00:07:41: Under the sea.

00:07:42: Under the sea, okay.

00:07:43: So the search for efficiency and proximity to renewables is driving some truly global architectural change.

00:07:50: And

00:07:50: we can't talk about data centers without talking about water.

00:07:53: It's an increasingly precious resource.

00:07:55: Absolutely.

00:07:56: Michael Renneker detailed Microsoft's work on their zero water data center design.

00:08:01: They're using tech like zero water evaporation cooling and closed loop liquid cooling.

00:08:06: And

00:08:06: the savings.

00:08:06: It

00:08:06: can save over a hundred and twenty five million liters of water annually.

00:08:10: per data center.

00:08:11: That

00:08:11: scale of water saving is hugely significant, especially with the AI boom putting so much pressure on local water supplies.

00:08:19: Yet despite all this incredible innovation, the integrity of the claims being made is under some pretty intense scrutiny.

00:08:24: It is.

00:08:25: Mark Butcher reacted very strongly to a public statement from a major tech company VP who claimed they had one hundred percent net zero data centers.

00:08:35: He called the claim scientifically impossible and damaging to trust.

00:08:40: Wait, why impossible?

00:08:41: I thought net zero just meant you, you know, balance what you use.

00:08:44: It's the scope.

00:08:45: three issue, mostly.

00:08:47: Today, no hyperscaler is truly net zero if you account for all their environmental impacts.

00:08:52: That means the embodied carbon, the emissions from the concrete and steel used to build the facility.

00:08:58: Okay, so everything that goes into it before you even switch it on.

00:09:01: Exactly.

00:09:01: Plus the supply chain emissions from manufacturing all the hardware and, critically, the emissions from running diesel generators for backup power.

00:09:09: So misleading claims just confuse customers and can give regulators a false sense of compliance.

00:09:14: So the industry really needs to focus on verifiable metrics and continuous improvement.

00:09:19: Right.

00:09:20: And we are seeing companies respond.

00:09:22: Steve Obstein mentioned the work Trane is doing with resilience and energy storage solutions.

00:09:26: And Karen Vander Zanden highlighted a research program called MISD.

00:09:30: What's their focus?

00:09:31: They're focused on creating modular, sustainable, secure data centers specifically for these new cloud-edge environments.

00:09:39: It's a real shift toward dynamic, energy-aware, and decentralized infrastructure.

00:09:45: Which means the future infrastructure will be much more dynamic than the big monolithic facilities we have now.

00:09:51: That's

00:09:51: the idea.

00:09:52: More responsive to grid conditions.

00:09:54: Okay, our final theme takes us up to the highest level.

00:09:58: The world of policy, reporting rigor, and Scope III Accountability.

00:10:04: The regulatory environment is opposed to forced transparency, but we're seeing some significant political pushback.

00:10:09: That pushback is what makes this whole topic so volatile.

00:10:12: I mean, on one hand, the direction of travel is clear.

00:10:15: Robert Metzky stressed that companies have to set science-aligned targets, and they absolutely must move forward on scope three accountability collaboratively.

00:10:22: Which, for the ICT sector, means going deep into the supply chain.

00:10:25: Exactly.

00:10:26: And Anita Falls noted that thanks to new methods like TM-XXV, it's a way to estimate embodied carbon.

00:10:32: that invisible scope-three data is actually becoming measurable and actionable for data centers.

00:10:37: Scope three reporting is notoriously difficult, though, because it relies on everyone in the supply chain collaborating.

00:10:43: That's the challenge.

00:10:45: It requires transparency from everyone.

00:10:47: But this is where the policy environment gets messy.

00:10:50: Igor Trishov raised a pretty serious alarm about Europe's omnibus legislation.

00:10:55: He explained that this bill has significantly reduced the number of companies required to report under the Corporate Sustainability Reporting Directive, the CSRD.

00:11:04: How significant a reduction are we talking about?

00:11:06: It

00:11:07: could potentially exempt up to ninety percent of firms that were previously scoped in in some places.

00:11:13: It's a massive reversal.

00:11:14: It effectively lets a huge chunk of the business community step back from mandatory reporting.

00:11:18: And that risks undermining the whole goal of comprehensive data collection?

00:11:22: Completely.

00:11:23: And what's really interesting is that we're seeing similar legal resistance here in the U.S Nolan, Goddard, and Gala DUEZ discussed the legal pushback in California against their landmark climate disclosure bills, SB- two fifty three and SB- two sixty one.

00:11:37: Right.

00:11:37: Those bills were designed to mandate comprehensive carbon reporting, including scope three.

00:11:43: The resistance really highlights this critical tension.

00:11:46: Big tech often prefers opacity.

00:11:49: They argue scope three reporting is too burdensome or the data is unreliable.

00:11:53: These legal challenges, they expose flaws in our current systems, but they also show this continuous effort to avoid radical transparency.

00:12:01: So on one hand, you have technical experts who are achieving forty, fifty percent cost and carbon reductions in their own operations.

00:12:09: And on the other hand, you have these global political forces trying to ease the regulatory pressure that pushes companies toward that kind of external accountability.

00:12:17: And if you connect that dichotomy all the way back to the beginning of our deep dive, the lesson is clear.

00:12:22: If the companies can't accurately report, especially on scope three, it just completely undercuts their ability to set credible science-based targets.

00:12:30: And it devalues all that incredible work being done in green software and data center efficiency.

00:12:35: Exactly.

00:12:35: Transparency is the non-negotiable bedrock of credibility here.

00:12:39: So what does this all mean for the ICT professional today?

00:12:42: I think our deep dive really showed that while the scale of the challenge is immense from, you know, the architecture shifts for AI to water demands, the solutions are tangible, technical, and financially

00:12:53: rewarding.

00:12:54: Yes, that's the key.

00:12:55: We're seeing the strategic convergence of cost control and carbon control everywhere.

00:13:00: The conversation feels mature now.

00:13:01: It's focused on implementation.

00:13:04: We know the challenge is massive, but The tools, things like neuromorphic computing, fine ops and green ops, zero water cooling, they're moving from the lab to operational reality.

00:13:14: So maybe here's a final provocative thought for you, the listener, to consider.

00:13:18: Given this

00:13:19: increasing political pushback against rigorous external sustainability reporting like the omnibus in Europe and the legal challenges in the US, how can the green ICT professional ensure that internal efficiency gains remain a core prioritized business mandate, even if that regulatory pressure temporarily goes

00:13:56: away.