Aleksandar Rodic writes about how energy, AI infrastructure and data sovereignty shape business and investment decisions. Rodic's posts connect developments such as European fuel supply and data-centre expansion to their practical effects on companies, governments and industry.
They are most useful for briefings on how energy, AI infrastructure and data sovereignty affect business and investment decisions.
Best for
An infrastructure investment firm briefing on AI compute, energy security and data sovereignty.
Not for
Paid LinkedIn series for consumer retail brokerages seeking broad investor sign-ups.
Key takeaways
AI is becoming an infrastructure race, not just a model race.
Data sovereignty, not model quality, is the next enterprise AI battle.
Fuel costs are a business problem that travels to the supermarket.
Check first: Confirm a recent briefing recording, audience data, and format, availability and terms.
Main ideas
AI is becoming an infrastructure race, not just a model race · Data sovereignty, not model quality, is the next enterprise AI battle · Energy security now constrains rearmament, AI and industrial competitiveness at once
Aleksandar Rodic writes about the business consequences of energy security, AI infrastructure and data sovereignty. Their posts connect big shifts—from European gas-supply pressures to Australian data-centre expansion—to the practical questions facing companies and governments: where energy and computing capacity come from, what they cost, and who controls them.
Rodic’s strength is tracing a headline through to its operational consequences.
In a post on rising diesel costs, they follow the impact from fuel bills into transport, construction and food, then make the trade-offs tangible for a small trucking company: raise prices, accept lower margins or delay investment.
Their AI writing applies the same practical lens.
They argue that the enterprise AI contest is not just about model quality: data retention, residency and control can shape which systems organizations trust, while power, land and grid access constrain where computing can be built.
The result is a useful view of how infrastructure and policy decisions can affect business costs, investment and industrial capacity.
Aleksandar Rodic's posts treat energy, computing power and data control as one connected subject: the physical and legal foundations that decide which companies, sectors and countries can compete.
Read more
The recurring argument is that AI is becoming an infrastructure race rather than a software race, that energy supply now constrains rearmament, industry and AI expansion at the same time, and that large institutions will buy AI on data-control terms rather than model quality.
Each post follows a similar shape: a development in the news, a translation into why it matters for boards, a list of who gains and loses power, and the indicators to watch next.
The posts draw on Reuters and other outlets' reporting and add an interpretation layer rather than original reporting.
AI is becoming an infrastructure race, not just a model race
Rodic argues the AI competition is shifting to physical inputs and jurisdiction: electricity, land, grid connections, data centres, cooling, chips and regulatory approvals.
Read more
Every model depends on physical infrastructure, and where that infrastructure sits increasingly matters.
Boards should ask where data is stored, where inference happens, who controls the infrastructure, which laws apply and whether the power supply can support the load.
Countries with abundant, reliable energy and predictable regulation gain an advantage; markets with constrained grids or slow approvals lose it.
Rodic cites Anthropic's first Australian data-centre lease: a planned 2.16-gigawatt campus about 250 km from Brisbane, expected to start coming online in 2027 and used for inference rather than training.
He pairs it with Nvidia's plans with Australian partners for up to 2 GW of AI-related capacity by 2027, against Australia's roughly 1.6 GW of existing computing capacity.
Australia's energy, land, political stability and familiar legal system make it strategically attractive, he argues.
The money may sit in the layers below the model: power generation, transmission, cooling, data centres, secure cloud, digital identity and controlled gateways.
Data sovereignty, not model quality, is the next enterprise AI battle
Rodic argues that sophisticated buyers are starting to define the conditions under which AI can enter their most valuable workflows.
Read more
Commercially, this matters more than the public debate about slowing frontier AI.
Procurement is moving toward architecture-level guarantees: data residency, zero retention, model isolation, identity, access control, auditability and contractual control over intellectual property.
The winning enterprise stack may not be the smartest public model, but the system institutions can trust with their most sensitive data, favouring sovereign and private AI infrastructure, secure cloud, zero-trust, model gateways and governance providers over model vendors that need broad data access.
Rodic cites reports that Palantir, Nvidia and Booz Allen are restricting or reconsidering advanced third-party AI models.
Palantir reportedly pressed Anthropic for irrevocable zero-data-retention guarantees; Nvidia limits Anthropic to less sensitive work and relies on its own models internally; and Booz Allen has barred Anthropic's commercial model for proprietary cybersecurity work.
Energy security now constrains rearmament, AI and industrial competitiveness at once
Rodic's gas post starts with a storage shortfall and makes a broader point: Europe's plans to rearm, expand AI and data-centre infrastructure, and remain industrially competitive with China all depend on reliable, affordable power.
Read more
Higher energy costs can weaken manufacturers and influence where factories, semiconductor plants and data centres are built.
Defence production also needs dependable electricity for metals, chemicals, machinery, electronics and munitions.
He names US LNG exporters, storage operators, power producers, grid developers and efficiency providers as potential winners, and European energy-intensive industries, defence manufacturers exposed to input costs and price-sensitive Asian importers as losers.
European gas storage is only about 67% full, well below normal seasonal levels, while Gulf disruption has removed major Qatari and UAE LNG volumes.
Industry estimates put Asian spot LNG near $40/mmBtu in a cold winter.
Rodic stresses that Europe is better prepared than during the 2022 crisis and faces no immediate supply-security risk.
Exposure depends on weather, global prices and competition for cargoes, not a present shortage.
The framing echoes 2022's vulnerability argument but acknowledges that the region enters this winter in a stronger position. That shifts the recommendation from crisis response to watching how energy prices influence where strategic assets are built.
Fuel costs are a business problem that travels to the supermarket
Rodic treats a diesel price move as an operating problem rather than a commodity story.
Read more
Diesel moves trucks, construction equipment and agricultural machinery, so a higher fuel bill passes through transport contracts, food, construction and logistics and eventually reaches supermarket prices.
Their key distinction is between temporary shocks that companies can absorb and permanent costs they cannot absorb indefinitely.
Rodic also draws a line between large firms with hedging teams, purchasing power and capital access and smaller operators who have far fewer options, and ends with a direct policy point that Europe should remember smaller businesses when energy policy is discussed.
European diesel futures reached record levels and have more than doubled since the beginning of 2026, while refinery disruptions in the Middle East and Russia have cut supply and many refineries elsewhere already run close to capacity. Rodic lays out the choices facing a small transport company with twenty trucks: raise prices, accept a smaller margin, postpone investment or cut costs elsewhere.
A repeatable method: translate the news into winners, losers and a watch list
Across the posts, Rodic uses a consistent structure: a headline claim, an explanation of why it matters, board-level implications, who gains and loses power, commercial or investment relevance, and a short watch list.
Read more
The writing targets decision-makers who need the business consequences of an energy, AI or policy development, not the technical detail. The interpretation is the product: several posts draw on reporting by Reuters and other outlets, which Rodic frames as questions about where factories, data centres and capital will go.
In the LNG post, Rodic applies the same list to gas storage levels and supply disruptions; in the diesel post, the pattern leads to a policy argument about small businesses. In the AI compute post, physical dependencies—electricity, land, grid connections, cooling, chips and approvals—become board-level questions about data location, inference, ownership and applicable law.
What does AI infrastructure mean beyond chips and models—and which countries, sectors and companies stand to gain?
Read more
Rodic starts with two developments: Anthropic's first Australian data-centre lease for a planned 2.16-gigawatt campus about 250 km from Brisbane, expected to begin coming online in 2027 for inference rather than training; and Nvidia's announced plans with Australian partners for up to 2 GW of AI-related capacity by 2027, against Australia's roughly 1.6 GW of existing computing capacity.
The post argues the AI race is becoming physical—electricity, land, grid connections, data centres, cooling, chips and regulatory approvals—and that the location of compute is now a board-level question: where data sits, where inference happens, who controls infrastructure, which laws apply and whether power supply can support it.
Countries with abundant reliable energy, data-centre developers, power generation and grid infrastructure, cooling and energy-management technology, and secure cloud providers gain power; constrained markets, slow-approval jurisdictions and AI companies that cannot secure long-term compute lose power.
For brands planning a campaign
For an institutional briefing on AI compute, energy security and data sovereignty, this is the strongest core piece: it offers an Australia-specific investment example, a physical-infrastructure frame and board questions to anchor a speaker session.
Why are large enterprises beginning to restrict third-party AI models, and what kind of infrastructure will they buy instead?
Read more
Rodic uses a Reuters report citing The Information to describe Palantir, Nvidia and Booz Allen limiting or reconsidering advanced third-party AI models over proprietary-data and intellectual-property concerns.
The post notes Palantir has reportedly pressed Anthropic for irrevocable zero-data-retention guarantees, Nvidia limits Anthropic use to less-sensitive work and relies on its own models internally, and Booz Allen has barred Anthropic's commercial model for proprietary cybersecurity work, while Microsoft promotes isolated cloud environments.
They argue enterprise AI procurement is moving toward architecture-level guarantees—data residency, zero retention, model isolation, identity, access control, auditability and contractual control over intellectual property—so the winning enterprise stack may be the system institutions can trust with crown-jewel data, not the smartest public model.
Sovereign and private AI infrastructure, secure cloud, identity, zero-trust, model gateways, auditability and governance providers gain power; model vendors that require broad data access or cannot enforce retention and use guarantees lose power.
For brands planning a campaign
This gives a speaker session a ready-made enterprise buying framework and a named cast of vendors, useful for translating data-sovereignty risk into procurement and investment criteria.
How does a gas-storage gap connect to rearmament, AI data centres and industrial competitiveness?
Read more
Rodic opens with European gas storage only about 67% full, well below normal seasonal levels, while disruption in the Gulf has removed major Qatari and UAE LNG volumes and industry estimates put Asian spot LNG near $40/mmBtu in a cold winter.
The post says Europe is better prepared than in 2022 and faces no immediate supply-security risk, but lower inventories leave the region exposed to weather, global LNG prices and cargo competition.
They connect energy security to three priorities that depend on reliable and affordable power: rearmament, AI and data-centre expansion, and industrial competitiveness with China.
Higher energy costs can weaken manufacturers and influence where factories, semiconductor facilities and data centres are built; defence production depends on dependable power for metals, chemicals, machinery, electronics and munitions.
U.S.
LNG exporters, storage operators, power producers, grid developers, energy-efficiency providers and resilient energy infrastructure gain power; European energy-intensive industries, defence manufacturers exposed to high input costs and price-sensitive Asian LNG importers lose power.
For brands planning a campaign
This is the clearest energy-security piece for a cross-sector audience: it links LNG inventories to defence production and AI infrastructure, framing energy as a constraint on industrial strategy rather than a utility-sector topic.
How does a diesel price spike become a small-business problem at the supermarket and on the construction site?
Read more
Rodic reports that European diesel futures reached record levels and have more than doubled since the beginning of 2026, with refinery disruptions in the Middle East and Russia reducing diesel supply while many refineries elsewhere already run close to capacity.
The post shows how the market move affects everyday operations: diesel powers trucks, construction equipment and agricultural machinery, so higher fuel bills flow through transport contracts, food, construction and logistics, eventually reaching supermarket prices.
The clearest example contrasts large companies with hedging teams, purchasing power and capital access with a small transport company with twenty trucks, where someone must decide whether to raise prices, accept a smaller margin, postpone investment or cut costs elsewhere.
Rodic ends with a direct policy point: Europe should remember smaller businesses when discussing energy policy.
For brands planning a campaign
This is the set's most concrete operational example, useful for audiences beyond energy specialists—especially in a speaker session on energy costs, small-business pressure and supply-chain inflation.
Recurring method: board-level framing with named winners and losers4
Subjects a brand can brief on4
Energy security as an industrial and defence constraint1
Diesel costs and the physical economy1
Data sovereignty and enterprise AI procurement1
AI compute as sovereign industrial infrastructure1
Energy security as an industrial and defence constraint
In the European gas post, Rodic starts with storage about 67% full and disrupted Qatari and UAE LNG volumes, then argues that rearmament, AI and data-centre expansion, and industrial competitiveness all depend on reliable, affordable power.
Read more
The post links higher energy costs to where factories, semiconductor plants and data centres get built. It names potential winners, including US LNG exporters, storage operators and grid developers, and losers, including energy-intensive European manufacturers and defence suppliers sensitive to input costs.
Macro event to business constraint to sector winners and losers to indicators to watch.
Rodic also states that Europe is better prepared than in 2022 and faces no immediate supply security risk; the warning depends on weather, prices and competition for cargoes.
For brands planning a campaign
Suited to briefings where a brand wants an energy-security frame connecting fuel supply to industrial strategy, defence production and AI investment, rather than a pure commodity-market update.
Rodic treats record European diesel futures and refinery disruptions in the Middle East and Russia as a business problem that flows through transport contracts, food, construction, logistics and supermarket prices.
Read more
The post contrasts large companies with hedging teams and a small transport company with twenty trucks deciding whether to raise prices, accept thinner margins or postpone investment. The key distinction is a temporary shock versus a permanent cost.
Moves from futures prices to invoices, then to the decisions managers face.
For brands planning a campaign
Useful for logistics, transport, construction, agriculture or SME-focused campaigns explaining cost pass-through and operational choices, not just headline inflation.
In the data sovereignty post, Rodic cites reported restrictions by Palantir, Nvidia and Booz Allen on advanced third-party AI models to argue that the enterprise battle is shifting from model quality to controls over proprietary data.
Read more
The post highlights zero-data-retention guarantees, data residency, model isolation, identity, access control, auditability and contractual control over intellectual property. They argue the winning enterprise stack may be the system institutions can trust with sensitive data, not the smartest public model.
For brands planning a campaign
Relevant for enterprise AI, cybersecurity, sovereign cloud and regulated-industry campaigns focused on procurement architecture and board risk rather than model benchmarks.
Rodic reads Anthropic's planned 2.16-gigawatt Australian data-centre lease, Nvidia's separate plan with Australian partners for up to 2 GW by 2027 and Australia's existing roughly 1.6 GW of computing capacity as evidence that AI is becoming an infrastructure race.
Read more
The post lists its physical dependencies: electricity, land, grid connections, data centres, cooling, chips and regulatory approvals. It asks where data sits, where inference happens, who controls infrastructure, which laws apply and whether power supply can support it.
Breaks AI down into physical inputs and jurisdiction questions, then maps who gains power—including energy-rich countries, data-centre and grid developers, cooling and secure cloud—and who loses it, including constrained markets, slow approval regimes and AI firms without long-term power.
For brands planning a campaign
Strong fit for infrastructure, energy, data-centre, sovereign AI or investment briefings showing how compute is an industrial and jurisdictional asset.
Recurring method: board-level framing with named winners and losers
Across the four posts, Rodic uses a consistent structure: a headline claim, why it matters, board-level implications, who gains or loses power, commercial or investment relevance, and a short watch list.
Read more
The posts target decision-makers seeking the business consequences of energy, AI or policy developments. The pattern holds whether the trigger is LNG storage, diesel futures, enterprise AI contracts or data-centre leases.
For brands planning a campaign
A brand can consider Rodic for panel moderation, executive briefings or explainer content built around a repeatable risk-and-opportunity frame, not a data dump.
The posts cover four practical topic areas: European energy security and LNG vulnerability; diesel and physical-economy cost pressure; enterprise AI data sovereignty and governance; and AI compute as sovereign infrastructure.
Read more
They reference Reuters reporting, named companies such as Anthropic and Nvidia, market prices, storage levels and infrastructure capacity, and connect those facts to company or government choices.
For brands planning a campaign
For the stated infrastructure investment briefing, the clearest angle is how energy and data sovereignty shape where AI compute is built, who controls it and what institutions require before adopting AI. A second angle is industrial cost pressure from energy and diesel.
These posts state no sponsorship, employment, client, affiliate, investment or disclosed partnership.
Read more
Anthropic, Nvidia, Palantir, Booz Allen and Microsoft appear as subjects of Rodic's analysis of AI infrastructure and data-sovereignty procurement, not as brands they work with.
Reuters and The Information appear only as cited reporting.
Ask Rodic to disclose any other commercial relationships.
Aleksandar Rodic's posts work the same seam: the physical and legal foundations that decide who can compete in AI and energy.
Read more
Start with the post on Anthropic's planned Australian data-centre campus, which gives the fullest version of the compute-as-infrastructure argument, then read the enterprise data-sovereignty post for the buying side and the two energy posts on European gas storage and diesel costs for the cost and security angle.
The format is short, structured commentary built on news reports: a headline claim, a why-it-matters translation, board-level implications, a list of who gains and loses power, and a watch list of indicators.
The posts work as briefings rather than technical or data-heavy treatments, with their value in framing reported facts.
For a different vantage point on these subjects, explore the creators below for delivery-side UK infrastructure, systems engineering or IT-channel strategy.
AI infrastructureEnergy securityData sovereigntyData centresPower and grid constraintsEnterprise AI procurementIndustrial cost pressureEnergy and compute infrastructureNews-driven analysisExecutive briefingExecutives and boardsInvestors and analysts
Allen covers the physical constraints behind digital growth — power equipment, cooling, chips, supply chains, grid capacity and skilled labour — and treats them as the factors that decide whether projects get delivered.
Read more
Their Capacity Europe panel, 'Scaling Under Scarcity: Delivering Projects Amid Shortages', draws on exactly the bottlenecks Aleksandar Rodic uses to map who gains and loses from AI compute. Both connect data-centre demand to energy and infrastructure strategy rather than software.
Allen writes from inside industry events, with a deliberately domestic UK lens on planning reform, regional growth, investment appetite and skills.
They treat quantum readiness as a practical infrastructure question involving power, cooling and site design.
Aleksandar Rodic works in the other direction, starting from reported news events and translating them into board and investment implications across Europe and Australia.
Allen's posts are convening and attendance notes on what the UK market is discussing; Rodic's are structured analysis with named winners, losers and watch items.
Freedman makes the same core claim from the engineering side: power has moved from an optimisation variable to the binding constraint on AI compute, and the physical plant — power, cooling, water, grid interconnects and backup generation — is now inseparable from systems architecture.
Read more
Both writers cite hyperscalers denying server expansion because of regional power limits, and see the compute and physical infrastructure layers as tightly coupled.
Freedman writes from a distributed-systems and time-series data background, focusing on the operational control loop, telemetry, workload placement and where compute can actually run.
Their work also spans conference and product activity, including the Ignition Community Conference and industrial control combined with Postgres.
Aleksandar Rodic stays at the market and policy layer — storage levels, LNG prices, zero-retention contracts and data-centre leases — rather than implementation detail.
Gitter argues the same asymmetry as Aleksandar Rodic: roughly half the AI build-out spend sits outside silicon, in power infrastructure, cooling and facilities, while hyperscalers lock up inputs — first GPUs, then memory, ultimately electricity — well beyond the current year.
Read more
Both see the binding constraint moving down the stack towards power and grid connection, and point to sovereign and edge deployments as territory hyperscalers will not own.
Gitter addresses the IT channel directly and ends with positioning advice: build a hyperscaler practice or serve constrained customers, with the promise that a hyperscaler's capacity runway becomes yours.
Aleksandar Rodic addresses boards and investors, building each case from a named news event, with energy security, fuel costs and data sovereignty supplying the argument.
Gitter's posts offer commercial guidance for resellers and integrators; Rodic's frame decisions for institutions.
Rodic's following is small and specialist, making a niche institutional briefing a reasonable fit to judge on the work itself.
Read more
Any placement that depends on measured reach or a verified audience profile needs more data.
On October 11, 2026, the LinkedIn account carrying these posts showed 6,988 followers.
That is the only measured audience number in the record: there are no engagement rates, follower geography, industry, age or seniority breakdowns, or results from past sponsored posts or briefings.
The posts are built like short board memos: a news hook, then sections headed "Why it matters", "Board level implication", "Who gains power", "Who loses power", "Commercial / investment relevance" and "Watch next".
They assume readers already track energy and AI infrastructure markets and want to understand the consequences for capital allocation, procurement or policy.
This shows who the writing is aimed at, not who actually follows the account.
A follower count near 7,000 is consistent with a specialist rather than a mass following, which fits a closed institutional briefing on AI compute, energy security or data sovereignty.
It does not show whether those followers are investors, operators, policymakers or general readers, and it is not a delivery forecast for a sponsored post.
Ask Rodic for LinkedIn audience data—country, industry and seniority breakdowns, or engagement figures from previous sponsored work—and for a recent talk or briefing on AI compute, energy security or data sovereignty with attendee feedback.
Request comparable topics, formats and depth rather than a single best-performing item.
LinkedIn followers
6,988
LinkedIn profile (the account publishing the reviewed posts) · 2026-10-11
The four posts by Aleksandar Rodic offer analytical commentary on energy security, AI compute and data sovereignty.
Read more
None carries a sponsorship disclosure, branded-post label, sponsored series, media kit, rate card or stated partnership or booking route.
They do not specify what a paid engagement would involve, its cost or available formats.
For an institutional briefing, contact Rodic through the LinkedIn account and ask about format, availability and terms; these are not stated in the posts.
Yes — book Rodic for this briefing, based on the published work, with delivery style still to be confirmed.
Read more
Brief
An infrastructure investment firm is choosing a speaker for a 45-minute institutional client briefing on AI compute, energy security and data sovereignty.
Reasons
The three requested themes are central to the published posts, and Rodic connects them rather than treating them separately. The data-centre piece examines Anthropic's planned 2.16-gigawatt Australian campus about 250 km from Brisbane, expected to begin coming online in 2027 for inference rather than training, alongside Nvidia's announced plans with Australian partners for up to 2 GW of AI capacity by 2027, against Australia's roughly 1.6 GW of existing computing capacity. It then reduces AI to its physical inputs — electricity, land, grid connections, data centres, cooling, chips and regulatory approvals. The energy piece links storage about 67% full and disrupted Qatari and UAE LNG volumes to rearmament, AI and data-centre expansion, and industrial competitiveness. The sovereignty piece turns enterprise AI governance into procurement architecture: data residency, zero retention, model isolation, identity, access control, auditability and contractual control over intellectual property.
The written format translates directly into a session. Every post follows the same shape: a headline claim, why it matters, board-level implications, who gains and loses power, commercial or investment relevance, and a short watch list. That gives a briefing a ready-made spine rather than a run of news items.
The posts are written for exactly the room described. They assume a reader who tracks energy and AI infrastructure markets and wants implications for capital allocation, procurement or policy, not a general-interest summary. A session on where compute gets built, who controls it and what institutions require before adopting AI closely matches the brief.
Open items
No talk, video, audio or transcript of Rodic speaking appears in their posts, and the posts do not mention attendee feedback from a previous briefing. The content match is strong, but live Q&A remains untested. Ask for a recent talk or briefing recording, or at minimum references from one, before locking the session.
The only audience figure available is 6,988 LinkedIn followers. Ask whether country, industry or seniority breakdowns and engagement data are available.
No media kit, rate card or booking route appears in Rodic's posts. Contact them through the LinkedIn account and confirm format, availability and terms in writing.
When to choose a specialist
No — decline or defer based on what Rodic's posts show. The subject and audience are a genuine mismatch, and commercial and delivery details remain unconfirmed.
Read more
Brief
A consumer-facing retail brokerage wants a paid LinkedIn series promoting account sign-ups, judged on reach against a target number of impressions and click-throughs, aimed at a broad investing audience.
Reasons
The subject matter points to institutions, not retail investors. The posts examine European gas storage levels, refinery disruptions in the Middle East and Russia, zero-data-retention guarantees demanded of AI vendors, and gigawatt-scale data-centre leases. None addresses individual investors, savings, portfolio construction or consumer products.
Rodic's posts provide no basis for a reach or conversion commitment. The only audience figure available is 6,988 followers. Ask for engagement rates, audience geography, seniority and industry breakdowns, and results from previous sponsored posts before setting targets.
The four posts carry no sponsorship disclosures or branded-post labels, and there is no media kit, rate card, published series or booking channel. Ask Rodic directly whether they accept paid work and what formats and terms they offer.
There is no video, audio or transcript evidence of Rodic presenting. Ask for a recent recording before assessing a brief that depends on demonstrated on-camera or on-stage delivery.
Contrast
The institutional briefing asks what the work already argues; this brief requires an audience and deliverable the work does not establish.
A worked campaign example
Angle
"Where AI compute actually gets built — and who gets to control it." A 45-minute session: roughly 25 minutes of framing, followed by 20 minutes of questions based on the client's own portfolio positions.
Brief
Institutional client briefing for an infrastructure investment firm — AI compute, energy security and data sovereignty.
Client takeaway
A one-page grid mapping the board questions — where data sits, where inference happens, who controls infrastructure, which laws apply, and power adequacy — to the firm's own holdings, plus the watch list above as the monitoring set. The firm assembles this from the session; it is not Rodic's deliverable.
Next steps
Ask Rodic for a recent talk, briefing or panel recording, or references from one, before confirming the date.
Ask whether audience data is available for the LinkedIn account — country, industry and seniority breakdowns, or engagement figures from previous sponsored work.
Ask directly whether Rodic offers commissioned briefings, and in what format and on what terms. Treat the engagement route as unconfirmed until they reply.
Run of session
Open on Australia as the concrete case: Anthropic's first Australian data-centre lease for a planned 2.16-gigawatt campus about 250 km from Brisbane, expected to start coming online in 2027 for inference; Nvidia's announced plans with Australian partners for up to 2 GW by 2027; and Australia's roughly 1.6 GW of existing computing capacity.
Break compute into what it consumes: electricity, land, grid connections, data centres, cooling, chips and regulatory approvals. The AI race is becoming an infrastructure race, making the location of compute a capital-allocation question, not a technical footnote.
Move to the jurisdiction questions facing the client's own portfolio: where does the data sit, where is inference happening, who controls the infrastructure, which laws apply, and can the electricity supply support it?
Turn to procurement using the sovereignty material: Palantir, Nvidia and Booz Allen restricting or reconsidering advanced third-party AI models over proprietary-data and intellectual-property concerns; Palantir reportedly pressing Anthropic for irrevocable zero-data-retention guarantees; Nvidia limiting Anthropic use to less-sensitive work and relying on its own models internally; Booz Allen barring Anthropic's commercial model for proprietary cybersecurity work; and Microsoft promoting isolated cloud environments. The conclusion for the room: the winning enterprise stack may be the one institutions trust with crown-jewel data rather than the smartest public model.
The LNG post adds an energy constraint: storage is about 67% full, major Qatari and UAE volumes have been removed from the market by Gulf disruption, and industry estimates put Asian spot LNG near $40/mmBtu in a cold winter. There is no immediate supply-security risk, but exposure to weather, prices and competition for cargoes is real. For infrastructure investors, the link is clear: rearmament, AI and data-centre expansion, and industrial competitiveness all depend on reliable, affordable power, influencing where factories, semiconductor plants and data centres get built.
The watch list: Australian foreign-investment approval, who pays for new grid capacity, the next hyperscale leases, energy prices around major AI clusters, and whether governments increasingly require sensitive AI workloads and data to stay inside their own jurisdictions.
Choose Simon Allen when the session is about how the build gets delivered rather than why it is happening.
Read more
Allen co-founded the Quantum Infrastructure Council and moderates Capacity Europe 2026's panel "Scaling Under Scarcity: Delivering Projects Amid Shortages", which examines power equipment, chips, cooling, supply chains and skilled resources — the constraints that decide whether a project lands on time.
Aleksandar Rodic's posts treat planning reform, grid constraints, regional growth, investment appetite and supply-chain delivery as UK-specific problems, and connect quantum readiness to power, cooling, site design and long-term strategy.
If the client's question is who can deliver this in the UK and what is slowing them down, Allen is the closer fit for the same briefing slot.
The UK delivery lens narrows the room.
Aleksandar Rodic's posts focus on conference attendance notes and panel promotion rather than a sustained argument.
They include no market figures, board-implication structure, or discussion of data sovereignty or enterprise AI procurement — two of the three themes in a campaign aimed at institutional investors.
An institutional investor seeking a capital-allocation frame around AI compute would get less of it here and should ask for the numbers and jurisdiction analysis.
Choose Michael Freedman when the audience is technical and the question is how compute and the physical plant became one system.
Read more
Freedman's background is in distributed systems, networking and storage — server utilization, network bisection, storage placement, scheduling and cluster orchestration.
Aleksandar Rodic's posts explain why power moved from an optimisation variable to the binding constraint: hyperscalers denying server expansion because of regional power limits, an operational control loop responding to thermal and workload conditions, and the shift from air to liquid cooling.
Freedman speaks at the Ignition Community Conference with Tiger Data and Inductive Automation about industrial control plus Postgres at the scale of the physical world.
For a briefing aimed at engineers, operators or a data-platform team, this is the deeper systems account.
The depth is technical and product-adjacent.
The material is written around Tiger Data and Inductive Automation work and reads as architecture commentary rather than institutional market analysis.
An investment audience looking for actionable evidence should ask for energy-market detail, sovereignty and procurement analysis, and specific deal figures.
Choose Eric Victor Gitter when the session should be about where the money goes outside silicon.
Read more
Aleksandar Rodic's posts break the roughly $6.7 trillion spend to build AI-ready data centres by 2030 into its non-compute halves — power infrastructure at 15–25% against 5–10x rack power density, a $1.3 trillion utility capital call, cooling at 10–15% as the second-largest capital expense, and about $800 billion in facilities build-out.
The Stargate post reads OpenAI's memory deals with Samsung, SK hynix and Micron as the hyperscalers locking up inputs long term, and describes the market splitting into companies that work with hyperscalers and everyone else.
If the client wants the supplier map and who wins the non-silicon spend, this is more directly on point than a compute-and-sovereignty session.
The framing is built for the IT channel, not institutional capital.
Both posts end with advice to MSPs, systems integrators and resellers about positioning against hyperscaler capacity, closing on questions like where you are building.
The spending splits are broad decade-level estimates rather than figures tied to disclosed projects.
A firm briefing its own clients on jurisdiction and procurement risk should ask Gitter to build that layer for the room, or draw on Rodic's material.
Four LinkedIn posts by Aleksandar Rodic observed on 11 October 2026 cover Australian AI data-centre and power capacity, enterprise data sovereignty, European gas storage and LNG risk, and diesel costs.
The posts do not show individual post dates.
They cite Reuters, other outlets and company announcements; figures and claims are reported as the posts present them.
The similar-creators comparison covers two posts each from Simon Allen, Michael Freedman and Eric Victor Gitter.
No video, audio or transcript of Rodic speaking, audience demographics, or documented sponsorship or partnership formats appear in these posts.